oEPA
EPA-231-B-24-001 | December 2024
Technical Guidance for Assessing
Environmental Justice in Regulatory Analysis
Second Edition
-------
Disclaimer: This document identifies internal Agency policies and recommended
procedures for EPA employees. This document is not a rule or regulation, and it may not
apply to a particular situation based upon the circumstances. This guidance does not
change or substitute for any law, regulation, or any other legally binding requirement and is
not legally enforceable. As indicated by the use of non-mandatory language such as
"guidance," "recommend," "may," "should," and "can," it identifies policies and provides
recommendations and does not impose any legally binding requirements.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
i
-------
Version
This document updates and supersedes the 2016 Technical Guidance for Assessing Environmental
Justice in Regulatory Analysis and reflects advancements in the state of the science, new peer-
reviewed Agency economic and risk assessment guidance, and lessons learned from conducting
environmental justice analysis for EPA rulemakings since publication of the original guidance.
Acknowledgements
We extend our appreciation to Stefani Penn, Cynthia Manson, Lena Flannery and others at Industrial
Economics, Incorporated for their technical support and diligent attention to detail. The U.S. EPA also
gratefully acknowledges the constructive comments and suggestions on how to improve the document
provided by EPA program office staff, the public, Tribes, and EPA Science Advisory Board members.
Contributors
The U.S. Environmental Protection Agency (U.S. EPA), Office of Policy staff that coordinated and
contributed to this revision are Ann Wolverton, Peiley Lau, Chris Dockins, Jennifer Bowen, and Wes
Austin. Industrial Economics, Incorporated provided technical and editing support on the revision. The
document was carefully reviewed internally by EPA program office staff and management.
Public Comment and Tribal Consultation
The draft revision of this document was published in the Federal Register for a 60-day public comment
period, which was then extended for an additional 14 days. The EPA also conducted a Tribal
consultation process concurrent with the public comment period.
Peer Review
The revised technical guidance was peer reviewed by external and independent experts through the
EPA's Science Advisory Board (SAB). For information on the SAB process and the final report, see
https://sab.epa.g0v/0rds/sab/r/sab apex/sab/advisorvactivitydetail?p18 id=2642&clear=18&session=66
57841334402.
Recommended Citation
U.S. Environmental Protection Agency (EPA). 2024. Technical Guidance for Assessing Environmental
Justice in Regulatory Analysis. Second Edition. EPA 231-B-24-001.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
-------
Table of Contents
Glossary vi
Acronyms and Abbreviations xii
Chapter 1: Introduction 1
1.1 How Is This Guidance Document Organized? 2
Chapter 2: Key Definitions 4
2.1 Disproportionate 4
2.2 Population Groups of Concern 4
2.2.1 Race, Ethnicity, Tribal Affiliated and Indigenous Populations, and National Origin 6
2.2.2 Low-Income Populations 7
2.2.3 Disability Status 9
2.2.4 Populations that Rely on Cultural and Subsistence Practices 10
2.3 Meaningful Involvement or Engagement 11
Chapter 3: Key Analytic Considerations 14
3.1 Questions to Evaluate When Analyzing EJ Concerns 14
3.2 Identifying Objectives, Data, and Other Information 14
3.3 Main Steps of an EJ Analysis 16
3.4 Recommendations for Analyses of EJ Concerns 18
Chapter 4: Contributors to Environmental Justice Concerns 21
4.1 Contributors to Higher Exposure to Environmental Hazards 22
4.1.1 Proximity to Emissions and Discharges from Nearby Sources 22
4.1.2 Unique Exposure Pathways 23
4.1.3 Physical Infrastructure 23
4.1.4 Exposure to Multiple Stressors and Cumulative Exposures 23
4.1.5 Monitoring, Compliance, and Enforcement 24
4.1.6 Community Capacity to Meaningfully Participate in Decision-Making 24
4.2 Contributors to Increased Consequences of Exposure 25
Chapter 5: Considering Environmental Justice Concerns When Planning a Human
Health Risk Assessment 27
5.1 Introduction 27
5.2 Overview of Modeling and Data Needs for Evaluating EJ Concerns 28
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
-------
5.2.1 HHRA Requires Sufficient Data and Modeling Capabilities to Incorporate EJ
Considerations 28
5.2.2 It Can Be Difficult to Incorporate the Cumulative Effects of Multiple, Dissimilar Stressors
into HHRA 30
5.2.3 HHRA Requires Effective Communication, Given its Highly Technical Nature 30
5.3 Considering EJ Concerns When Planning a HHRA 31
5.3.1 Overview of the HHRA Process 31
5.3.2 Planning and Scoping 35
5.3.3 Problem Formulation 39
5.4 Multiple Exposures and Cumulative Effects 46
5.4.1 Cumulative Risk Assessment 46
5.4.2 Cumulative Impact Assessment 47
Chapter 6: Conducting Regulatory Analyses to Assess Environmental
Justice Concerns 49
6.1 Preliminary Analysis of EJ Concerns 50
6.2 Characterizing the Baseline and Regulatory Options for In-Depth EJ Analysis 52
6.3 Data and Information to Assess EJ Concerns 54
6.4 Analytic Methods 59
6.4.1 Proximity-Based Analysis 60
6.4.2 Exposure and Risk Modeling 63
6.4.3 Combining Qualitative and Quantitative Approaches 64
6.5 Analytic Considerations 65
6.5.1 Geographic and Temporal Scope 65
6.5.2 Comparison Population Group 66
6.5.3 Spatial Identification and Aggregating Effects 68
6.5.4 Facility vs Community-Based Perspectives 72
6.5.5 Evaluating Underlying Variability and Identifying Potential Hot Spots 72
6.6 Characterizing Analytic Results 75
6.6.1 Choice of Summary Metrics 75
6.6.2 Displaying Results Visually 76
6.6.3 Statistical Significance and Other Considerations 79
6.7 Assessing the Distribution of Costs and Other Effects 80
6.7.1 Distribution of Economic Costs 80
6.7.2 Considering Compliance and Enforcement 82
6.7.3 Other Effects and Considerations 83
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition iv
-------
Chapter 7: Research Priorities to Fill Key Data and Methodological Gaps 84
7.1 Building Consistency in Terminology and Definitions 85
7.2 Enhancing Consideration of EJ in Human Health Risk Assessment 85
7.2.1 Planning and Scoping and Problem Formulation 85
7.2.2 Effects Assessment 85
7.2.3 Exposure Assessment 86
7.2.4 Risk Characterization 86
7.3 Cross-Cutting Analytic Needs 87
7.4 Meaningful Engagement and Risk Communication 88
7.5 Incorporating EJ into Regulatory Analysis 88
7.5.1 Data Gaps for Evaluating EJ Concerns 88
7.5.2 Methodological Gaps for Evaluating EJ Concerns in the Baseline and for Regulatory
Options 89
7.5.4 Other Analytic Considerations 90
7.6 Other General Needs 90
7.7 Next Steps 90
References 92
Appendix A: Select EPA Guidance and Other Documents A-1
Appendix B: Incorporating EJ Concerns into Exposure and Effects Assessments B-1
Planning for an Exposure Assessment B-1
Questions and Key Considerations B-1
Planning For an Effects Assessment B-6
Questions and Key Considerations B-7
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition v
-------
Glossary
The purpose of the Glossary is to aid analysts in understanding terms used throughout the Technical
Guidance for Assessing Environmental Justice in Regulatory Analysis. The EPA recognizes that many
definitions will continue to evolve and may be updated.1
Agency action: includes rulemakings and issuance of policy statements, orders, licenses, permits, and
other actions (Administrative Procedure Act, 1946).
American Indian or Alaska Native: an individual with origins in any of the original peoples of North,
Central, and South America (OMB, 2024).
Ancestry: a person's ethnic origin or descent, "roots," or heritage, or the place of birth of the person or
the person's parents or ancestors before their arrival in the United States (U.S. Census Bureau, 2024).
Background exposures: potential exposures to stressors due to background levels of both naturally
occurring and anthropogenic sources (U.S. EPA, 2024I).
Baseline: an analytically reasonable forecast of the way the world would look absent the regulatory
action being assessed, including any expected changes to current conditions over time (OMB, 2023b).
Bioaccumulation: a process by which chemicals are taken up by an organism either directly from
exposure to a contaminated medium or by consumption of food containing the chemical (U.S. EPA,
2010b).
Comparison population group: the effects of a regulatory action on population groups of concern
need to be presented in relation to another group, which can be defined as individuals with similar
socioeconomic characteristics in areas of the state, region, or nation unaffected by the regulatory action
(i.e., within-group comparison) or as individuals with different socioeconomic characteristics within the
affected areas (i.e., across-group comparison).
Contaminant: any physical, chemical, biological, or radiological substance found in air, water, soil, or
biological matter that can have a harmful effect on people, animals, or plants. Also, see stressor.
Cumulative impact assessment: the process of accounting for cumulative impacts in the context of
problem identification and decision-making (U.S. EPA, 2022c).
Cumulative impacts: the totality of exposures to combinations of chemical and nonchemical stressors
and their effects on health, well-being, and quality of life outcome throughout a person's lifetime (U.S.
EPA, 2022c).
Cumulative risk assessment: an analysis, characterization, and possible quantification of the
combined risks to human health and/or the environment from multiple agents and/or stressors (U.S.
EPA, 2003a).
1 Additional information about relevant EJ-related terminology is available at https://www.epa.aov/environmentaliustice/learn-
about-environmental-iustice.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition vi
-------
Difference or differential: in this document, an analytically discernible (or measurable) distinction in
effects or risks across population groups.
Disability: an individual with a disability is someone who has a physical or mental impairment that
substantially limits one or more major life activities, has a record of such an impairment, or is regarded
as having such an impairment (Americans with Disabilities Act, 1990).
Disproportionate: in this document, differences in effects or risks that are extensive enough that they
may merit Agency action and should include consideration of cumulative impacts or risks where
appropriate (U.S. EPA, 2022a).
Distributional analysis: characterizing how benefits and costs of a regulation are ultimately
experienced across the population and economy, divided up in various ways (OMB, 2023b).
Dose: the amount of a substance that enters a target in a specified period of time after crossing an
exposure surface (U.S. EPA, 2019a).
Dose-response assessment: analysis of the relationship between the total amount of an agent
administered to, taken up by, or absorbed by an organism, system, or target population and the
changes developed in that organism, system, or target population in reaction to that agent, and
inferences derived from such an analysis with respect to the entire population (U.S. EPA, 2019a).
Effects: risks, exposures, and outcomes caused by a chemical, activity, or process as it comes into
contact with humans or the environment, sometimes used interchangeably with impacts.
Effect-modifier: another variable that when present causes the effect of the exposure on the outcome
to differ (U.S. EPA, 2005a).
Environmental justice: the just treatment and meaningful involvement of all people, regardless of
income, race, color, national origin, Tribal affiliation, or disability, in agency decision-making and other
Federal activities that affect human health and the environment so that people are fully protected from
disproportionate and adverse human health and environmental effects (including risks) and hazards,
including those related to climate change, the cumulative impacts of environmental and other burdens,
and the legacy of racism or other structural or systemic barriers; and have equitable access to a
healthy, sustainable, and resilient environment in which to live, play, work, learn, grow, worship, and
engage in cultural and subsistence practices (E.O. 14096).
Environmental justice concern: the actual or potential lack of just treatment or meaningful
involvement of any population group, community, or geographic area (e.g., associated with differences
in income, race, color, national origin, Tribal affiliation, or disability status) in the development,
implementation, and enforcement of environmental laws, regulations, and policies (U.S. EPA, 2015a).
For analytic purposes, it refers specifically to disproportionate and adverse health and environmental
effects that may exist prior to or be created by the regulatory action. See environmental justice.
Environmental stressor: In this document, the range of chemical, physical, or biological agents,
contaminants, or pollutants that may be subject to a rulemaking.
Equity: the consistent and systematic treatment of all individuals in a fair, just, and impartial manner,
including individuals who belong to communities that often have been denied such treatment (E.O.
14091).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
vii
-------
Exposure: human contact with environmental contaminants in media including air, water, soil, and food
through inhalation, ingestion, or direct contact with the skin or eye.
Exposure assessment: the process of estimating or measuring the magnitude, frequency, and
duration of exposure to an agent and the size and characteristics of the population exposed (U.S. EPA,
2019a).
Exposure pathway: the course a chemical or contaminant takes from its source to the person being
contacted (U.S. EPA, 2019a).
Extrinsic factors: factors or conditions acquired over a person's lifetime (e.g., socioeconomic status,
disease status, stress, nutrition, lifestyle, workplace, geography, previous or ongoing exposure to
multiple chemicals) that may contribute to increased vulnerability (U.S. EPA, 2022b).
Fit-for-purpose: the concept that risk assessments and associated products should be suitable and
useful for their intended purpose(s), particularly for informing choices among risk management options
(U.S. EPA, 2014b).
Hazard: inherent property of an agent, contaminant, or situation having the potential to cause adverse
effects when an organism, system, or population is exposed to that stressor (U.S. EPA, 2011a).
Hazard identification: the process of determining whether a stressor has the potential to cause harm
to humans and/or ecological systems, and if so, under what circumstances (U.S. EPA, 2014b).
Health impact assessment: a systematic process that uses an array of data sources and analytic
methods and considers input from stakeholders to identify the potential effects of a proposed policy,
plan, program, or project on the health of a population and the distribution of those effects within the
population. HIA also provides recommendations on monitoring and managing those effects (NRC,
2011a).
Hot spot: a geographic area with a high level of pollution/contamination within a larger geographic area
of more "normal" environmental quality (U.S. EPA, 2022i).
Human health risk assessment (HHRA): the process to estimate the nature and probability of
adverse health effects in humans who may be exposed to chemicals or other stressors in contaminated
environmental media, now or in the future (https://www.epa.gov/risk/human-health-risk-assessment).
Indigenous Knowledge/Traditional Ecological Knowledge: a body of observations, oral and written
knowledge, innovations, practices, and beliefs developed by Tribes and Indigenous Peoples through
interaction with and experience with the environment (OSTP and CEQ, 2022).
Indigenous Peoples: includes state-recognized Tribes; Indigenous and Tribal community-based
organizations; individual members of federally recognized Tribes, including those living on a different
reservation or living outside Indian country; individual members of state-recognized Tribes; Native
Hawaiians; Native Pacific Islanders; and individual Native Americans (U.S. EPA, 2014a).
Intrinsic factors: biological conditions or factors that cannot be altered (e.g., age, gender, genetic
conditions) that may contribute to increased vulnerability (U.S. EPA, 2022b).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
viii
-------
Lifestage: a temporal stage of life with distinct anatomical, physiological, behavioral, or functional
characteristics that contribute to potential differences in vulnerability to environmental exposures (U.S.
EPA, 2019a).
Low-income: individuals, households, or populations characterized by limited economic resources.
The OMB has designated the U.S. Census Bureau's annual poverty measure as the official metric for
program planning and analysis (OMB, 1978), although other definitions exist and may be appropriate
as supplemental measures.
Meaningful involvement or Engagement: actions that agencies take to engage persons or
communities with EJ concerns that are potentially affected by Federal activities by: providing timely
opportunities for members of the public to share information and concerns and participate in decision-
making processes; fully considering public input provided as part of decision-making processes;
seeking out and encouraging the involvement of persons and communities affected by Federal
activities; and providing technical assistance, tools, and resources to assist in facilitating meaningful
and informed public participation, whenever practical and appropriate (E.O. 14096).
Non-environmental stressor: factors found in the built, natural, and social environments that can
directly or indirectly adversely affect health or increase vulnerability to environmental stressors (Tulve et
al., 2016)
Overburdened: in this document, population groups, communities, or geographic areas that potentially
experience disproportionate environmental harms and risks due to greater vulnerability to
environmental hazards, lack of opportunity for public participation, or other factors.
Participatory science: engages the public in advancing scientific knowledge by formulating research
questions, collecting data, and interpreting results. This includes a broad and inclusive range of
activities, from those originating in academic and government institutions that enlist the public in data
collection to create knowledge, to community-led projects intended to identify potential EJ issues and
community concerns (U.S. EPA, 2022e).
Peer review: a documented process conducted to ensure that activities are technically supportable,
competently performed, properly documented, and consistent with established quality criteria (U.S.
EPA, 2014b).
People of color: populations or individuals who list their racial status as a race other than White alone
and/or list their ethnicity as Hispanic or Latino. That is, all people other than non-Hispanic White-alone
individuals. The word alone in this case indicates that the person is of a single race, not multiracial. The
term people of color is intended to be interchangeable with the term minority populations, as that term
appears in numerous federal documents and statutes. (U.S. EPA, 2024c).
Plain language: writing that is clear, concise, well-organized, and follows other best practices
appropriate to the subject or field and intended audience (The Plain Writing Act of 2010, Public Law
111-274).
Private costs: the costs that the purchaser of a good or service pays the seller (U.S. EPA, 2024f).
Pollutant: any substance introduced into the environment that may adversely affect the usefulness of a
resource or the health of humans, animals, or ecosystems. For most environmental media, this term is
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
ix
-------
commonly understood to refer to substances introduced by human activities. Also, see stressor (U.S.
EPA, 2021 e).
Population groups of concern: in this document, any population groups being analyzed for actual or
potential EJ concerns related to the regulatory action, including cases where specific demographic or
socioeconomic characteristics correlate with increased vulnerability to environmental exposure, such as
race, ethnicity, national origin, low-income, Tribal affiliated and Indigenous populations, and disability
status in the United States and its territories and possessions.
Proximity-based analysis: analytic approach using geospatial data that uses proximity to or distance
from the source(s) of an environmental stressor to indicate a population group's likelihood of risk or
exposure when direct measurement is unavailable.
Quantitative methods: analysis of numerical data using statistical techniques that include simply
describing the phenomenon of interest or looking for differences between groups or relationships
among variables (Tashakkori et al., 2020).
Qualitative methods: encompasses a wide range of methods, such as interviews, case studies,
discourse analysis, and ethnographic research. A key distinction from quantitative methods is that
qualitative methods do not necessarily collect numerical data, and therefore frequently cannot provide
numerical results (Taherdoost, 2022; Tashakkori et al, 2020).
Regulatory action: any substantive action by an agency (normally published in the Federal Register)
that promulgates or is expected to lead to the promulgation of a final rule or regulation, including
notices of inquiry, advance notices of proposed rulemaking, and notices of proposed rulemaking (E.O.
12866; E.O. 14094).
Regulatory analysis: a tool used to anticipate and evaluate the likely consequences of regulatory
actions. It compares the baseline scenario to one or more regulatory or policy scenarios. Economic and
other effects of policies or regulations are then measured as the differences between these two
scenarios (OMB, 2023b).
Regulatory options: the alternatives under consideration for a specific regulatory action.
Risk: the probability of an adverse effect in an organism, system, or population caused under specified
circumstances by exposure to a contaminant or stressor (U.S. EPA, 2024I).
Risk analyst/assessor: one who plans and conducts a risk assessment. In particular, the risk analyst
provides a transparent description of all aspects of the risk assessment (e.g., default assumptions, data
selected, and policy choices) to make clear the range of plausible risk associated with each risk
management option.
Risk characterization: the integration of information on hazard, exposure, and dose-response to
provide an estimate of the likelihood that any identified adverse effects will occur in exposed people
(U.S. EPA, 2024I).
Risk management: in the context of human health, a decision-making process that accounts for
political, social, economic, and engineering implications together with risk-related information in order to
develop, analyze, and compare management options and select the appropriate managerial response
to a potential chronic health hazard (U.S. EPA, 2024I).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
x
-------
Social context: refers to all social and political mechanisms that generate, configure, and maintain
social hierarchies. These mechanisms can include the labor market, the educational system, political
institutions, and cultural and societal values (WHO, 2024).
Social cost: the total burden a regulation will impose on the economy. The bearers of social costs can
be either specific individuals or society at large (U.S. EPA, 2024f).
Source: the origin of potential contaminants or environmental stressors; frequently a facility or site
(U.S. EPA, 2019a).
Statistical significance: the probability that a result is not likely to be due to chance alone (U.S. EPA,
2024I).
Stressor: any chemical, physical, or biological entity that induces an adverse response (U.S. EPA,
2019a).
Structural and systemic racism: forms of racism that are pervasively and deeply embedded in
systems, laws, written or unwritten policies, and entrenched practices and beliefs that produce,
condone, and perpetuate widespread unfair treatment and oppression of people of color, with adverse
health consequences (Braverman et al., 2022).
Subsistence: dependence on indigenous fish, vegetation, and/or wildlife as the principal portion of
one's diet (CEQ, 1997).
Summary statistics: descriptive statistics that provide an overview of available data and may include
the mean, median, mode, interquartile mean, range, and/or standard deviation, etc.
Uncertainty: imperfect knowledge or lack of precise knowledge of the real world, either for specific
values of interest or in the description of the system. Although numerous schemes for classifying
uncertainty have been proposed, most focus on parameter uncertainty and model uncertainty (U.S.
EPA, 2014b).
Underserved: populations as well as geographic communities that have been systematically denied
the opportunity to participate fully in aspects of economic, social, and civic life (E.O. 14091).
Variability: inherent natural variation, diversity, and heterogeneity across time and/or space or among
individuals within a population (U.S. EPA, 2014b).
Vulnerability: characteristics of individuals or populations that place them at increased risk of an
adverse health effect (U.S. EPA, 2019a).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
xi
-------
Acronyms and Abbreviations
ACS
-
American Community Survey
AHS
-
American Housing Survey
BIE
-
Bureau of Indian Education
CCR
-
Coal Combustion Residuals
CEQ
-
Council on Environmental Quality
CIA
-
Cumulative Impact Assessment
CPI-U
-
Consumer Price Index for all Urban Consumers
CRA
-
Cumulative Risk Assessment
E.O.
-
Executive Order
ECHO
-
Enforcement and Compliance History Online
EDGE
-
Education Demographic and Geographic Estimates
EJ
-
Environmental Justice
EPA
-
Environmental Protection Agency
HFC
-
Hydrofluorocarbon
HHRA
-
Human Health Risk Assessment
HIA
-
Health Impact Assessment
HOLC
-
Home Owner's Loan Corporation
IQ
-
Intelligence Quotient
IRP
-
Integrated Review Plan
ISA
-
Integrated Science Assessment
MU
-
Management Unit
NAAQS
-
National Ambient Air Quality Standards
NEJAC
-
National Environmental Justice Advisory Council
NHANES
-
National Health and Nutrition Examination Survey
NOX
-
Nitrogen Oxide
NRC
-
National Research Council
NSTC
-
National Science and Technology Council
OMB
-
Office of Management and Budget
ORD
-
Office of Research and Development
OSTP
-
Office of Science and Technology Policy
PFAS
-
Per- and Polyfluoroalkyl Substances
PM
-
Particulate Matter
RCRA
-
Resource Conservation and Recovery Act
RIA
-
Regulatory Impact Analysis
RSEI
-
Risk Screening Environmental Indicators
SAB
-
Science Advisory Board
SES
-
Socioeconomic Status
SI
-
Surface Impoundment
TRI
-
Toxics Release Inventory
U.S.
—
United States
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
xii
-------
Chapter 1: Introduction
The purpose of this document, the Technical Guidance for Assessing Environmental Justice in
Regulatory Analysis (EJ Technical Guidance), is to outline analytic expectations and technical
approaches that can be used by Agency analysts to evaluate environmental justice (EJ) concerns for
United States Environmental Protection Agency (EPA) regulatory actions.2 The term analyst is
intentionally broad because an evaluation of EJ concerns may be conducted by economists, risk
assessors, and others, often as a team. Senior EPA managers will also find this document useful for
understanding what role analysis can play in ensuring that EJ concerns are appropriately considered in
the development of regulatory actions, to the extent practicable and consistent with applicable law.
Recommendations and best practices highlighted in the guidance are intended to bring greater
consistency across EJ analyses, while recognizing the need for flexibility to reflect policy considerations
and technical challenges within a particular regulatory context. Conducting EJ analysis is valuable,
whether it informs further Agency action or is conducted for informational purposes only, as it provides
the public with information on the distribution of environmental and health effects and risks and furthers
the EPA's mission of integrating environmental justice into its activities and processes.3
The guidance recommends that early in the rulemaking process analysts identify the extent to which a
regulatory action may raise EJ concerns that need further evaluation, including the level of analysis that
is feasible and appropriate (see Sections 3.2 and 6.1). This helps ensure that EJ concerns are given
due consideration, including informing how to avoid, minimize, or mitigate disproportionate and adverse
human health and environmental effects through regulatory design and the proposed options;
information provision; opportunities for retrospective analysis; the leveraging of statutory authorities;
and monitoring, compliance, and enforcement, among others. Based on the initial evaluation, analysts
can choose from a suite of methods depending on the data and resources available, time needed to
conduct the analysis, and other technical challenges. When conducting these evaluations, the EPA not
only considers the distribution of reductions in environmental and health effects and risks but also the
distribution of positive outcomes (U.S. EPA, 2015a).
This document is intended for use alongside other Agency guidance, including guidance on human
health risk assessment (HHRA) and economic analysis (see Appendix A).4 It will evolve with advances
in the state of the science, data, and analytic methods available to Agency analysts. Regarding risk
assessment, this technical guidance currently is limited to a discussion of how to integrate EJ into the
planning of an HHRA. The EPA has developed and continues to refine methods and guidance on a
variety of topics relevant to conducting analyses of EJ concerns in the context of a regulatory action.
2 Though not the focus of this guidance, the approaches and methods described here may also be useful for analyses that
precede and are ultimately used in support of regulation.
3 Circular A-4 (2023) states that regulatory analysis "provides a formal way of organizing the evidence on the key effects of the
various alternatives that should be considered in developing regulations. A high-quality regulatory analysis is designed to
inform policymakers, other government stakeholders, and the public about the effects of alternative actions. Regulatory
analysis can help agencies in developing regulations by clarifying the likely effects of a regulation under consideration, and it is
meant to inform the public about the anticipated consequences of government action (and alternatives)."
4 See also EPA Legal Tools to Advance Environmental Justice: Cumulative Impacts Addendum (U.S. EPA, 2023a), EPA Legal
Tools to Advance Environmental Justice: Executive Order 14096 Addendum (U.S. EPA, 2023b), and EPA Legal Tools to
Advance Environmental Justice (U.S. EPA, 2022a),, available at: https://www.epa.gov/oqc/epa-leqal-tools-advance-
environmental-iustice.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
1
-------
Such references are noted throughout the document, and future updates to the EJ Technical Guidance
may include more detail on these topics.
The Executive Orders (E.O.s) that underpin the EPA's efforts to incorporate EJ analyses into
rulemakings are summarized in Table 1.1.
Table 1.1 Executive Orders with Implications for EJ Analysis of Federal Rulemakings
Executive Order
Year
Main Directives Pertaining to EJ Analysis for Rulemakings
12898
1994
"To the greatest extent practicable and permitted by law, make environmental justice part of
their mission by identifying and addressing, as appropriate, disproportionately high and
adverse human health or environmental effects of their programs, policies, and activities on
minority populations and low-income populations in the U.S."
14094
2023
"Regulatory analysis, as practicable and appropriate, shall recognize distributive impacts
and equity, to the extent permitted by law."
14096 2023 Identify, analyze, and address:
(1) "disproportionate and adverse human health and environmental effects..., including
those related to climate change and cumulative impacts of environmental and other burdens
on communities with environmental justice concerns;"
(2) "historical inequities, systemic barriers, or actions related to any Federal regulation,
policy, or practice that impair the ability of communities with environmental justice concerns
to achieve or maintain a healthy and sustainable environment;" and
(3) "barriers related to Federal activities that impair the ability of communities with
environmental justice concerns to receive equitable access to human health or
environmental benefits..."
Note: This is not an exhaustive list of the directives contained within these E.O.s that pertain to
rulemakings or to Federal activities more generally.
1.1 How Is This Guidance Document Organized?
The first four chapters of this guidance establish the objectives, definitions, main analytic
considerations, and context for an assessment of EJ concerns in support of EPA regulatory actions:
• Chapter 1: Introduction (this chapter) provides background and outlines the main objectives of
the EJ Technical Guidance.
• Chapter 2: Key Definitions reviews key EJ concepts that are expected to influence analytic
considerations. In particular, the chapter discusses how to define the concepts of an EJ
concern; disproportionate effects; race, ethnicity, national origin, low-income, and disability;
Tribal affiliated and Indigenous populations; subsistence practices; and meaningful involvement
or engagement.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
2
-------
Text Box 1.1 Overarching Recommendations to Analysts
1. When risks, exposures, outcomes, or benefits are quantified, some level of quantitative EJ
analysis is recommended.
2. Analysts should integrate EJ into the planning of a risk assessment conducted for the regulatory
action.
3. Analysts should strive to characterize the distribution of risks, exposures, or outcomes within
each population group, not just average effects.
4. Analysts should follow best practices appropriate to the analytic questions at hand.
5. As relevant, analysts should consider any economic challenges that may be exacerbated by the
regulatory action for relevant population groups of concern.
• Chapter 3: Key Analytic Considerations discusses three questions analysts should strive to
answer when evaluating EJ concerns, outlines the key steps of an EJ analysis, and presents
overarching recommendations (Text Box 1.1) and best practices to guide assessments of EJ
concerns for EPA regulatory actions. Appendix A provides links to additional EPA guidance that
may be helpful when assessing EJ concerns.
• Chapter 4: Contributors to Environmental Justice Concerns identifies factors that can result
in differential patterns of exposure to environmental hazards for some population groups, and/or
a greater response of some individuals to a given level of exposure.
By design, Chapters 5 and 6 present some overlapping information about key concepts and methods
under the assumption that analysts may consult only one of the technical chapters. These technical
chapters provide guidance for considering EJ in two specific contexts:
• Chapter 5: Considering Environmental Justice when Planning a Human Health Risk
Assessment provides guidance on incorporating EJ concerns into the planning of an HHRA,
including descriptions of available methodologies and tools. Appendix B provides examples of
approaches for incorporating EJ concerns into the planning of exposure and effects
assessments.
• Chapter 6: Conducting Regulatory Analyses to Assess Environmental Justice Concerns
discusses how to identify and evaluate the feasibility and appropriateness of different analytic
approaches and tools for assessing EJ concerns; the types of information that should be
included in the assessment; other analytic considerations that could affect results; and how to
consider costs and non-health effects in the assessment.
The final chapter describes identified near-term research needs related to the analysis of EJ concerns:
• Chapter 7: Research Priorities to Fill Key Data and Methodological Gaps provides
information on research goals to improve assessment of EJ at the EPA.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
3
-------
Chapter 2: Key Definitions
For purposes of this document, an EJ concern is the actual or potential lack of just treatment or
meaningful involvement of any population group, community, or geographic area (e.g., associated with
differences in income, race, color, national origin, Tribal affiliation, or disability status) in the
development, implementation, and enforcement of environmental laws, regulations, and policies.5 For
analytic purposes, this concept refers specifically to disproportionate and adverse health and
environmental effects that may exist prior to or be created by the regulatory action.6
This chapter defines and discusses key terms and concepts from E.O.s 12898 and 14096 that are
important for analysts to understand before conducting an analysis of EJ concerns. These key terms
and concepts include disproportionate; race, ethnicity, national origin, low-income, and disability, Tribal
affiliated and Indigenous populations; subsistence practices; and meaningful involvement or
engagement1
2.1 Disproportionate
For this technical guidance, the term disproportionate is used to refer to differences in effects or risks
that are extensive enough that they may merit Agency action and should include consideration of
cumulative impacts or risks where appropriate and consistent with applicable law (U.S. EPA, 2022a).8
In general, the determination of whether a difference in effects or risks is disproportionate is ultimately a
policy judgment which, while informed by analysis, is the responsibility of the decision-maker.9 The
terms difference or differential indicate an analytically discernible (or measurable) distinction in effects
or risks across population groups. It is the role of the analyst to assess and present differences in
anticipated effects across population groups for both the baseline and regulatory options, using the best
available information (both quantitative and qualitative) to inform the decision-maker and the public.
See Text Box 2.1 for examples of how differences in effects have been characterized for a regulatory
action.
2.2 Population Groups of Concern
At an early stage of the analysis, analysts need to identify the population groups of concern relevant to
a specific regulatory context.10 The concept of vulnerability can be used to help identify population
5 The term EJ concern was first defined in U.S. EPA (2015a) and used in the 2016 version of this technical guidance. While
consistent with those earlier definitions, it has been revised slightly here to reflect E.O. 14096.
6 What constitutes an adverse effect will likely vary by regulatory context and statute.
7 While the Justice40 Initiative uses the term disadvantaged, it is not typically used at the EPA for purposes of identifying
and/or evaluating populations or communities with EJ concerns in the context of rulemaking. For information on the Justice40
Initiative, see: https://www.whitehouse.qov/wp-content/uploads/2021/07/M-21-28.pdf.
8 The definition of disproportionate is consistent with U.S. EPA (2022a) and builds on what was used in the 2016 version of
this technical guidance.
9 A finding of disproportionate and adverse effects is neither necessary nor sufficient for the EPA to address environmental
risks. The Agency's statutory and regulatory authorities provide a broad basis for protecting human health and the
environment and do not require a demonstration of disproportionate effects to protect the health or environment of any
population.
10 The term population groups of concern is used instead of the term subpopuiations to include "population groups that form a
relatively fixed portion of the population (e.g., based on ethnicity)." See the EPA's Early Lifestages website:
http://www.epa.qov/children/earlv-life-staqes.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
4
-------
groups of concern.11 For example, analysts can combine available data on baseline health,
demographic, socioeconomic, or other relevant indicators (including those related to cumulative
impacts, historic inequities and systemic barriers, and lack of access) to identify characteristics in
affected communities that correlate with increased vulnerability to environmental exposure or lack of
opportunity for public participation (Fann et al., 2011). See Chapter 4 for a discussion of underlying
contributors to increased vulnerability. Further discussion about considering vulnerability and exposure
factors in risk assessment is found in Section 5.3 and Appendix B.
While the EPA does not have rigid criteria for identifying population groups of concern, this section
defines and describes characteristics from E.O.s 12898 and 14096: race, ethnicity, national origin, low-
income, disability status, Tribal affiliated and Indigenous populations, and those engaged in cultural or
subsistence practices. Note that population groups of concern may be clustered within specific
communities or geographically dispersed (e.g., unhoused populations, migrant workers). Underserved
communities or populations also may warrant consideration.12
It may be useful in some contexts to analyze population groups in combination or to evaluate additional
aspects of diversity within a specific population group of concern (e.g., by lifestage, gender), particularly
when some individuals within a population group may be at greater risk for experiencing
disproportionate and adverse effects (e.g., due to unique exposure pathways).
In addition to the information below, analysts should rely on the Office of Management and Budget
(OMB) or other federal statistical agencies (e.g., U.S. Census Bureau), when available, to define
relevant population groups (or combinations thereof) for a specific regulatory action. Note that analysis
of additional population groups is not a substitute for examining the population groups explicitly
mentioned in the E.O.s.
11 Note that specific terminology and definitions related to the term vulnerability may be provided by statute.
12 Examples of other characteristics that may be relevant in some regulatory contexts include linguistic isolation, occupation,
rurality, and employment status, among others.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
5
-------
Text Box 2.1 Characterizing Differences in Effects for a Regulation
Regulatory actions have described differences in the size, type, or distribution of environmental and
health effects among population groups or communities, both in the baseline and for the regulatory
options. Terminology varies with specific context. Below are three examples.
The final rule for the Phasedown of Hydrofluorocarbons (HFC) (U.S. EPA, 2021a) states:
"EPA finds evidence of environmental justice concerns near HFC production facilities from
cumulative exposure to existing environmental hazards in these communities. However, given
uncertainties about where and in what quantities HFC substitutes will be produced, EPA cannot
determine the extent to which this rule will exacerbate or reduce existing disproportionate adverse
effects on communities of color and low-income people."
The final rule for the National Emission Standards for Hazardous Air Pollutants: Ethylene Oxide
Emission Standards for Sterilization Facilities Residual Risk and Technology Review (U.S. EPA,
2024a) describes its baseline demographic analysis as follows:
"A total of 17.3 million people live within 10 km of the 88 facilities that were assessed. The percent of
the population that is Hispanic or Latino is substantially higher than the national average (36 percent
versus 19 percent), driven by the seven facilities in Puerto Rico, where an average of 99 percent of
the 658,000 people living within 10 km of the facilities are Hispanic or Latino. The proportion of other
demographic groups living within 10 km of commercial sterilizers is similar to the national average.
The EPA also conducted a risk assessment of possible cancer risks and other adverse health
effects, and found that prior to the implementation of this regulation, cancer risks are unacceptable
for several communities."
The final Per-and Polyfluoroalkyl Substances (PFAS) National Primary Drinking Water Standard rule
(U.S. EPA, 2024b) states that the EPA's EJ analysis:
"...demonstrates that some communities of color are anticipated to experience elevated baseline
PFAS drinking water exposures compared to the entire sample population... The EPA believes that
this action is likely to reduce existing disproportionate and adverse effects on communities with EJ
concerns. Across all hypothetical regulatory thresholds, elevated exposure—and thus reductions in
exposure under the hypothetical regulatory scenarios—is anticipated to occur in communities of
color and/or low-income populations."
2.2.1 Race, Ethnicity, Tribal Affiliated and Indigenous Populations, and National Origin13
The OMB provides minimum reporting standards for "maintaining, collecting, and presenting race and
ethnicity data for all Federal information collection and reporting purposes. The categories in these
standards are understood to be socio-political constructs and are not an attempt to define race and
13 Where data shows a higher concentration of health and environmental effects among specific population groups—such as
among groups of a particular race, national origin or sex—the EPA uses race-neutral criteria to address such actual or
potential risks/harms and to ensure environmental justice for all. The EPA does not distribute governmental benefits or
burdens based on race, color, national origin, or sex; however, such data are important for understanding the distribution of
environmental and health effects and risks, as well as to measure the impact of race-neutral EPA decision-making on these
distributions.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
6
-------
ethnicity biologically or genetically. These standards... provide a common language for uniformity and
comparability in the collection and use of race and ethnicity data by Federal agencies" (OMB, 2024).
Within this context, the OMB defines seven minimum categories for data on race and ethnicity:
• American Indian or Alaska Native;
• Asian;
• Black or African American;
• Hispanic or Latino;
• Middle Eastern or North African;
• Native Hawaiian or Pacific Islander; and
• White.
Note that these categories are not mutually exclusive and cannot simply be added to estimate a total
population. For example, Hispanic or Latino is an ethnic category and, as such, an individual may
identify as both Hispanic or Latino and as one or more races.
While the OMB does not use the terms Tribal affiliation or Indigenous, it defines someone who identifies
as an American Indian or Alaska Native as an individual "with origins in any of the original peoples of
North, Central, and South America" (OMB, 2024). The EPA defines Indigenous Peoples in the EPA
Policy on Environmental Justice for Working with Federally Recognized Tribes and Indigenous Peoples
(U.S. EPA, 2014a) to include state-recognized Tribes; Indigenous and Tribal community-based
organizations; individual members of federally recognized Tribes, including those living on a different
reservation or living outside Indian country; individual members of state-recognized Tribes; Native
Hawaiians; Native Pacific Islanders; and individual Native Americans.
National origin refers to where a person or their family is from originally and may encompass birthplace,
ethnicity, ancestry, culture, and language. In addition, national origin may refer to a specific country or
to a part of the world. While potentially inclusive of national origin, the race and ethnicity categories in
the U.S. Census do not distinguish individuals based on national origin (U.S. Census Bureau, 2024).
The U.S. Census asks about an individual's ancestry, which is defined as "a person's ethnic origin or
descent, 'roots,' or heritage, or the place of birth of the person or the person's parents or ancestors
before their arrival in the United States." In addition, it may encompass identities that originate in
geographic areas outside the United States or from within the United States (U.S. Census Bureau,
2024). Up to two ancestries are tabulated per respondent. Note that some ancestries may not be
reported to protect confidentiality. In addition, the U.S. Census Bureau collects data on foreign-born
individuals living in the United States, which includes anyone who is not a U.S. citizen at birth (U.S.
Census Bureau, 2024).
2.2.2 Low-Income Populations
The OMB has designated the U.S. Census Bureau's annual poverty measure, produced since 1964, as
the official metric for program planning and analysis by all Executive branch federal agencies in
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
7
-------
Statistical Policy Directive No. 14, though the OMB does not preclude the use of other measures (OMB,
1978). The Council on Environmental Quality (CEQ) also suggests analysts use "annual statistical
poverty thresholds from the Census Bureau's Current Population Reports, Series P-60 on Income and
Poverty" to define low-income populations (CEQ, 1997).14
The U.S. Census Bureau's annual official poverty measure uses a set of income thresholds that vary by
family size and composition to determine the households that live in poverty. If a family's total income
falls below the threshold, then that family and every individual in it is defined as being in poverty. This
measure of poverty has remained essentially unchanged since its inception.15 It does not vary
geographically, though it is updated for inflation using the Consumer Price Index for All Urban
Consumers (CPI-U).16 It also does not account for capital gains or non-cash benefits such as public
housing, Medicaid, and food stamps (U.S. Census Bureau, 2022).
The ability of the official poverty measure to adequately capture regional and other differences in
economic well-being has been the subject of ongoing debate. In particular, the National Research
Council (NRC) recommended that the official measure be revised because "it no longer provides an
accurate picture of the differences in the extent of economic poverty among population groups or
geographic areas of the country, nor an accurate picture of trends over time" (Citro and Michael, 1995).
In response, the OMB convened an interagency group in 2009 to define a supplemental poverty
measure based on the NRC recommendations. The U.S. Census Bureau has produced a
Supplemental Poverty Measure annually since 2011 (Fox and Burns, 2021). Unlike the official poverty
measure, it accounts for "co-resident unrelated children" (such as foster children), any cohabiters and
their children, and uses a broader resource measure to account for out-of-pocket medical expenses
and in-kind benefits. It also improves on the traditional measure of poverty by adjusting for differences
in housing prices and family size by metropolitan statistical area.17 However, since the Supplemental
Poverty Measure is based on survey data and available only at a relatively aggregate geographic
spatial scale, it should not supplant the use of official poverty measures in an analysis of EJ concerns.18
Unlike its treatment of poverty, the U.S. Census Bureau does not provide an official definition of low
income. For screening purposes, the EPA's EJScreen tool uses the "number or percent of a block
group's population in households where the household income is less than or equal to twice the federal
poverty level" to identify low-income populations (U.S. EPA, 2024c). However, analysts may
characterize low-income populations more broadly than just those that fall below a certain income level
14 Federal agencies may also identify low-income populations relevant to their mission. For example, the Federal Reserve
Board, under the Community Reinvestment Act, designates a low-income community as one where median family income is
less than 50 percent of the area median income. The U.S. Department of Agriculture defines as limited resource a farmer or
rancher "with direct or indirect gross farm sales not more than the current indexed value in each of the previous two years, and
who has a total household income at or below the national poverty level for a family of four, or less than 50 percent of county
median household income in each of the previous two years." See
https://www.federalreserve.gov/consumerscommunities/cra resources.htm and https://lrftool.sc.eqov.usda.gov/.
15 The U.S. Census Bureau produces single-year estimates of median household income and poverty by state and county and
poverty by school district in its Small Area Income and Poverty Estimates. It also provides estimates of health insurance
coverage by state and county in its Small Area Health Insurance Estimates. These data are broken down by race at the state
level and by income categories at the county level.
16 See https://www.bls.gov/cpi/auestions-and-answers.htm for a discussion of the limitations of the Consumer Price Index.
17 The NRC recognizes that income-based measures such as the official or supplemental poverty thresholds are not
necessarily the best measure of relative poverty since they do not account for differences in accumulated assets across
households. The Supplemental Poverty Measure tries to capture inflows of income and outflows of expenses, which are likely
correlated with short-term poverty since many assets are not easily convertible to cash in the short run (Short, 2012).
18 See Congressional Research Service (2022) for more information on the Supplemental Poverty Measure, including
limitations and outstanding issues.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
8
-------
(e.g., include families with income above the poverty threshold but still below the average U.S.
household income). Additional socioeconomic characteristics typically collected by U.S. statistical
agencies, such as educational attainment, baseline health status, and health insurance coverage, may
also be useful for characterizing low-income populations. Measures that capture the dynamics of
poverty, such as the percent of people who are chronically poor versus those who experience poverty
on a more episodic basis, are also available in other U.S. Census Bureau data products (Iceland,
2003).19
Finally, the U.S. Census Bureau makes available several cross-tabulations between poverty measures
and other socioeconomic characteristics of interest such as race, ethnicity, age, sex, education, and
work experience; these can be useful in developing more specific population descriptions.
2.2.3 Disability Status
The Federal government defines an individual with a disability as someone who has a physical or
mental impairment that substantially limits one or more major life activities, has a record of such an
impairment, or is regarded as having such an impairment (Americans with Disabilities Act, 1990).
Beginning in 2008, the U.S. Census Bureau asked respondents of the American Community Survey
(ACS) about six types of disability:20
• Hearing difficulty: deaf or having serious difficulty hearing.
• Vision difficulty: blind or having serious difficulty seeing, even when wearing glasses.
• Cognitive difficulty: having difficulty remembering, concentrating, or making decisions because
of a physical, mental, or emotional problem.
• Ambulatory difficulty: Having serious difficulty walking or climbing stairs.
• Self-care difficulty: Having difficulty bathing or dressing.
• Independent living difficulty: having difficulty doing errands alone such as visiting a doctor's
office or shopping because of a physical, mental, or emotional problem.
Several other agencies also collect statistical information on disability status based on the same six
disability categories listed above. For instance, the Bureau of Labor Statistics collects information on
the employment status of persons with disabilities as part of the Current Population Survey. The
Centers for Disease Control and Prevention includes disability status in several population surveys
about a wide range of demographic, socioeconomic and health indicators.21
19 This type of measure is reported in the U.S. Census Bureau's Survey of Income and Program Participation. For more
information, see https://www.census.gov/proarams-survevs/sipp.html.
20 In 2013, the U.S. Census Bureau produced the first set of five-year estimates on disability status for all geographies
including tracts and block groups. See https://www.census.qov/topics/health/disabilitv/quidance/data-collection-acs.html for
more information. Note that these data exclude people in institutions such as nursing or retirement homes, correctional
facilities, and inpatient hospice care.
21 For more information, see https://www.cdc.gov/ncbddd/disabilitvandhealth/datasets.html.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
9
-------
2.2.4 Populations that Rely on Cultural and Subsistence Practices
E.O. 12898 identifies the need to analyze the human health effects and risks of Agency programs,
policies, and activities for "populations with differential patterns of subsistence consumption of fish and
wildlife ... whenever practical and appropriate." E.O. 14096 also highlights the importance of analyzing
differences in consumption patterns related to the cultural and subsistence practices of Tribal and
Indigenous populations. Tribal and Indigenous populations often rely on traditional diets of indigenous
fish, vegetation, and/or wildlife. Access to medicinal plants and natural resources is also often integral
to Tribal cultural practices, traditions, and customs (also referred to as lifeways) and helps define them
as peoples (U.S. EPA, 2019a).
The CEQ (1997) describes the two main components of subsistence practices: differential patterns of
consumption of natural resources and subsistence consumption. Differential patterns of consumption of
natural resources are "differences in rates and/or patterns of fish, water, vegetation and/or wildlife
consumption among minority populations, low-income populations, or Indian Tribes, as compared to
rates and patterns of consumption of the general population." The term subsistence consumption is
defined as dependence "on indigenous fish, vegetation and/or wildlife, as the principal portion of their
diet." Both differential patterns and subsistence consumption can result in higher exposure to
environmental stressors. For example, Tribal and Indigenous populations may be exposed to higher
concentrations of environmental contaminants due to greater reliance on indigenous fish, vegetation
and/or wildlife in their diets and cultural practices compared to the general population. See Section
4.1.2 for a discussion of unique exposure pathways.22 Note that this category identifies populations
based on specific pathways of exposure and may overlap with those defined based on income,
race/ethnicity, and national origin.23
While federal statistical agencies do not specifically track the cultural and subsistence practices of
individuals and population groups, the EPA has conducted consumption surveys and exposure
assessments in specific geographic areas to inform policy formulation (see U.S. EPA (2011a) for
examples). If differential patterns of indigenous fish, vegetation, and wildlife consumption are a
substantial concern for a specific regulatory action, analysts should refer to existing EPA guidance
when collecting and using these data for analysis (e.g., U.S. EPA, 2019a; 2016a; 2011a).
Analysts may also investigate whether data collected via consumption surveys or exposure
assessments are available from other federal agencies or from state, Tribal, or local governments. Per
EPA guidance, it is important to verify that any survey data used in an EJ analysis accords with
appropriate parameters and methodology for that specific analysis (U.S. EPA, 2016a). Note, it is
essential to gain permission from a Tribe to gather and use information on cultural practices (e.g.,
Indigenous Knowledge, also referred to as Traditional Ecological Knowledge).24 Because Tribes retain
22 For example, over 40% of non-Hispanic Asian populations in the United States eat seafood at least twice per week
compared to a national average of 20% between 2013 and 2016 (Terry et al., 2018). This can result in elevated mercury levels
that can affect neurodevelopment in children and the risk of cardiovascular disease in adults (e.g., Buchanan et al., 2015).
23 The overlap between populations that principally subsist on indigenous fish, vegetation, and wildlife and other population
groups based on race, ethnicity, income, or other factors is an important consideration when evaluating EJ concerns in a risk
assessment. As part of a risk assessment, analysts are encouraged to evaluate as appropriate all consumption/contact
patterns and rates that are relevant from an EJ perspective, including those associated with populations that subsist on
indigenous fish, vegetation, and wildlife.
24 While Indigenous Knowledge is used here, a variety of terms, including Traditional Ecological Knowledge, Traditional
Knowledge, Indigenous Traditional Knowledge, Native science, and related terms are used and preferred by different Tribes
and Indigenous Peoples. For more information, see OSTP and CEQ (2022).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
10
-------
ownership of these data and can decide who and how they are used, it is important that the EPA be
transparent about statutory requirements that may limit its ability to protect sensitive information from
public disclosure.25 For these reasons, data collection, storage, and sharing should be discussed with
Tribes early in the process (see Section 2.3).26
2.3 Meaningful Involvement or Engagement
The EPA strives to meaningfully involve or engage members of the public that are potentially affected
by Federal activities by:
• Providing timely opportunities for members of the public to share information and concerns and
participate in decision-making processes,
• Considering public input provided as part of decision-making processes,
• Seeking out and encouraging the involvement of persons and communities potentially affected
by federal activities; ... [and]
• Providing technical assistance, tools, and resources to help facilitate meaningful and informed
public participation, whenever practical and appropriate (U.S. EPA, 2024d).
The EPA can seek out and encourage public involvement by sharing information in a way that is
accessible to individuals with limited English proficiency and people with disabilities; providing notice of
and reaching out to communities or groups of people who are potentially affected and who might not
regularly participate in federal decision-making; and addressing other barriers to participation that
individuals may face, to the extent practicable and appropriate. (U.S. EPA, 2024d).27 In addition, the
EPA continues to respect Tribal sovereignty and support self-governance by ensuring that Tribal
Nations are consulted on Federal actions that have Tribal implications. The EPA consults with federally
recognized Tribes consistent with the EPA Policy on Consultation with Indian Tribes (U.S. EPA, 2023f)
and E.O. 13175.28
While opportunities for engaging the public differ across agency activities and take time and resources,
there are three steps recommended for planning and providing opportunities to meaningfully engage
the public that can be tailored to the decision context: (1) understand the EPA action and key issues the
public can inform; (2) identify the expected level of participation; and (3) identify the appropriate
engagement tools and practices. Details on how to actualize these steps, as well as best practices,
25 Statutory requirements include the Freedom of Information Act and Administrative Procedures Act.
26 See the EPA's Policy on Environmental Justice for Working with Federally-Recognized Tribes and Indigenous Peoples
(https://www.epa.aov/environmentaliustice/epa-policv-environmental-iustice-workina-federallv-recoanized-tribes-and). For
more information on practices around Tribal data sovereignty, see Carroll et al. (2020).
27 The terms meaningful involvement and meaningful engagement are often used interchangeably. For the remainder of this
document, we use the term meaningful engagement to align with U.S. EPA (2024d).
28 The EPA's policy is to consult on a government-to-government basis with federally recognized Tribes when EPA actions or
decisions may affect Tribes. Tribal consultations are separate and distinct from the EPA's obligations to engage the public as
required by federal environmental laws and regulations and the EPA's work to address the environmental concerns of non-
federally recognized Tribes, individual Tribal members, Tribal community-based organizations, and other Indigenous
stakeholders. The EPA Policy on Consultation with Indian Tribes (U.S. EPA, 2023f) is available at:
https://www.epa.aov/svstem/files/documents/2023-12/epa-policv-on-consultation-with-indian-tribes-2023.pdf.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
11
-------
examples, and other resources are available in the EPA's Meaningful Engagement Policy (U.S. EPA,
2024d).29
While the EPA's Meaningful Engagement Policy does not create new legal requirements or mandatory
obligations, the recommendations are designed to enhance meaningful engagement opportunities
within Agency actions, including rulemaking. The EPA is committed to providing timely, accessible, and
accurate information to the public about its regulatory actions and recognizes the importance of
providing "opportunities for participation that foster a spirit of mutual trust, confidence, and openness
between the EPA and the public" (U.S. EPA, 2024d). Within the regulatory context, the OMB (2023a)
recommends that agencies engage communities through trust-based, long-term, and two-way
relationships throughout the regulatory process.30
Meaningful engagement intersects with analytic considerations in several important respects. First,
providing the analysis of EJ concerns in a timely fashion and using plain language to improve
accessibility makes key assumptions, methods, and results more transparent and easier to
understand.31 This can further a clear understanding of the EJ implications of a regulatory action and
allow for more substantive engagement by community members and other interested parties during the
rulemaking process. Second, analysts play a role in ensuring meaningful engagement by evaluating
possible differences in opportunities for ongoing public input and feedback, the ability to identify and
resolve compliance issues, and ways implementation may be improved once a regulation is in place for
the regulatory options under consideration. Third, it may be possible for analysts to request data and
other information early in the process regarding unique exposure pathways or end points of concern
that could improve the analysis of EJ concerns. This includes the possible use and consideration of
Indigenous Knowledge (also referred to as Traditional Ecological Knowledge) and data and information
collected via participatory science (U.S. EPA, 2024d).32 Communities with EJ concerns have unique
knowledge of their goals, needs, and vulnerabilities.
Through early involvement, the EPA can obtain information and improve understanding of issues
affecting these populations in the context of the regulatory action.33 Text Box 2.2 highlights several
examples of activities taken to ensure meaningful engagement on EJ issues for regulatory actions.
Section 5.3.1.2 discusses meaningful engagement in the context of a human health risk assessment.
29 The EPA's National Environmental Justice Advisory Council (NEJAC) issued updated recommendations on public
participation in 2013, Model Guidelines for Public Participation, available at https://www.epa.qov/environmentaliustice/model-
quidelines-public-participation. See also U.S. EPA (2015b).
30 EPA staff should also hold early, transparent discussions about Freedom of Information Act with public participants prior to
seeking public input, exchanging information, obtaining recommendations, entering into collaborations or agreements,
conducting community-based participatory research, and working together. Sharing the EPA's limitations with participants
enables communities to assess how they intend to share information or knowledge with the Agency prior to engagement as
well as facilitates trust and relationship building.
31 The Plain Writing Act of 2010 (Public Law 111-274) defines plain language as "writing that is clear, concise, well-organized,
and follows other best practices appropriate to the subject or field and intended audience." For Federal plain language
guidelines, see https://www.plainlanquaqe.gov/quidelines/.
32 The Agency's network of Tribal Partnership Groups facilitates the exchange of technical information and communication
between Tribes and the EPA. The National EPA-Tribal Science Council works to integrate and increase tribal involvement in
the EPA's scientific activities, while the National Tribal Toxics Council provides Tribal input on issues related to toxic chemicals
and pollution prevention. Assessors should engage with these partnership groups, through the EPA's Tribal Program, to better
understand Tribal lifeways and to discuss and collaborate on data and research needs.
33 Note that the Paperwork Reduction Act requires that an Information Collection Request be submitted for collecting
information (e.g., focus groups, interviews, surveys) from more than nine people (44 U.S.C. 3501).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
12
-------
Text Box 2.2 Examples of Meaningful Engagement for EPA Regulatory Actions
To inform the proposed Supplemental Effluent Limitations Guidelines and Standards for the Steam
Electric Power Generating Point Source Category, the EPA characterized the environmental and
socioeconomic characteristics of communities that are expected to be impacted by discharges from
steam electric plants to prioritize outreach and engagement. The EPA then met with community members
in five identified communities either virtually or in a hybrid format.
"Two broad themes were conveyed consistently across communities. First, community members
conveyed several perceived harmful impacts from steam electric power plants and their desire for more
stringent regulations to reduce these harmful impacts. Second, community members expressed the
desire for more transparency and communication to overcome their decreasing trust in the regulated
power plants and state regulatory agencies and, thus, a corresponding skepticism that their community
would be protected from these harmful impacts. In addition to these broad themes, commenters also
raised concerns unique to each community. For example, members of the Navajo Nation discussed with
EPA the spiritual and cultural impacts to the community from pollution related to steam electric power
plants. In Jacksonville, Florida, community members raised concerns regarding tidal flows of pollution
upstream and storm surges during extreme weather events which cause additional challenges in their
community" (U.S. EPA, 2023c).
In addition to imposing more stringent requirements to limit discharges from steam electric power plants,
the final rule responds to the feedback received at these outreach meetings by:
"...requiring posting of required reports to a publicly available website to improve transparency. In
addition, the EPA recently added a new feature called ECHO Notify to the Enforcement and Compliance
History Online (ECHO) website. ECHO Notify provides weekly email notifications of changes to
enforcement and compliance data in ECHO. Notifications are tailored to the geographic locations, facility
IDs, and notification options that users select" (U.S. EPA, 2024e).
For the Petroleum Refineries Risk and Technology Review, the EPA conducted EJ-relevant outreach
activities with communities living near refineries in the two years prior to proposing the rule. It was in the
context of this outreach that the possibility of fenceline monitoring was first raised as a regulatory option.
In response to public comments asking for more data on refinery emissions that may affect nearby
communities, the EPA changed several aspects of its proposed electronic reporting requirements: it
required data collection no later than two years from the effective date of the final rule instead of the
originally proposed three years and required submission of fenceline monitoring data on a quarterly
instead of a semi-annual basis.
In addition, the EPA stated that it "will continue to work with communities to better understand their
unique concerns and needs. We will seek opportunities to enhance education and engagement around
our rules, including the best way to make the monitored data required by this rule accessible and
digestible by those who need to understand what the data means" (U.S. EPA, 2015c). Subsequent public
outreach and solicitation of input specifically asked for feedback from the public:
• "What information needs to be included on the web page to understand the data?
• Within context of the rule, what information and data on benzene do you want to see on the web
page?
• Do you have ideas for other ways to share the information besides a web page?
• What other information would be helpful for you? Other EPA information about air toxics? Links to
EPA environmental justice tools?
• Is there other training or support that might be helpful?" (U.S. EPA, 2016b).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
13
-------
Chapter 3: Key Analytic Considerations
This chapter provides an overview of the questions analysts should aim to evaluate and the main steps
of an analysis of EJ concerns. It also offers five broad recommendations and a list of best practices to
enhance consistency across assessments.
3.1 Questions to Evaluate When Analyzing EJ Concerns
The analysis of EJ concerns for regulatory actions should address three questions:34
• Baseline: Are there existing EJ concerns associated with environmental stressors affected by
the regulatory action for population groups of concern?
• Regulatory options: For the regulatory option(s) under consideration, are there potential EJ
concerns associated with environmental stressors that are affected by the regulatory action for
population groups of concern?
• Mitigation or exacerbation of effects: For the regulatory option(s) under consideration, are EJ
concerns exacerbated, mitigated, or unchanged compared to the baseline?
The term environmental stressor encompasses the range of chemical, physical, or biological agents,
contaminants, or pollutants that may be subject to a regulatory action. Baseline is defined as "an
analytically reasonable forecast of the way the world would look absent the regulatory action being
assessed, including any expected changes to current conditions over time" (OMB, 2023b).35 Section
6.2 of this document provides more information on characterizing the baseline for a regulatory action.
Note that characterizing the baseline may include consideration of other relevant environmental or non-
environmental stressors that increase a community or population group's vulnerability. This may include
considering the cumulative effects of exposure to multiple stressors on human health and well-being
and those related to climate change. See Chapter 4 for a discussion of vulnerability.
3.2 Identifying Objectives, Data, and Other Information
The purpose of a regulatory analysis is to "anticipate and evaluate the likely consequences" of a
regulatory action in a way that informs the public and decision-makers (OMB, 2023b).36 To help fulfill
34 These analytic questions are intended to prompt assessment of differences in anticipated effects across population groups
for the baseline and regulatory options, and to prompt presentation of these results to decision-makers to support
determinations of potentially actionable disproportionate and adverse effects. Differences in effects or risks may include
differential exposures, differential health and environmental outcomes, or other relevant effects.
35 "In some cases, it may be reasonable to forecast that the world absent the regulation will resemble the present. In other
cases, particular attention should be paid to ways in which conditions will change absent the regulation—e.g., technological
advances, demographic changes, changes in the economy, or alterations to the climate—that will significantly affect the
estimated effects of the regulation... If a harm addressed by a regulation is expected to become more severe overtime, the
baseline should reflect that trend" (OMB, 2023b).
36 E.O. 12866 (1993) directs agencies to consider distributional effects when choosing among alternative regulatory
approaches, unless prohibited by statute. A distributional analysis is one that characterizes "how the benefits and the costs of
a regulatory action are ultimately experienced across the population and economy, divided up in various ways (e.g., across
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
14
-------
this purpose, analysts need to communicate with decision-makers early in the process on how the
relevant Executive Orders and other applicable EPA policies or statutes interact with the evaluation of
EJ concerns for a regulatory action. Identifying the objectives of the EJ analysis is a key input into the
analytic plan, including how to frame the questions, which datasets to use, which analyses to perform,
and how to synthesize and interpret results.
To help decide what level of analysis of EJ concerns is feasible and appropriate, analysts need to
identify the quantitative and qualitative data and methodological needs to ensure that they are duly
considered and reasonably accommodated. Data and methods availability influence the scope and
complexity of an assessment and may inform the extent to which EJ concerns are considered in the
decision-making process (see Sections 5.3, 6.3, and 6.4). Feasibility is based on a technical evaluation
of the data and methods that are available or can be reasonably collected (e.g., the availability of
disaggregated data, data quality, and the quality of evidence from the peer-reviewed literature,
community input, and other information). Appropriateness is informed by relevant policy, budgetary,
and statutory considerations (see Chapter 6). Text Box 3.1 provides an overview of the criteria by which
the EPA assesses the quality of scientific and technical information used in support of Agency
decisions. When possible, it is recommended that analysts also identify the population groups of
greatest relevance within the context of a specific regulatory action early in the process to inform data
collection and analysis.
In some circumstances, available data may not be sufficient to perform a quantitative evaluation, but it
may be possible to conduct a meaningful qualitative analysis or rely on a mix of quantitative and
qualitative information (see Sections 6.1 and 6.3).37 Documentation of the process for identifying what
level of analysis is feasible is encouraged and ensures transparency when communicating with the
public. It is also recommended that analysts coordinate with the Office of Policy when determining the
level of EJ analysis undertaken for a specific regulatory action.
income groups, race or ethnicity, sex, gender, sexual orientation, disability, occupation, or geography; or relevant categories
for firms, including firm size and industrial sector). The benefits and costs of a regulation may also be distributed unevenly over
time, resulting in regulatory benefits and costs falling on different individuals or different groups of individuals" (OMB, 2023b).
Note that while the concepts and objectives of distributional analysis and EJ analysis overlap, they are distinct. See Wolverton
(2023) for more discussion.
37 The EPA also receives information that is voluntarily submitted by or collected from other federal, state, Tribal, local, and
international agencies; national laboratories; academic and research institutions; business and industry; and public interest
organizations. Although the EPA's existing quality systems are not applied at the time this information is generated, the EPA
"does apply appropriate quality controls when evaluating this information for use in Agency actions and for its dissemination
consistent with the EPA" data quality guidelines (U.S. EPA, 2012a).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
15
-------
Text Box 3.1 The Quality and Relevance of Scientific and Technical Information
To evaluate the quality and relevance of scientific and technical information used to support Agency
actions, the EPA assesses five factors:
Soundness - The extent to which the scientific and technical procedures, measures, methods, or
models employed to generate the information are reasonable for, and consistent with, the intended
application.
Applicability and Utility - The extent to which the information is relevant for the Agency's intended
use.
Clarity and Completeness - The degree of clarity and completeness with which the data,
assumptions, methods, quality assurance, sponsoring organizations, and analyses employed to
generate the information are documented.
Uncertainty and Variability - The extent to which the variability and uncertainty (quantitative and
qualitative) in the information or in the procedures, measures, methods, or models are evaluated
and characterized.
Evaluation and Review - The extent of independent verification, validation, and peer review of the
information or of the procedures, measures, methods, or models.
See U.S. EPA (2012a) for examples of the types of questions that could be considered to assess
each of these factors. OMB (2019) and U.S. EPA (2002b) also are helpful resources.
3.3 Main Steps of an EJ Analysis
Conducting a preliminary analysis may be a useful first step to identifying what level of assessment is
feasible and appropriate to support the regulatory action. In addition, it can help identify the extent to
which a regulatory action may raise EJ concerns that need further evaluation. While there is no single
recommended approach for conducting a preliminary analysis, Section 6.1 discusses the types of
factors and tools analysts may consider as part of a preliminary analysis.
In cases where there is a need for further evaluation, and the data and methods support it, an in-depth
analysis should be conducted. The extent to which an in-depth analysis can address all three questions
from Section 3.1 will vary due to data limitations, time and resource constraints, and other technical
challenges that vary by media and regulatory context. The EPA encourages analysts to document key
reasons why a particular question cannot be addressed to help identify future priorities for filling key
data and research gaps. A lack of existing data and methods does not mean there is no EJ concern, so
identifying such gaps is important.
While Offices may develop program-specific approaches to evaluating EJ concerns for regulatory
actions to reflect their specific methodological, regulatory, and resource constraints, an EJ analysis
typically includes five main steps (Figure 3.1). We briefly describe each step and refer the reader to
more detailed discussion in Chapter 6.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
16
-------
Figure 3.1 Main Steps of an EJ Analysis
1. Identify the sources being regulated: Before analysts can identify the populations and
communities being affected by a regulatory action, it is important to first characterize the regulated
sources: where are they located? Are there particular characteristics of the regulated sources that
contribute to higher exposure and/or risk of health effects? Do the regulatory options vary with
these characteristics? For instance, are some sources subject to greater stringency or other
regulatory requirements that would be important to account for in the EJ analysis?
2. Describe the environmental stressor: The spatial distribution of health and welfare outcomes is
a relevant consideration for some regulatory actions. In these cases, evidence on the fate and
transport of the environmental stressor can help determine the populations and communities
potentially exposed. In other cases, the regulatory action's effects may be more widespread (see
Section 6.5.3). It is also important to understand which specific health effects are of greatest
relevance for a given regulatory context. The benefits analysis and, when conducted, the human
health risk and exposure assessments can be important sources for this information (see Chapter
5 and Appendix B).
3. Characterize affected populations: It is important to understand what factors may contribute to
EJ concerns. How are individuals being exposed? Are there unique pathways or other factors that
drive higher exposures for some population groups? Recognizing underlying contributors within a
specific regulatory context is important for properly assessing EJ concerns and can aid in the
design of regulatory options. This may include evidence of already overburdened communities,
including the cumulative effects of exposure to multiple environmental or non-environmental
stressors on human health and well-being (see Chapter 4, Chapter 5, and Appendix B).
4. Compare the affected and comparison groups: To answer each of the three questions in
Section 3.1, analysts need to characterize the exposure and risk of health effects for population
groups of concern in the baseline and for the regulated action relative to a comparison population
group (see Section 6.5.2). This allows analysts to gauge the extent to which effects for the
affected population are similar or different than they are for the comparison group and how they
vary across population groups.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
17
-------
5. Conduct sensitivity analysis: Due to the inherent limitations and uncertainties associated with
analyses of EJ concerns, conducting sensitivity analysis around key assumptions is particularly
important for clearly communicating results to the public (see Chapter 6).
In addition, Figure 3.1 identifies four overarching questions that are relevant throughout the EJ analytic
process:
How can meaningful engagement inform the EJ analysis?
As discussed in Section 2.3, meaningful engagement can help analysts to identify and sometimes help
fill information and data needs. It can also help analysts to identify factors such as unique pathways or
pre-existing vulnerabilities that may contribute to exposure and/or risk for affected populations.
What data are available and at what spatial scale?
As mentioned in Section 3.2, the quality and availability of data are key determinants in the scope and
complexity of the EJ analysis. In some cases, analysts will have data at the individual level for the
environmental stressor being regulated, allowing for a detailed, rigorous analysis. In other cases,
analysts may need to rely on proxies for individual-level effects. Data relevant to the EJ analysis may
include, but are not limited to, the demographic and socioeconomic characteristics of populations that
may be exposed to environmental stressors from regulated sources, what each regulated source is
emitting or discharging, and pre-existing health conditions or other environmental and non-
environmental stressors that increase the vulnerability and therefore the risk of experiencing a health
effect for some population groups (see Sections 5.3.3.2 and 6.3).
What tools are available to model exposure and/or risk?
Analysts have a choice among several scientifically defensible methods to assess EJ concerns
associated with a regulatory action (see Section 5.3.3.2, Section 6.4, and Appendix B), including
proximity-based analysis, exposure and risk modeling, and combining qualitative and quantitative
approaches. The choice of a specific analytic method for the EJ analysis is often driven by data
availability. Together, the data and methods utilized directly influence what conclusions can be drawn
regarding EJ concerns for specific population groups or communities.
How are the effects distributed across sources and communities?
In some cases, extensive differences in effects among population groups of concern may occur in only
a few geographic locations. Referred to as hot spots, these locations are typically exposed to localized
concentrations of emissions from one or more sources along with other stressors. In these cases, it
may be appropriate to tailor the analysis to evaluate effects in a few specific areas. Identifying the
potential for hot spots early helps analysts develop appropriate sources of data and analytic
approaches, which may differ from those used for a broader analysis (see Section 6.5.5).
3.4 Recommendations for Analyses of EJ Concerns
This technical guidance makes five overarching recommendations to ensure a high-quality EJ analysis,
while also recognizing the need for flexibility to reflect policy considerations and technical challenges
within a particular regulatory context. The recommendations are intended to bring greater consistency
across EJ analyses as they strive to answer the three analytic questions from Section 3.1 but are not
prescriptive and do not mandate the use of a specific approach. Analysts should use their best
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
18
-------
professional judgement to decide on the type of analysis that is feasible and appropriate within a
specific regulatory context.
While these recommendations and best practices are intended as a starting point, they should not be
interpreted as limiting the scope of the EJ analysis. It is recommended that analysts thoughtfully tailor
their analysis to the rule context and incorporate new data and methods as they become available.
Ultimately, the EPA strives to innovate and improve upon EJ analyses as the state of science continues
to evolve. The five overarching recommendations are:
1. When risks, exposures, outcomes, or benefits of the regulatory action are quantified, some level
of quantitative EJ analysis is recommended (see Chapter 6).
• Analysts should present information on estimated health and environmental risks,
exposures, outcomes, benefits, or other relevant effects disaggregated by race, ethnicity,
income, and other relevant demographic and socioeconomic categories when feasible and
appropriate.
• When such data are not available, it may still be possible to evaluate potential risk or
exposure using other metrics (e.g., proximity to affected facilities, cancer or asthma
prevalence, or evidence of unique exposure pathways for specific population groups) in a
scientifically defensible way.
• When health and environmental outcomes or benefits are not quantified or disaggregated
by race, ethnicity, income, or other relevant demographic and socioeconomic categories,
analysts should present available quantitative and/or qualitative information that sheds light
on EJ concerns that may arise.
2. Analysts should integrate EJ into the planning of a risk assessment conducted for the regulatory
action (see Chapter 5).
3. Analysts should strive to characterize the distribution of risks, exposures, or outcomes within each
population group, not just average effects (see Section 6.5.5).
• In particular, analysts should pay attention to whether populations in the upper tail of the
distribution face the highest adverse risks, exposures or health effects.
4. Analysts should follow best practices appropriate to the analytic questions at hand.
• Text Box 3.2 outlines best practices for evaluating EJ concerns. If it is not feasible for
analysts to follow these best practices, analysts should explain their use of different
approaches.
5. As relevant, analysts should consider any economic costs or challenges that may be exacerbated
by the regulatory action for relevant population groups of concern.
• For instance, it may be appropriate to consider how low-income populations are affected
by price changes or to consider the distribution of economic costs (i.e., private and social
costs) more broadly from an EJ perspective (see Section 6.7.1).38
38 See the EPA's Guidelines for Preparing Economic Analyses, hereafter referred to as the Economic Guidelines (U.S. EPA,
2024f), for information on defining costs.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
19
-------
Text Box 3.2 Current Best Practices for Evaluating EJ Concerns
• Use the best available science while relying on current, generally accepted Agency procedures for
conducting risk assessment and economic analysis.
• Use existing frameworks and data from other parts of the regulatory analysis, supplemented as
appropriate.
• Be consistent with the basic assumptions underlying other parts of the regulatory analysis, such as
using the same baseline and regulatory option scenarios.
• Use the highest quality and most relevant data available. Discuss the overall quality and main
limitations of the data (see Text Box 3.1 and Section 6.3).
• Identify relevant population groups of concern and discuss available evidence of factors that make
them vulnerable to adverse effects (e.g., unique pathways; cumulative exposure to multiple
stressors; behavioral, biological, or environmental factors) (see Chapter 4).
• Consider unique pathways for individuals that rely on cultural or subsistence practices and
relevance for Tribal or Indigenous populations, when practicable (see Section 2.2.4).
• Carefully select and justify the choice of comparison population group (see Section 6.5.2).
• Carefully select and justify the choice of the geographic unit of analysis and discuss any challenges
or aggregation issues related to the choice of spatial scale (see Section 6.5.3).
• Analyze and compare effects in baseline and across policy scenarios to show differences in effects
(see Section 6.2).
• Present summary metrics for each population group and the comparison population group and
characterize differences between them (see Section 6.6.1).
• When data allow, characterize the distribution of risks, exposures, or outcomes within each
population group, not just average effects (see Section 6.5.5).
• Disaggregate data to reveal important spatial differences (e.g., demographic information for each
source/place) when feasible and appropriate (see Sections 6.5.4 and 6.5.5).
• Clearly describe data sources, assumptions, analytic approaches, and results (see Sections 6.4,
6.5, and 6.6).
• Summarize the main conclusions of differences in exposure or health risk between analyzed
population groups based on the available evidence (see Section 6.6).
• Discuss key sources of uncertainty or potential data biases (e.g., sample size, proximity as a
surrogate for exposure) and how they may influence the results (see Section 6.6.3).
• When possible, conduct sensitivity analysis for key assumptions or parameters that may affect
findings (see Section 6.5).
• Qualitatively describe behavioral responses not accounted for in the analysis that could affect the
level or distribution of exposure or health risks (e.g., dynamic spatial or temporal effects, averting or
adaptive behavior) (see Section 6.5).
• Make elements of the EJ analysis as straightforward and easy for the public to understand as
possible (see Section 2.3).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
20
-------
Chapter 4: Contributors to Environmental
Justice Concerns
Increased vulnerability may be attributable to differences in intrinsic (meaning, biological) factors (e.g.,
age, lifestage, sex, pre-existing disease, genetic predisposition) or extrinsic (meaning, acquired or
external) factors over a person's lifetime. Extrinsic factors consist of both environmental stressors (e.g.,
previous or ongoing exposure to chemicals; the impacts of climate change) and non-environmental
stressors (e.g., socioeconomic status, stress, nutrition, lifestyle, racism, discrimination, culture) from the
built, natural, and social environment across the locations in which individuals spends their time (e.g.,
work, home, school, community).39'40'41
Figure 4,1 Examples of Sources of Human Variability and Vulnerability to Health Risks from
Exposure to Environmental Stressors
(Source: Varshavsky et al., 2023)42
Intrinsic and extrinsic factors interact in complex ways that can result in differential patterns of exposure
to environmental hazards for some population groups, and/or a greater response of some individuals to
a given level of exposure to an environmental stressor (Figure 4.1). These factors can ultimately result
39 It is important to note that race/ethnicity is a social construct that captures the complex interplay of social vulnerability
factors that drive environmental health risk. Belonging to a race/ethnic or low-income group does not on its own influence how
a stressor causes adverse health effects; rather, it is an upstream factor in a causal chain for which there may be little or no
data (Morello-Frosch et al.. 2011).
40 Extrinsic factors may relate to current and historical mechanisms that operate through the labor market, real estate market,
educational system, political institutions, and cultural and societal values to reinforce social hierarchies based on race,
ethnicity, income, occupation, age, or other characteristics (NASEM, 2016; Solar and Irwin, 2010).
41 Differences in outcomes due to extrinsic factors related to socioeconomic, demographic, cultural, psychological, and
physical factors are sometimes also referred to as non-chemical or, more broadly, non-environmental stressors (NASEM,
2023).
42 Used under Creative Commons license (https://creativecommons.Org/licenses/bv/4.0/). No changes were made to the
original figure.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition 21
-------
in higher cumulative risk and incidence of adverse health effects for these population groups and
communities.4344
4.1 Contributors to Higher Exposure to Environmental Hazards
Important extrinsic factors that contribute to higher exposure among population groups of concern
include:
• Proximity to emissions and discharges from nearby sources (U.S. EPA, 2022b; Morello-Frosch
et al., 2011);
• Unique exposure pathways (Burger and Gochfeld, 2011; Solar and Irwin, 2010);
• Physical infrastructure (e.g., housing conditions, water infrastructure) (Solar and Irwin, 2010);
• Exposure to multiple stressors/cumulative exposures (U.S. EPA, 2022c; Morello-Frosch et al.,
2011; Brenderetal., 2011);
• Differential monitoring, compliance, and regulatory enforcement (Banzhaf et al., 2019a); and
• Community capacity to meaningfully participate in decision-making (U.S. EPA, 2011b).
4.1.1 Proximity to Emissions and Discharges from Nearby Sources
It is well documented that sources of environmental hazards are often concentrated in communities
with higher proportions of people of color, low-income populations, Indigenous populations, or persons
with disabilities (Chakraborty et al., 2020; Chakraborty et al., 2016; Mohai et al., 2009).45 Researchers
have pointed to a variety of explanations for these spatial patterns, including the legacy of historically
discriminatory land-use siting decisions (e.g., highway placement, redlining, other zoning practices) and
other systemic barriers (Lane et al., 2022; Shkembi et al., 2022; Grove et al., 2017; Mohai and Saha,
2015).
Not surprisingly, given their proximity to concentrated sources of environmental hazards, these
populations often experience higher levels of exposure.46 Exposure is a function of what or how much is
being emitted or discharged from a nearby source, how and where the pollutant travels as it moves
through the environment (i.e., fate and transport), the time-activity patterns of individuals, as well as
other key determinants (Banzhaf et al., 2019a; NRC, 1991). See Section 6.4.1 for further discussion of
proximity-based analysis.
43 See Cain et al. (2024); WHO (2023); Knapp et al. (2023); Ding et al. (2023); Tessum et al. (2021); Bekkar et al. (2020);
Colmeretal. (2020); Deere and Ferdinand (2020); U.S. EPA (2019a); McHale et al. (2018); Manuck (2017); Akinbami et al.
(2016); Tulve et al. (2016); Wilson et al. (2015); Currie et al. (2013); and Morello-Frosch etal. (2011).
44 The World Health Organization (2006) defines vulnerability as "a matrix of physical, chemical, biological, social, and cultural
factors that result in certain communities and subpopulations being more susceptible to environmental factors because of
greater exposure to such factors or a compromised ability to cope with and/or recover from such exposure."
45 Other studies documenting environmental stressors in communities with higher proportions of non-White populations include
Bullard et al. (2008), Faber and Krieg (2005, 2002), Wilson et al. (2002), and Maantay (2001).
46 See Morello-Frosch and Obasogie (2023), Di Fonzo et al. (2022), Jbaily et al. (2022), Pace et al. (2022), Grineski and
Collins (2018), and Ash and Boyce (2018).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
22
-------
4.1.2 Unique Exposure Pathways
Exposure pathways describe the means by which exposure to a given stressor occurs. Environmental
hazards and risks are not uniformly distributed throughout a population; biological factors and social
context intersect to create unique exposure pathways that put some individuals at higher exposure risk
(Burger and Gochfeld, 2011). At the community level, groups of individuals may be exposed to certain
stressors through shared cultural or social practices, learned traditions, values, and life experiences.
For example, subsistence fishing is more prevalent in some communities, leading to potential exposure
through handling and ingesting fish with high levels of mercury or other chemicals (U.S. EPA, 2023d).
Occupation-related pathways are also relevant to consider, such as potential exposures from pesticide
drift faced by farmworkers or from "take home" chemical residues on clothing (Kalweit et al., 2020;
Hyland and Laribi, 2017).
Exposure pathways are also related to lifestages (U.S. EPA, 2011a). For example, object-to-mouth and
crawling behaviors in infants and toddlers could increase exposure to contaminants such as lead dust
that accumulate on floors or carpets (U.S. EPA, 2013a). Other lifestages, such as persons of
childbearing age or old age, are associated with other exposure pathways; the EPA provides exposure
assessment tools for different lifestages.47
4.1.3 Physical Infrastructure
For some environmental stressors, physical infrastructure (or lack thereof) may contribute to increased
exposure. For instance, housing in the United States built before1978 is more likely to contain lead-
based paint, exposure to which can impair cognitive function in children (U.S. EPA, 2024g). Likewise,
older homes may have leaded pipes and result in exposure via drinking water (Triantafyllidou et al.,
2021). Substandard structural and building conditions such as dampness, poor ventilation, dust
collection, and pest infestation can also trigger asthma or other negative health effects (U.S. EPA,
2021b; Stephens, 2016). Living near highways, airports, railroads, or pipelines may also result in
increased exposure to specific stressors (e.g., particulate matter, lead, noise, odors) (Tessum et al.,
2021; Woodburn, 2017; Rowangould, 2013).
4.1.4 Exposure to Multiple Stressors and Cumulative Exposures
People of color and low-income populations are often impacted by exposure to environmental hazards
from multiple industrial sources, such as contaminants from manufacturing facilities, landfills, and
leaking underground tanks; transportation-related air pollution; and consumer products (e.g.,
Bakkensen et al., 2024; Banzhaf et al., 2019a; California Environmental Protection Agency, 2015).
Rural and Tribal communities may be exposed to higher levels of contaminants that naturally occur in
the soil or groundwater (e.g., radon, uranium, arsenic) via groundwater or economic activities such as
mining or energy production (Erickson et al., 2024). The uneven distribution of the effects of climate
change, such as increased risk of wildfires, droughts, flooding, and other extreme weather events, can
further compound these differences in exposure (Nolte et al., 2018; Lall et al., 2018).
An analysis that considers risks from only one source can inaccurately characterize the potential health
risks faced by a population group if they are also exposed to stressors from other sources. The
presence of other stressors, such as lost access to culturally important subsistence practices, structural
47 See https://wvwv.epa.qov/expobox/exposure-assessment-tools-lifestaqes-and-populations-lifestaqes.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
23
-------
and systemic racism, or other life events, may also exacerbate the effects of some chemical exposures
for vulnerable communities (e.g., increased likelihood of adverse health outcomes due to increased
presence of stress hormones from psychological factors) (Swope et al., 2022; Padula et al., 2020).
4.1.5 Monitoring, Compliance, and Enforcement
The monitoring activities, compliance efforts, and enforcement of existing environmental regulations
can also contribute to differences in exposure. The difficulty and cost of siting and maintaining
monitoring equipment often limits the amount of environmental sampling performed. This can lead to an
underestimate of the emissions generated by regulated sources (Hoyt and Raun, 2015). Given that
industrial activity tends to be clustered in communities with EJ concerns, insufficient information about
what is being emitted by both regulated and unregulated sources can further mask the magnitude of
potential exposure faced by these communities.
Differential compliance across sources can also exacerbate pre-existing disparities (e.g., Fedinick et
al., 2019; Allaire et al., 2018; Balazs et al., 2012). For instance, drinking water systems that serve low-
income, Indigenous, and rural communities and communities of color have been found to have higher
levels of drinking water violations and poorer water quality (Martinez-Morata et al., 2022; Mueller and
Gasteyer, 2021; McDonald and Jones, 2018). The way environmental policies are enforced may also
differ across local governments, states, and Tribal lands (Switzer, 2019; Teodoro et al., 2018).48
4.1.6 Community Capacity to Meaningfully Participate in Decision-Making
The capacity to meaningfully participate in decision-making varies widely across communities and
depends on a variety of factors such as leadership, skills, resources, community power, and social and
organizational networks (Freudenberg et al., 2011). Removing barriers related to disability, language
access, and lack of resources is particularly important for facilitating meaningful engagement in
decision-making. For example, community planning meetings that make facility siting and permitting
decisions without translating key materials and discussion limit the ability of a non-English speaking
community to participate. Other factors that can affect a community's capacity to participate are lack of
broadband internet access, particularly in rural areas and Tribal lands (U.S. FCC, 2020), and lack of
technology fluency. When communities are unable to participate effectively in decision-making due to
these types of barriers, they may be more likely to experience negative environmental consequences.
Though meaningful engagement is related to a community's capacity to participate in the decision-
making process, these topics are not discussed in depth in this guidance document. Information about
how meaningful engagement can inform EJ analysis is found in Sections 2.3 and 5.3.1.2.
48 Enforcing federal environmental regulations is a shared responsibility of federal and state governments and Tribal Nations.
This requires cooperative, periodic, and early joint planning and regular communication between the EPA, states, and Tribal
governments on the sharing of enforcement responsibilities. Several federal environmental laws authorize the EPA to treat
eligible federally recognized Tribes as a state for the purpose of implementing and managing certain federal environmental
programs and functions. That said, the EPA is ultimately "responsible for fair and effective enforcement of federal
requirements. If a state partner is not taking timely or appropriate action to address threats to public health and the
environment, [the] EPA has the authority and responsibility to take direct action" (U.S. EPA, 2023e).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
24
-------
4.2 Contributors to Increased Consequences of Exposure
An individual's biological response to an environmental stressor is an important determinant of both the
occurrence and severity of an adverse health effect. Potentially relevant current and historical intrinsic
and extrinsic factors that may influence this response include, but are not limited to:
• Pre-existing diseases and health conditions (e.g., asthma, disability) (Varshavsky et al., 2023);
• Material circumstances (e.g., neighborhood quality and housing conditions, green space,
walkability, access to fresh foods and high-quality schools; access to credit) (Jimenez et al.,
2021);
• Behavioral and biological factors (e.g., nutrition, smoking, genetic factors) (Varshavsky et al.,
2023);
• Access to health care (e.g., interaction with health care providers and resources; lack of health
insurance or access to preventative care) (Varshavsky et al., 2023);
• Psychosocial circumstances (e.g., stressful living conditions and relationships, low
socioeconomic status, racism, discrimination, lack of coping and support mechanisms) (Padula
et al., 2020; McEwen and Tucker, 2011; Couch and Coles, 2011);
• Limited capacity to adapt to the effects of climate change and other natural disasters (World
Health Organization, 2023); and
• Co-exposure to similarly acting toxics or chemicals, and cumulative burden of disease resulting
from exposure to all stressors throughout the course of life (McPartland et al., 2022; Schwartz et
al., 2011a, 2011b).49
Also known as risk- or effect-modifiers, these factors may influence health outcomes from exposure
through biological interactions at the individual level. Socioeconomic status, which does not by itself
elicit a biological interaction, has a complex and robust association with many health states (Mani et al.,
2013), and may influence factors such as diet, nutrition, and access to health care and consequently
health status (Christensen et al., 2022; Munoz-Pizza et al., 2020; Clougherty et al., 2014).50
Some individuals within population groups may be more vulnerable to the effects of some stressors due
to their stage of physiological and behavioral growth and development, referred to as lifestage (U.S.
EPA, 2011a). Vulnerable individuals based on lifestage can include children, the elderly, and pregnant
women. Workers in certain occupations (e.g., farmworkers) may also be more vulnerable depending on
the health endpoint and stressor. These groups may also have unique exposure pathways or may be
exposed to multiple exposure sources (e.g., workers that are both exposed occupationally and also
reside in neighborhoods with high ambient concentrations of air pollution) that, when combined with
higher responsiveness, can further increase the risk for adverse health effects. Text Box 4.1 provides
an overview of the literature on increased vulnerability to the effects of climate change.
49 Several conceptual frameworks explicitly integrate social context into the exposure-disease paradigm to highlight how these
factors may interact with environmental exposures to yield health differences (Morello-Frosch and Jesdale, 2006; Gee and
Payne-Sturges, 2004).
50 See Schwartz et al. (2011a, 2011 b) for several examples of how these risk- or effect-modifiers may increase risk.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
25
-------
Text Box 4.1 Increased Vulnerability to the Effects of Climate Change
Changes in global temperatures due to climate change are expected to result in changes in average
annual temperatures in the U.S. For instance, global warming of 2°C by 2100 is projected to result in
average annual temperature increases of between 3°C and 4°C for large portions of the country
(U.S. EPA, 2021c). These changes in temperature and other changes to our natural systems are
expected to affect human health in myriad ways over the coming decades. These effects are due to
changes in the frequency, duration, intensity, timing, and location of extreme events such as heat
waves, floods, wildfires, and droughts. These extreme events can also affect human health through
their effects on vector-, food-, and waterborne infectious diseases, through changes in temperature-
related mortality, climate-driven air pollution exposure, flooding-related property damage, and effects
on labor productivity, to name a few. These effects are not expected to be evenly distributed across
the U.S. population (U.S. EPA, 2021c; USGCRP, 2018).
Low-income and predominantly non-White communities are especially vulnerable to these and other
effects of climate change because of their limited adaptive capacity; dependence on climate-
sensitive resources, such as local water, fish, and food supplies; and inadequate access to
information and recovery resources (USGCRP, 2023; USGCRP, 2018; IPCC, 2018; NASEM, 2017;
USGCRP, 2016; IPCC, 2014; NRC, 2011b). For example, low-income households typically have
limited access to healthcare and often do not have adequate insurance. Workers in outdoor
occupations such as agriculture or construction may not be able to avoid working on high-
temperature days without significant loss of income. Non-English speaking and disabled individuals
may have more difficulty accessing flood or fire hazard alerts, evacuating safely, or accessing aid
after natural disasters (U.S. EPA, 2021c; USGCRP, 2016).
In addition, health conditions such as cardiovascular or respiratory illnesses that occur at higher
rates in many socially and economically vulnerable communities may also be exacerbated by the
effects of climate change (USGCRP, 2023; USGCRP, 2016). Outdoor workers, who frequently are
comprised of already at-risk groups, are also more likely to be exposed to poor air quality and
extreme temperatures (U.S. EPA, 2021c). Low-income households may also face increased food
insecurity as climate change reduces food availability and increases prices (USGCRP, 2018;
USGCRP, 2016).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
26
-------
Chapter 5: Considering Environmental
Justice Concerns When Planning a Human
Health Risk Assessment
This chapter provides guidance to Agency analysts on integrating EJ concerns into the planning of a
human health risk assessment (HHRA) conducted to support a regulatory action. As noted in the EPA's
Framework for Human Health Risk Assessment to Inform Decision-Making (referred as the HHRA
Framework) (U.S. EPA, 2014b), EJ concerns are a key consideration in the early stages of HHRA,
including planning and scoping, and problem formulation.
5.1 Introduction
Human health risk assessment is a complex, iterative, and multidisciplinary process intended to inform
decision-makers about the effects of environmental stressors on human health and to support the
formulation of policy actions that impact these stressors. EPA risk assessment has well-established
procedures, methods, and tools to inform regulatory decision-making, but it also evolves with
experience and the advance of new scientific capability (U.S. EPA, 2014b). The following questions,
outlined in Section 3.1 (and repeated here), are important to consider during HHRA planning to help
ensure that the HHRA provides relevant information about differences in risks for population groups of
concern that can then be used as inputs into the EJ analysis for a regulatory action (see Chapter 6):
• Baseline: Are there existing EJ concerns associated with environmental stressors affected by
the regulatory action for population groups of concern?
• Regulatory options: For the regulatory option(s) under consideration, are there potential EJ
concerns associated with environmental stressors that are affected by the regulatory action for
population groups of concern?
• Mitigation or exacerbation of effects: For the regulatory option(s) under consideration, are EJ
concerns exacerbated, mitigated, or unchanged compared to the baseline?
Even when methods and data relevant to these questions are not available, their consideration can
highlight where additional data or research are needed. HHRA and the science and practices that
support it continue to evolve. As new tools and information become available, incorporating EJ
considerations into HHRA should also evolve to reflect improved risk assessment methodologies and
guidance.
This chapter first provides an overview of modeling and data needs for evaluating EJ concerns (Section
5.2), then discusses how EJ concerns can be considered when planning a HHRA (Section 5.3). This
discussion is guided by EPA's existing HHRA Framework and highlights considerations during the initial
phases of risk assessment: planning and scoping, and problem formulation. The chapter concludes
with a brief summary of possible approaches for considering the cumulative effects of multiple
exposures (Section 5.4).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
27
-------
5.2 Overview of Modeling and Data Needs for Evaluating EJ Concerns
HHRA seeks to characterize the nature, probability, and magnitude of current or future risks of adverse
human health effects related to exposure to environmental stressors. This can include both quantitative
and qualitative characterizations of risk (U.S. EPA, 2014b; NRC, 1983) and may incorporate different
approaches, methods, and metrics, depending on the nature of the decision that the assessment is
intended to inform. The EPA has published guidance on all steps of the HHRA process, as well as the
HHRA Framework (U.S. EPA, 2014b).
Analyzing EJ concerns as part of HHRA requires spatially and temporally resolved data, including
environmental exposures, health data, and demographic information. However, information on EJ
concerns is often qualitative and place specific, which results in several challenges when considering
EJ in HHRA.
5.2.1 HHRA Requires Sufficient Data and Modeling Capabilities to Incorporate EJ
Considerations
As noted by the NRC "the goal [of HHRA] of achieving accurate, highly quantitative estimates of
risk... is hampered by limitations in scientific understanding and the availability of relevant data..."
(NRC, 2009). These data requirements may be even greater when it comes to risk estimates to inform
EJ considerations due to complexities of exposure and the existence of additional stressors. For
example, modeling how non-environmental stressors (such as nutritional deficits and stress) interact
with specific chemicals to exacerbate or mitigate individual health endpoints is still at a nascent stage of
development. However, population-level vulnerability due to these types of non-environmental stressors
can be informative for establishing increased susceptibility to exposures to pollutants.
Figure 5.1 illustrates how considering intrinsic and extrinsic factors, including both environmental and
non-environmental stressors, can impact risk. The first row, baseline risk, depicts how exposure above
a clinically defined threshold leads to adverse health outcomes in the general population. As we also
consider intrinsic factors such as age and genetics, a greater proportion of the population experiences
an adverse health effect. Adding extrinsic factors, such as systemic racism and exposure to hazardous
chemicals results in a more accurate estimation of the population with an adverse health outcome.
Though not depicted in the figure, this concept could be extended to broader characterizations of
adverse outcomes (e.g., lost workdays).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
28
-------
Figure 5.1 How Intrinsic and Extrinsic Factors Influence Adverse Health Outcomes in a
Population
Clinically-defined
threshold
No Adverse Outcome Adverse Outcome
********M
i
Intrinsic Factors
(ex. age. genetics)
Baseline Risk +
Intrinsic Factors
(ex. age. genetics) +
Extrinsic Factors
(ex. systemic racism, exposure
to hazardous chemicals)
Physiological
P Parameters
(ex. blood glucose
**********
levels, blood pressure)
(Source: Varshavsky et al., 2023)51
Data needed for HHRA to more fully consider EJ may be difficult to find, particularly given the
importance of considering it within a holistic framework. For example, robust national studies of health
outcomes, and their variability, are often limited because specific measurements of non-environmental
stressors vary locally. Many studies target populations whose socioeconomic status is higher than the
national average (e.g., White, male adults) and who reside in urban areas that have adequate
monitoring and therefore may not be generalizable to populations with EJ concerns (Pelacho et al.,
2021; Payne-Sturges, 2011).52 That said, a growing literature is focused on vulnerable population
groups, including Medicare populations (Di et al., 2017), Hispanics and Latinos (Letellier et al., 2022),
Native Americans, Asians, and those who experience low income (Josey et al., 2023; Liu and Eichen-
Miller, 2021).
National data can have limited utility for informing health disparities and there are limitations to
extrapolating community-level data from national surveys (Dosemagen and Williams, 2022; Nweke et
al., 2011). See Chapter 7 for a discussion of identified research priorities for improving the EPA
analysis of EJ concerns, which elaborates upon these modeling and data gaps.53
51 Used under Creative Commons license (https://creativecommons.Org/licenses/bv/4.0/). No changes were made to the
original figure.
52 In the absence of scientific data to fully characterize the range of responses to chemical exposures, the EPA employs
default assumptions, such as uncertainty factors used in non-cancer risk assessments, to account for human variability. See
the EPA's risk assessment portal (www.epa.gov/risk) for guidance documents on these uncertainty factors. As noted by the
EPA's Science Advisory Board (SAB, 2015), however, "...the use of uncertainty factors in developing dose-response
assessments for an individual level chemical might address the general population as a whole but does not specifically
address differential or disproportionate vulnerability."
53 Varshavsky et al. (2023) notes that emerging tools and data "that better account for human variability and susceptibility
include probabilistic methods, genetically diverse in vivo and in vitro models, and the use of human data to capture underlying
risk and/or assess combined effects from chemical and non-chemical stressors."
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
29
-------
5.2.2 It Can Be Difficult to Incorporate the Cumulative Effects of Multiple, Dissimilar
Stressors into HHRA
Communities with EJ concerns are often exposed to many environmental and non-environmental
stressors through multiple pathways. HHRA is most often conducted on a chemical-by-chemical basis
using single exposure-to-effect pathways.54 Broadening the scope of HHRA to incorporate other
environmental and non-environmental stressors requires addressing challenging scientific and technical
questions (e.g., what effect does exposure to pesticides have on rural or Indigenous populations? How
does living near an environmental hazard affect vulnerability to other environmental concerns?).55
In addition, adoption and use of a framework for cumulative risk assessment (CRA) is rare among
practitioners (Clougherty and Rider, 2020). Cumulative risk assessment that incorporates both
environmental and non-environmental stressors requires more comprehensive data than traditional
HHRA (e.g., information on background exposure or health status among specific populations).56 While
the EPA's SAB (SAB, 2015) continues to recommend use of HHRA, it encourages the EPA to develop
further guidance for quantitative and/or qualitative evaluation of cumulative effects of exposure to
multiple stressors on human health and well-being. See Section 5.4 for additional discussion.
5.2.3 HHRA Requires Effective Communication, Given its Highly Technical Nature
HHRA has been criticized for often having limited consideration of public perceptions of risk and, more
broadly, for limiting public input into the process. Both factors are critical to the assessment of EJ
concerns (Sexton and Under, 2010).57 The technical complexity of HHRA can also lead to a lack of
transparency and accountability (SAB, 2015).
Effective communication about HHRA is therefore necessary for meaningful public involvement, which
is not only a critical aspect of the HHRA process (see Figure 5.2) but is also essential for decision-
making. Payne-Sturges (2011) notes that "when affected citizens actively participate in the process to
better understand science and inform policy responses, better decisions emerge as a result." Risk
communication can increase community involvement in decision-making processes, better inform risk
assessors regarding community perceptions of risk, and provide opportunity for dialogue about the
limits and opportunities of policymaking efforts (Barzyk et al., 2015). See Section 2.3 for further
discussion of meaningful engagement and Section 5.3.1.2 for further discussion of risk communication
in the context of environmental justice.
54 Some assessments have also evaluated the risk associated with exposure to multiple chemicals that act by similar
mechanisms (Cattaneo et al., 2023; Backhaus and Faust, 2012).
55 While there is limited quantitative data available on non-environmental stressors, how they interact with chemicals is a
growing area of research (Aker et al., 2020; Payne-Sturges et al., 2018), including methods to combine multiple and dissimilar
stressors (Bakkensen et al., 2024).
56 While the literature often uses the terms chemical and non-chemical stressors, we use the terms environmental and non-
environmental stressors in this report to capture the range of chemical, physical, or biological agents, contaminants, or
pollutants that may be subject to a rulemaking.
57 HHRA is framed in terms of the risk of an adverse outcome from a specific stressor or stressors, but EJ advocates and
analysts are often interested in broader concepts of health (Barzyk et al., 2015). For example, HHRA as traditionally practiced
does not quantify factors such as fairness, voluntariness, responsibility, control, trust, reversibility, and identifiable victims
(Sexton and Under, 2010; Corburn, 2002), though these factors may be identified and characterized as part of risk
management discussions. HHRA methods also do not typically consider public attitudes toward risk or incorporate community
insights and priorities around sources of risk (Sandman, 1989).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
30
-------
5.3 Considering EJ Concerns When Planning a HHRA
When considering EJ concerns, it is important that analyses conducted in support of regulatory actions
explicitly consider both baseline health risks and whether changes in risk due to policy actions may
accrue differentially within population groups of concern. Specific demographic and socioeconomic
attributes (e.g., age, poverty) can be correlated with increased vulnerability to environmental stressors.
Also, the burden of health problems and differential environmental exposures associated with race,
ethnicity, income, or other relevant demographic and socioeconomic characteristics may overlap with
other factors that influence vulnerability such as lifestage, genetic predisposition, or pre-existing health
conditions (Shao et al., 2022). See Chapter 4 and Section 6.2 for further discussion.
5.3.1 Overview of the HHRA Process
The EPA's HHRA Framework provides an overview of the risk assessment process (Figure 5.2). The
HHRA process is not strictly sequential; steps are often performed together in an integrative fashion,
and EJ can be considered at multiple points in the HHRA process. Ultimately, the final step—risk
characterization—synthesizes information from the other steps and provides the basis for
communicating the results to decision-makers and the public.
The planning and scoping and the problem formulation phases are key initial elements of the HHRA
Framework. At the planning and scoping phase, analysts define the process for conducting the risk
assessment and establish its analytic scope. The problem formulation phase informs the HHRA
technical approach. Important outcomes of this step are a conceptual model that describes the
relationship between stressors, exposure pathways, exposed lifestages, populations, and health
endpoints that will be addressed in the risk assessment, as well as an analysis plan for conducting the
assessment (U.S. EPA, 2014b). The consideration of EJ throughout the risk assessment planning and
scoping and the problem formulations phases is important to ensure an effective assessment.
Figure 5.2 also illustrates two key elements to consider throughout the risk assessment process, from
initiation to informing decisions: fit-for-purpose and public and community involvement.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
31
-------
Figure 5.2 Framework for Human Health Risk Assessment to Inform Decision-Making
Adapted from U.S. EPA (2014b)
5.3.1.1 Fit-for-Purpose
Fit-for-purpose refers to the concept that risk assessments and associated products should be suitable
and useful for their intended purpose(s), particularly for informing choices among risk management
options (U.S. EPA, 2014b). Accordingly, throughout the process of planning and performing HHRAs, it
is important to evaluate whether the assessment effectively addresses the information needs of
decision-makers. This is primarily accomplished through transparent dialogue between risk assessors
and risk managers early in the assessment process.
While the nature and scope of a HHRA may vary by statute and the type of problem being addressed
(and will be limited by available data, methods, and resources), a robust fit-for-purpose process that
includes consideration of its purpose in relation to evaluating EJ concerns will maximize the usefulness
of the assessment. A clearly defined purpose informs many aspects of HHRA planning, including how
to frame the questions, which datasets to use, which analyses to perform, and how to synthesize and
interpret results.
Because fit-for-purpose will be continually evaluated throughout the risk assessment process, it is
important to raise key questions early and even revisit questions that bear on whether the resulting risk
characterization will be able to inform EJ concerns. These questions include:
1. What types of individuals or population groups face higher risks in the baseline relative to the
average or comparable individuals in the general population?
Technical Guidance for Assessing Environmentai Justice in Regulatory Analysis - Second Edition
32
-------
2. What types of individuals or population groups could experience higher risks relative to the
average or comparable individuals in the general population as a result of a regulatory option?
3. What are the reasons why an identified population group (or lifestage within a population group)
may potentially experience higher risk than the average person?
4. What tools and data are available to estimate and characterize the potential for differences in
risk for affected groups? (This is especially relevant when developing the HHRA analysis plan.)
5. How can information about differences in risk for affected groups, or the potential for these
differences, be effectively communicated in the risk characterization?
5.3.1.2 Public and Community Involvement
Public and community involvement is integral to both the HHRA process and the broader consideration
of EJ concerns. For example, engaging Tribes early in the process, as relevant, may lend new insights
into how they define human health (U.S. EPA, 2019a) and allow for consideration of Indigenous
Knowledge (also referred to as Traditional Ecological Knowledge). See Section 2.3 for more discussion
of meaningful engagement.
To enable communities to meaningfully participate, it is important to recognize and address conditions
that could reduce or hinder their ability to participate in the HHRA and regulatory process. These could
include time and resource constraints; barriers due to limited English proficiency or disability; lack of
trust; lack of information; and difficulty in accessing and understanding complex scientific, technical,
and legal resources. Public participation may also be hindered by socio-political dynamics that serve to
weaken democratic processes, which raises the importance of finding approaches that function in
contexts of social distrust and creating positive participatory experiences (Webler and Tuler, 2021). For
instance, some forms of public participation may favor individuals with more political influence and do
not necessarily lead to more equitable outcomes. It is therefore paramount that the EPA is intentional
about creating opportunities for meaningful public involvement so that risk-based decisions do not lead
to unsatisfactory community outcomes (Sexton, 2013).
See Chapter 3 of the HHRA Framework (U.S. EPA, 2014b) on how to involve the public and the
broader community in the risk assessment process. Text Box 5.1 also highlights the potential role of
participatory science in further enhancing data available for HHRA. See Section 2.3 for more discussion
of meaningful involvement in the context of analysis.
In planning a HHRA, information gained through public involvement can be particularly valuable for
constructing exposure scenarios in scoping (Section 5.3.2) and for identifying both data and data gaps
as part of the analysis plan (Section 5.3.3.2).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
33
-------
Text Box 5.1 Participatory Science
Participatory science engages the public in advancing scientific knowledge by formulating research
questions, collecting data, and interpreting results. This includes a broad and inclusive range of
activities, from those originating in academic and government institutions that enlist the public in
data collection to create knowledge to community-led projects intended to identify potential EJ
issues and community concerns.
For HHRA planning and scoping, participatory science may be able to fill information gaps and
provide useful data that informs analysis. For example, projects may collect monitoring data in
overburdened communities. For data generated by the public to have a meaningful impact, quality
assurance during data gathering is critical. The EPA has developed a Handbook for Citizen Science
Quality Assurance and Documentation (U.S. EPA, 2019b) to help guide data collection efforts and
emphasize the high standards required for data to be used in regulatory decisions. See
https://www.epa.qov/participatorv-science/qualitv-assurance-handbook-and-toolkit-participatorv-
science-projects for the handbook and related materials.
Participatory science may also facilitate meaningful community involvement in the HHRA process.
As described in the EPA's Vision for Participatory Science (U.S. EPA, 2022e), it can create a
stronger, more inclusive, and collaborative network of individuals dedicated to environmental
problem solving. It can also contribute to effective risk communication by improving public
understanding of environmental issues and actions to address them.
Note: Other terms sometimes used to refer to participatory science include citizen science, community
science, volunteer monitoring, or public participation in scientific research.
5.3.1.3 Risk Communication
Risk communication is intended to provide the public with the information it needs to make informed,
independent judgments about risks to health, safety, and the environment. The EPA has extensive
resources on effective risk communication, including frameworks, tools, and case studies.58
To identify potentially affected members of the public, the Presidential/Congressional Commission on
Risk Assessment and Risk Management suggests using the following questions:59
• Who might be affected by the risk management decision?
• Who has information and expertise that might be helpful?
• Who has been involved in similar risk situations before?
• Who has expressed interest in being involved in similar decisions before?
• Who might reasonably feel they should be included?
Analysts and risk managers can consult the Achieving Health and Environmental Protection Through
EPA's Meaningful Engagement Policy (U.S. EPA, 2024d) for general guidance when designing a
meaningful engagement and public participation process. To ensure that the public can participate
58 See the EPA's Risk Communication page at https://www.epa.qov/risk-communication.
59 See the EPA's Presidential Commission on Risk Assessment and Risk Management website:
http://cfpub.epa.gov/ncea/risk/recordisplav. cfm?deid=55006&CFID=55036505&CFTQKEN=43224210.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
34
-------
meaningfully in the HHRA process, the approach for soliciting information needs to be specific, involve
interactive dialogue that is designed to elicit specific responses, and include accommodations for
population groups facing additional barriers to participation, such as those in rural areas, with limited
English proficiency, or a disability. Elements of such a dialogue may include specific questions about
the types of data or models that are needed to adequately reflect EJ concerns.60
5.3.2 Planning and Scoping
Planning and scoping is the stage in which the risk assessor defines the process for conducting the risk
assessment and identifies its general scope. This activity contributes to the development of a sound
risk assessment that will eventually accomplish its intended purpose. It also assists those interested in
the risk assessment process in understanding the broader context and how the anticipated results will
satisfy the intended purpose. A broad range of technical experts (e.g., toxicologists, epidemiologists,
statisticians, economists, and other social scientists) working as a team may be involved in this stage.
The HHRA Framework identifies several aspects of planning and scoping. Of note for EJ
considerations are the context, purpose, and scope of the risk assessment; and planning scientific peer
(and other) review. As described throughout this chapter, informing EJ considerations requires greater
emphasis on the human receptor characteristics in many of these steps, which is a shift from a more
conventional focus on stressor characteristics.
5.3.2.1 Defining the Context, Purpose, and Scope of the Risk Assessment
EPA human health risk assessments conducted for rulemaking are intended to inform and support risk
management decisions and the development of regulatory actions, focusing on the changes in health
risk that are anticipated. When planning a risk assessment for any purpose, analysts should clearly
identify the decision(s) that will be supported by the analysis, specify the boundaries for the
assessment (i.e., the scope), and detail what will not be addressed in the risk assessment.
Context of the Risk Assessment
A key first step in a HHRA is documenting the context for the analysis. This includes the regulatory and
statutory context as well as the physical and cultural context for the investigation. To provide regulatory
and statutory context that considers EJ issues, analysts need to identify any complementary
requirements between the triggering statutory authority and E.O.s 12898 and 14096 that focus on
identifying and addressing potentially disproportionate and adverse health effects and risks. For
example, a new regulation of coal combustion residuals (CCR) and how they are managed will have an
impact on the community surrounding the waste site, as well as along the roadways traveled by large
vehicles moving CCR waste.
60 When EPA actions or decisions may affect Tribes, the EPA Tribal consultation policy provides clear guidelines for when,
how, and the types of actions or decisions for which consultations with Tribal governments should occur (U.S. EPA, 2023f).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
35
-------
In addition to the specific policy context, other contexts may also be relevant and important to consider.
For example, background exposure to chemicals from multiple sources that might affect responses, or
an enhanced background risk for a relevant adverse health outcome due to other factors that may
complicate measurements, are important for assessing differences in risk. In addition, communities with
EJ concerns may experience differential risks due to higher vulnerability (e.g., due to lifestage or pre-
existing health conditions) or other factors influencing exposures (e.g., behavioral patterns or proximity
to sources of exposure).61 Together, these existing conditions help frame and clarify the context of the
HHRA.
Purpose of the Risk Assessment
Clearly articulating the overall purpose of an assessment is an iterative process and may involve
extensive interaction between the assessment team and affected members of the public to establish a
common understanding and ensure that relevant community concerns are considered. In addition, in
this step analysts may conduct a review of data needs and limitations to ensure that the results may
adequately inform decision-makers and members of the public (NRC, 2009).
The assessment's purpose and scale (e.g., regional or national) will have significant implications for the
assessment's scope, level of detail, and approach. Key considerations at this stage include:
• What specific policy or regulatory decision is to be informed by the risk assessment? When is
the decision anticipated? What are the risk management options being considered?
• What legal or statutory requirements affect risk management options and the level or type of
analysis? (U.S. EPA, 2014b)
To ensure that a HHRA generates useful information, risk managers and analysts should develop
concise statements of risk management and analytic objectives that incorporate EJ concerns. See Text
Box 5.2 for an example. As risk managers and analysts develop these objectives, it is important to
frame them such that the data generated by the HHRA can be used to respond to the main EJ analytic
questions from Section 3.1. Analytic objectives for incorporating EJ concerns should concisely identify
the evidence to be collected; the direction and structure of the planned consideration of EJ concerns;
the analytic methods to be employed (e.g., between socioeconomic group comparisons); the type of
data required; and the scope of the analysis (e.g., national versus local scale).
Scope of the Risk Assessment
Scoping establishes the boundaries of the HHRA (e.g., what population groups, health effects,
chemicals, and exposure pathways will be included in the assessment). These boundaries should align
with and support the stated purpose of the risk assessment.
61 As an example, primary National Ambient Air Quality Standards (NAAQS) are required to protect public health, including the
health of sensitive or at-risk population groups, with an adequate margin of safety. The Integrated Science Assessment (ISA)
and its supplement cite extensive evidence indicating that "both the general population as well as specific populations and
lifestages are at risk for PIVh s-related health effects" (U.S. EPA, 2022d; U.S. EPA, 2019c). Factors that may contribute to
increased risk of PIVhs-related health effects include lifestage (e.g., children and older adults), pre-existing diseases (e.g.,
cardiovascular disease and respiratory disease), race/ethnicity, and socioeconomic status.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
36
-------
Text Box 5.2 Considering EJ Concerns for the 2015 Coal Combustion Residuals
Generated by Electric Utilities Final Rule (U.S. EPA, 2015d) - Examples of
Risk Management and Analytic Objectives
Regulatory Context: The Resource Conservation and Recovery Act (RCRA) gives the EPA
authority to regulate non-hazardous wastes. RCRA solid waste regulations ban open dumping of
waste and set minimum federal criteria for the operation of solid waste landfills and impoundments.
The EPA established requirements for disposal of coal combustion residuals (CCRs or coal ash) at
landfills and surface impoundments in the 2015 CCR Final Rule. The regulation includes structural
integrity requirements to reduce catastrophic failure risk; groundwater monitoring and corrective
action; location restrictions; liner design and operating criteria; closure and post-closure
requirements; and record keeping and notification, including public websites documenting monitoring
and corrective action planning. These requirements reduce releases of arsenic and other
contaminants to water and air, help ensure actions taken by power plants to comply with the rule are
transparent, and that the communities impacted by the disposal of CCRs have the information they
need to understand risks.
Risk Management Objective: Evaluate the effect of the 2015 CCR rule on the potential for
increased risk to human health and the environment.
Example Questions:
• What hazards from corrective action, operations and closure and post-closure requirements,
including accidental releases of CCR, could pose hazards resulting in differential risks to
population groups and communities of concern?
• How does the rule affect the likelihood of such hazards compared to pre-2015 regulations?
• Are there potential CCR management hotspots of concern?
Analytic Objectives: (1) Evaluate whether baseline demographic characteristics of population
groups potentially affected by the rule differ from the broader population; (2) Evaluate whether other
factors that affect the potential for differential risk are present under the rule; and (3) Evaluate
whether the potential for exposure to communities of concern is exacerbated or mitigated by the
2015 CCR rule.
Example Questions:
• Do communities near facilities subject to the rule have higher percentages of population
groups of concern relative to national, state, or regional populations?
• Are the communities potentially affected by the 2015 CCR rule also affected by other sources
of pollution (e.g., industrial facilities, landfills, leaking underground storage tanks)?
• Do other factors contribute to increased vulnerability (e.g., vulnerable lifestages, nursing
mothers) among population groups of concern?
• Does the 2015 CCR rule reduce or improve the ability for potentially impacted communities to
participate in the decision-making process?
• Does the regulation reduce existing (baseline) disparities in risk?
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
37
-------
At this step, most EPA assessments focus on identifying and considering information available in areas
such as the sources of contaminants, stressors, and associated effects; exposure routes and pathways;
and vulnerable populations or lifestages. Incorporation of public concerns is also important at this stage
(and should be built upon throughout the assessment process).
Depending on the nature of the assessment, consulting with affected communities can provide valuable
information when identifying exposure routes, duration, pathways, and other information for
constructing exposure scenarios (U.S. EPA, 2014b). This information may help reduce the need for
default assumptions in the assessment. The EPA has developed extensive guidance on community
and public involvement for this purpose and continues to update its guidance specifically for EJ-related
contexts (U.S. EPA, 2024d; 2015b; 2013b). See Section 2.3 for a general discussion of meaningful
engagement.
In the scoping stage of a HHRA that supports a regulatory action, analysts should consider several
questions that can aid in identifying EJ concerns:
• Which population groups, as characterized by geographic location, ethnicity or race,
gender, occupation, age, baseline health status or other factors, should be part of the
assessment? In some instances, the presence of risk- or effect-modifiers may mean that some
types of individuals or communities are at greater risk for experiencing adverse effects, and
there can be interactions among these modifiers. In identifying population groups for the
assessment of differential risks, analysts should consider the extent to which risk- or effect-
modifiers, by themselves or in combination, may explain differences that vary by demographic
or socioeconomic characteristics. If analysts decide to assess population groups defined by risk-
or effect modifiers, the rationale for this decision and the associated methods needs to be
transparently documented.
• What health endpoints are to be addressed by the assessment? Defining health endpoints
clearly in the planning and scoping phase focuses the risk assessment and increases the
transparency of the process. When selecting health endpoints, analysts should consider
whether specific health endpoints may be significant in population groups of concern. In making
this selection, it is important to evaluate whether baseline health endpoints for a given exposure
differ across population groups and to consider what intrinsic and extrinsic factors might
contribute to variation. This type of information is most often found in epidemiology and
toxicology studies, such as those focused on the modifying effects of social context on
environmental risk. It may not be possible to identify all relevant health endpoints at the
beginning of the HHRA. Some information found in toxicity assessments may only define the
potential for an adverse health outcome for specific stressors. Information from sources such as
EJScreen as well as community insights may also identify areas of further research as the
HHRA progresses.
• What exposure routes and pathways are relevant? Do specific exposure pathways
potentially lead to specific effects? What exposure scenarios should be modeled?
Analysts should evaluate whether population groups of concern may have different exposure
routes, pathways, or contact scenarios from the general population. Scoping for an exposure
assessment should include timing and duration of historical, current, and future exposure.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
38
-------
Unique exposure pathways based on lifestages, cultural practices, and other relevant categories
may also be considered.62 Different pathways of exposure (e.g., inhalation, dermal, ingestion)
may produce different effects with varying levels of severity. See Appendix B.
At the completion of the scoping step, analysts will have a well-defined context for the analysis, a set of
key analytic priorities to evaluate key policy questions, and a set of boundaries for the HHRA that
reflect how the analysis will address its analytic priorities. All of these elements can be incorporated into
the problem formulation phase to produce a detailed plan for the assessment.
5.3.2.2 Scientific Peer Review or Other Reviews
During the planning and scoping phase of a HHRA, analysts should also consider the need for and
timing of peer review. Peer review is a documented process conducted to ensure that activities are
technically supportable, competently performed, properly documented, and consistent with established
quality criteria (U.S. EPA, 2014b).63 When a HHRA that incorporates EJ concerns is subject to scientific
peer review, subject matter experts may include community representatives with technical expertise
and scientists with community and EJ experience. Peer review usually involves a one-time or limited
number of interactions by the independent peer reviewers with the authors of the work product.
An assessment also may benefit from other types of input, such as peer involvement and public
comment. Planning and scoping for the assessment includes discussion of whether and what types of
reviews will be included in light of the context and constraints for the assessment, including schedule
and resources (U.S. EPA, 2014b). In addition, risk assessors may rely on existing peer-reviewed
literature to consider topics such as cumulative risk or differences in effects across communities.
5.3.3 Problem Formulation
Problem formulation builds on the planning and scoping phase to identify major factors to consider in
the HHRA and informs its technical approach. Two important products from problem formulation are:
• A conceptual model to describe the linkages between stressors and adverse human health
effects, including the stressor(s), exposure pathway(s), exposed lifestage(s) and population(s),
and health endpoint(s) that will be addressed in the risk assessment.
• An analysis plan based on the conceptual model to describe the approach for conducting the
risk assessment, including its design, methods, key inputs, intended outputs, and assessments
of uncertainty and variability that might specifically affect communities with EJ concerns.
Like the planning and scoping phase, problem formulation is often an interactive, nonlinear process,
and substantial re-evaluation is anticipated in the development of resulting products.
In considering EJ, problem formulation focuses on identifying whether population groups of concern
experience elevated risks relative to the broader population or other appropriate comparison population
group (see Section 6.5.2), both in the baseline and in response to policy changes. Specifically, this
62 An important consideration is to develop exposure scenarios for Tribal and Indigenous populations that draw from historical
(pre-European settlement) time frames and account for the many irreversible changes that have occurred in the United States
that have fundamentally altered the practice of traditional Tribal lifeways (see U.S. EPA, 2019a).
63 Guidelines for the peer review process are available in the EPA's Peer Review Handbook', http://www.epa.gov/osa/peer-
review-handbook-4th-edition-2015.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
39
-------
involves: (1) clarifying the relevant source and characteristics of the stressors; (2) identifying factors
that may influence exposures that contribute to those risks; and (3) characterizing vulnerabilities of the
population groups of concern that may exacerbate exposure or risk.
Text Box 5.3 provides examples of EJ-related questions that may be raised during problem formulation.
For additional sample problem formulation questions, see U.S. EPA (2014b).
5.3.3.1 EJ Considerations when Developing the Conceptual Model
Conceptual models consist of (1) a set of risk hypotheses that describe predicted relationships among
stressor(s), exposure(s), and health endpoints and/or responses, along with the rationale for their
selection; and (2) a diagram that illustrates the relationships presented in the risk hypotheses.64
Generally, the conceptual model addresses the following with respect to EJ concerns:
• How and to what degree identified risk factors contribute to differences in exposure and/or risk;
• The strength and direction of relationships between these risk factors and exposure and/or risk;
• Identification of data needs by characterizing these relationships as low, medium, and high
uncertainty; and
• Scope of the assessment as to EJ concerns given current scientific understanding.
Characterizing the Stressor and its Sources
The properties and sources of the stressor(s), and how these may drive differential risks, are important
to consider in the context of EJ. This includes the source(s) of regulatory concern - e.g., what is the
likelihood that the sources of the stressor(s) are located in areas where population groups of concern
live or experience exposure? But it can also include, where relevant and appropriate, identifying the
distribution of additional sources of the stressor(s) that are not the focus of the regulatory action
because they may contribute to differential risks. For example, a stressor may be present in
environmental media due to background concentrations (e.g., resulting from historical or past industrial
activity, or natural occurrence) in areas with population groups of concern.
64 The HHRA Framework (U.S. EPA, 2014b) provides descriptions, resources, and examples of conceptual models.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
40
-------
Text Box 5.3 Examples of EJ-Related Questions to Consider During Problem Formulation
Characteristics Related to Proximity to a Stressor or Source
• What are the sources of the stressor?
• Is the source located in geographic areas with greater proportions of population groups of
concern?
• Are other sources of the stressor more prevalent in these geographic areas?
• Are there historical releases or uses of the stressor in such areas?
• Is the concentration of the stressor in the relevant ambient media higher in these geographic
areas?
• Does each stressor have multiple sources that should be evaluated?
Differential Exposures to a Stressor
• Do population groups of concern have higher body burdens of the contaminant (e.g., higher blood
concentrations)?
• Are these population groups more likely to experience current or historically higher exposures to
the stressor from sources other than the one under consideration?
• Are there particular lifestages within these population groups that may be more at risk to higher
exposure to the stressor?
• Are there products/consumer goods that contain the stressor?
• Are these products/consumer goods used at noticeably higher rates among population groups of
concern (e.g., use of personal care or cleaning products that contain harmful chemicals)?
• Are there cultural practices or other activities that are unique to these population groups (e.g., fish
consumption for sustenance or cultural reasons)?
• What is the frequency and duration of occurrence of the unique cultural practice or atypical
activity?
• Is proximity to the emitting source an important factor in assessing differential exposure?
• What geographic scale is important to highlight differential exposures between population groups
for the pollutant in question?
Population Characteristics
• What are the rates of the adverse health outcome among population groups of concern?
• Are the rates of the adverse health outcome higher among these population groups?
• What factors or conditions are known to modify the effect of the contaminant?
• How are these modifying factors or conditions distributed across population groups?
• Do population groups of concern have a higher prevalence of modifying effects or conditions?
• Are members of these population groups employed in specific professions known to have higher
risks of the adverse health outcome?
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
41
-------
Identifying Exposure Pathways
It is important to clearly articulate how population groups may be exposed to one or more stressors.
That is, it is key to describe the exposure pathways experienced by population groups and to identify
unique exposure pathways relevant to assessing EJ concerns.65 Burger and Gochfeld (2011) suggest
that the first step in improving the risk methodology is to recognize and account for unique exposure
sources (e.g., hand-to-mouth behavior of small children; use of personal care or cleaning products that
contain harmful chemicals; fish consumption for subsistence or cultural reasons) and the corresponding
exposure pathways. Examples of questions helpful for extracting information about unique exposure
pathways are presented in Text Box 5.3.66
New pathways can be identified during or after planning as new data become available. For example,
biomonitoring data acquired during the assessment may provide evidence of unexpected health
differences, resulting in additional analyses of exposure pathways that may cause these differences. It
also may be useful to seek new information about certain exposure pathways to ensure a
comprehensive evaluation of the range of exposures in the population groups of concern.
In more data limited settings, it may be helpful to use national databases and screening tools to identify
the potential for differential exposure for specific population groups of concern. Screening-level
assessments can also be used to exclude exposure pathways of minor importance from further
consideration or to determine where additional data and information is needed to evaluate key
pathways (U.S. EPA, 2019a). See Appendix B for additional discussion.
Identifying Differences in Exposures that May Lead to Differential Risks
Differences in exposures can be an important indicator of differential risks across population groups of
concern. For example, if the regulated sources are co-located with other sources of the same stressor,
this may contribute to significant differences in patterns of exposure to the stressor.
Evaluated patterns of exposure can be location-specific or population group-specific, depending on the
scale of the assessment and the types of data available. Differences in cultural practices, use of
specific consumer products, and behaviors can lead to differences in exposures. Considering other
characteristics, such as lifestage, gender, or income, can further clarify which population groups may
face higher exposures. For example, children living in older housing, lower income households, and
individuals identifying as Non-Hispanic Black have higher blood lead levels in the United States (U.S.
EPA, 2024g). In addition, due to exposure via personal care products, Black women are more highly
exposed to endocrine disrupting chemicals than White women (Helm et al., 2018).
Identifying Population Characteristics that May Lead to Differential Risks
Population characteristics refer to those attributes shared by individuals within a population group that
influence not only the likelihood of exposure to a stressor but may also affect the risk of adverse health
outcomes from this exposure. These characteristics range from those with direct health effects, such as
pre-existing disease conditions, chronic disease, age, medication status, and immune status, to those
with more indirect influences, such as a lack of access to resources (e.g., health care, transportation),
age of housing, occupation, income, and educational status. Group differences in body burdens of the
65 Examples of such pathways include exposure to heavy metals from the use of non-traditional medicines, mercury from high
fish consumption, pesticides tracked into homes from places of work, and inorganic mercury from cosmetics.
66 The Exposure Factors Handbook (U.S. EPA, 2011a) also contains exposure factor data stratified by race and ethnicity. The
Guidelines for Human Exposure Assessment (U.S. EPA, 2019a) contains additional information on exposure assessment for
population groups of concern, including Tribal and Indigenous populations.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
42
-------
contaminant (e.g., blood concentrations) and co-exposures to multiple stressors that may affect the
body's ability to detoxify a particular contaminant (e.g., metabolism) can be factors to consider.
Characterizing the distribution of relevant demographic or socioeconomic characteristics for the
affected population helps identify factors that may affect communities' ability to withstand or recover
from exposure to a stressor. Appendix B provides examples of integrating these characteristics into a
dose-response assessment.
5.3.3.2 Analysis Plan
The analysis plan provides details on technical aspects of the risk assessment and how the hypotheses
about the relationships described in the conceptual model will be assessed. While the conceptual
model may identify a larger set of pathways and relationships, the analysis plan focuses on the
pathways and relationships that will be pursued in the risk assessment. The plan includes the rationale
for selecting or omitting pathways, the relationships between stressors and outcomes, and
acknowledgements of data gaps and uncertainties (see Text Box 5.4.)
The analysis plan also may consider how the level of confidence (or precision) needed for the
management decision compares with that expected from available analytic approaches. This informs
how data are used, the preferred analytic approach, and the extent to which new data are needed and
may be obtained.
The analysis plan may be divided into specific components: (a) the assessment design and rationale for
selecting specific pathways to include in the risk assessment; (b) a description of the data, information,
methods, and models to be used to assess exposures, effects, and risk (including associated
uncertainty and variability), as well as intended outputs (e.g., risk metrics); (c) quality assurance and
quality control measures; and (d) the associated data gaps and limitations. The analysis plan may also
describe scientific review and specify actions for community involvement (U.S. EPA, 2014b).
A central challenge for HHRA planning is identifying the data, tools, and models needed to inform EJ
considerations. Data selection should be based on the context, risk management and analytic
objectives, and scope of the analysis. (Appendix B provides sample questions to help identify data and
model needs when planning for exposure assessment and dose-response assessment.)
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
43
-------
Text Box 5.4 Example: Considering Factors Related to EJ Concerns in Planning Integrated
Science Assessments
The Clean Air Act requires periodic review of the science upon which national ambient air quality
standards (NAAQS) are based. The EPA performs an Integrated Science Assessment (ISA) as the
scientific foundation for these NAAQS reviews by evaluating and synthesizing the latest science,
including the kind and extent of identifiable effects on public health. At the beginning of an ISA, the
public is invited to submit research studies in identified subject areas via a "call for information."
At its early stages, an ISA is also guided by an Integrated Review Plan (IRP) that identifies the key
science issues relevant to review of the criteria for setting the NAAQS. Among other things, an IRP
generally enumerates how the ISA will evaluate factors that may affect and inform the risk
assessment, including those relevant to EJ concerns. The IRP is produced as a public information
document and management tool and therefore may be modified to reflect information developed
during the review and in response to advice and comments received during peer and public review.
The Integrated Review Plan for the National Ambient Air Quality Standards for Particulate Matter
(PM) provides an example:
"The PM ISA will evaluate an array of factors that may contribute to increased risk of PM related
health effects for various lifestages or populations. These factors reflect the multiple avenues
through which a lifestage or population may be at increased risk of an air pollutant related health
effect, specifically: intrinsic factors (e.g., biological factors such as age or genetic variants), extrinsic
factors (nonbiological factors such as diet, low [socioeconomic status]), and/or factors affecting dose
or exposure (e.g., sex, age, outdoor activity or work, low [socioeconomic status], physical activity). It
is also important to recognize the interconnectedness among these factors that may also confer
increased risk, an example being pre-existing diseases or conditions and socioeconomic status...
...Further, to the extent that evidence is available, the at-risk chapter of the PM ISA will discuss what
evidence is available regarding interrelationships among risk factors in a particular lifestage or
population as described in the preceding section that may add to the understanding of the public
health impact of exposure to PM." (U.S. EPA, 2016c).
See Section 3.4.9, At-Risk Lifestages and Populations and Public Health Impact, for more
information: https://www3.epa.qov/ttn/naaqs/standards/pm/data/201612-final-inteqrated-review-plan.pdf.
Identifying Data
A key planning element for identifying data relevant to EJ concerns is consulting with the public,
including communities that may have access to data useful for improving the characterization of
exposure and risk. Relevant data can be location-specific or population group-specific, or, ideally, both.
It may also include ambient concentration data (e.g., from air monitoring stations and water quality
measures) or public health data such as disease incidence.
Exposure data may include information on consumption or contact rates, routes, and duration of
exposure, behavior data for estimating contact rates, concurrent exposures to other stressors that are
of toxicological relevance, biomonitoring, or emissions. There are many sources of exposure data.
Some exposures can be evaluated using bio-monitoring data on chemical hazards such as the National
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
44
-------
Health and Nutrition Examination Survey (NHANES), although NHANES provides exposure data for
limited geographic areas.67
Health risk data could include toxicological data as well as incidence data specific to population groups
with EJ concerns, and historical population-specific disease or illness rates.68 More generally, multiple
data sources may be considered for an HHRA to inform EJ concerns, including chemical-specific
epidemiological evidence and potentially the broader literature on intrinsic and extrinsic factors that
influence vulnerability. State, Tribal, and local governments may also have relevant monitoring data for
use in HHRA. Appendix B provides more detailed information on using bio-monitoring data and an
example of estimating exposure using ambient concentration data.
Identifying Models and Tools
Risk assessment employs a range of models and tools to estimate ambient concentrations of stressors,
exposure, amounts of stressors likely to reach the target organ(s) (e.g., effective dose), risks for a
specific health endpoint(s), locational vulnerability to health effects, and other key factors. A challenge
for developing a HHRA that can inform EJ considerations is ensuring that input parameters for models
are representative of population groups of concern. Traditional defaults used as inputs for HHRA may
not adequately reflect the demographic or socioeconomic characteristics of these population groups.
While data may be insufficient to resolve this issue, it can be useful to discuss the extent to which and
reasons why defaults may not be representative.
Identifying Data Quality Limitations and Data Gaps
Consideration of EJ concerns may be aided by rapidly developing data and tools; thus, it is important
that the HHRA planning process include a clear discussion of data available to characterize uncertainty
and variability, data quality, and lack of data that may affect methodology development and/or results.
In some cases, lack of data may prompt a decision to limit the scope of planned analysis within a
HHRA. It is recommended that such decisions be clearly documented, and where possible affected
communities be consulted because they can often provide input into how to proceed when there is a
lack of data. In some instances, clear documentation of lack of data may lead to changes to the
regulatory design to facilitate better monitoring in vulnerable communities.69
To further promote the integrity, objectivity, and quality of data used in planning risk assessments,
analysts should review the OMB and EPA data quality guidelines (OMB, 2019; U.S. EPA, 2012a; U.S.
67 Some limitations of data available through NHANES can be addressed by location-specific surveys (such as the New York
City Health and Nutrition Examination Survey), and other site- and population-specific surveys that may be conducted for
reasons other than EJ considerations. Some limitations to the availability of primary site- and population-specific surveys are
cost and the amount of time required to conduct these surveys.
68 Analysts are advised to assess historical population-specific disease or illness rates to understand whether these data are
built on potentially flawed assumptions or population selection that could exacerbate or negatively affect the accuracy of the
HHRA results. For example, some population groups are less likely to seek medical care due to lack of health insurance or
access to care.
69 For example, public comments during the nitrogen oxide (NOx) NAAQS rulemaking process resulted in siting additional
monitors in vulnerable communities (U.S. EPA, 2010a). Likewise, outreach to communities living near refineries during a risk
and technology review resulted in fenceline monitoring of benzene emissions to provide access to data on what is being
released into nearby communities (U.S. EPA, 2015c).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
45
-------
EPA, 2002b).70 Text Box 3.1 describes the factors the EPA uses to evaluate the quality and reliability of
scientific and technical information used in support of its actions.
5.4 Multiple Exposures and Cumulative Effects
The science supporting assessments of the cumulative effects of exposure to multiple stressors on
human health and well-being is evolving, and the data and analytic tools needed to develop richer and
more informative analyses may become available as research continues. Given the importance of
cumulative effects in the context of environmental justice, we briefly summarize cumulative risk and
cumulative impact assessment (CIA) approaches below. In the meantime, even when utilization of a
more formal approach to assessment is not feasible, analysts are advised to consider the potential
implications of exposure to multiple stressors, both environmental and non-environmental, when
planning and scoping for a HHRA.
5.4.1 Cumulative Risk Assessment
HHRA often focuses on characterizing risk from a single stressor or contaminant. Recognizing the
potential harm associated with multiple stressors from one or more pollution sources or exposure
pathways, the EPA has described a framework for assessing the cumulative risk of adverse effects
associated with multiple stressors (U.S. EPA, 2003a). Cumulative risk assessment (CRA) is an
analysis, characterization, and possible quantification of the combined risks to health and/or the
environment from aggregate exposure to multiple agents and/or stressors. Specific elements and
implementation of CRA may differ according to programmatic needs. Because of data and methodology
limitations, applications of CRA at the EPA have mainly focused on chemical mixtures and/or single
chemicals from multiple sources.71
Conducting a CRA is possible when there is adequate information to identify the relevant risk
relationships among multiple environmental and non-environmental stressors, as well as the means to
identify those most important to characterizing relevant health endpoints. This approach is consistent
with the existing paradigm underlying risk assessment generally, which relies upon scientifically
rigorous methods for assessing hazard and exposure and the resulting adverse outcomes. See
Appendix A for links to other U.S. EPA guidance on the conduct of CRA.
As with HHRA, understanding the purpose of CRA is fundamental to its design. Clarity on the
management decision context, and what information is necessary to support it, as well as data quality,
are the first steps in determining if a CRA may be suitable, and if so, determining an assessment
design. This step is described as fit-for-purpose and is common to all assessment processes. See prior
sections of this chapter and U.S. EPA (2014b) for more information.
In planning a CRA, attention may be given to epidemiology studies that can indicate multiple chemical
exposures and other factors such as non-environmental stressors that may modify or increase the risk
of an adverse outcome from the target contaminant. It may be useful to use epidemiological data to
70 For information, visit the EPA's Information Quality Guidelines website
-------
focus on diseases or health conditions with a higher prevalence within or across population groups of
concern. Studies that employ stratification can provide insight into how co-exposure to additional
environmental or non-environmental stressors may affect the risk of an adverse health outcome for a
given population group of concern.
Incorporating non-environmental stressors in CRA can be complicated by uncertainties in exposure
evaluation, potential health consequences, potential modification of effects by other stressors (e.g.,
chemical exposures) and variation in individual (or community scaled) vulnerabilities. When the
suspected stressor estimates of risk are difficult to quantify, a qualitative or semi-quantitative
characterization of suspected relationships (e.g., an empirical correlation between the stressor and
adverse outcomes or the existence of a notably large number of potential stressors) can improve
understanding of the range of potential stressors contributing to adverse health outcomes. See U.S.
EPA (2003a) for more information on considering uncertainty and variability in CRA.
5.4.2 Cumulative Impact Assessment
Cumulative impact assessment is the process of accounting for cumulative impacts in the context of
problem identification and decision-making. The term cumulative impacts refers to the totality of
exposures to combinations of environmental and non-environmental stressors and their effects on
health, well-being, and quality of life outcome throughout a person's lifetime (U.S. EPA, 2022c).
There are linkages between CRA and CIA, both may address both environmental and non-
environmental stressors, but these analyses are not the same and may serve different decision
contexts; more than one approach may also be used in combination. Each approach may employ a
variety of methods to provide information to decision-makers about health effects associated with
exposures to multiple environmental and non-environmental stressors. Methods employed are not
exclusive to an assessment approach. Determinations about which approaches and analytic methods
to use depend on factors such as statutory requirements, the scope of an assessment, types of data
needed and available, and applicability for the evaluation and needs of the decision-maker (U.S. EPA,
2024d). For example, total burden in CIA encompasses direct health effects, but also considers a wide
set of outcomes that fall outside the purview of CRA, some of which may not be conducive to
quantification (NRC, 2009). CIA may be able to "use information supported by relationships among
stressors, exposures, effects, and/or health, well-being, and quality of life outcomes for which cause-
and-effect linkages may not be well understood" (U.S. EPA, 2022c).
The EPA does not currently have guidance on the use of CIA in the context of rulemaking. However,
the EPA has released a report on recommendations for CIA research (U.S. EPA, 2022c) and an interim
cumulative impacts framework document (U.S. EPA, 2024h). The EPA has also used a community-
engaged approach to assessing cumulative impacts called Health Impact Assessment (HIA) in other
decision contexts (See Text Box 5.5).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
47
-------
Text Box 5.5 Health Impact Assessment
Health impact assessment (HiA) is "a systematic process that uses an array of data sources and
analytic methods and considers input from [affected individuals, communities, and other members of
the public] to determine the potential effects of a proposed policy, plan, program, or project on the
health of a population and the distribution of those effects within the population" (NRC, 2011a).
Given the broad purview of this approach, HIAs may consider a wider spectrum of health
determinants than a typical HHRA, such as housing quality, access to services, and social cohesion,
as well as exposure to contaminants or the effects of climate change.
The HIA process typically emphasizes meaningful public engagement that focuses on empowering
vulnerable and affected populations to participate in decisions that have the potential to affect their
daily lives. Effective input from the public can provide local knowledge of existing conditions; identify
areas of concern and issues of interest that are not readily apparent to those outside the community;
offer contextual/cultural perceptions and experiences; and assist in identifying and refining the HIA
scope and recommendations.
HIA is most often applied in the context of a specific local community. However, while the EPA has
not used HIA in support of national regulatory actions, it could serve as a complement to HHRA in
the national context, for instance to evaluate cumulative effects and EJ concerns related to hot
spots. The figure below illustrates how an HIA can be tailored to available time and resources.
DESKTOP
RAPID
INTERMEDIATE
COMPREHENSIVE
2-6 weeks for one
person full time.
6 to 12 weeks for one
person full time.
12 weeks to 6 months
for one person full
time.
6 to 12 months for one
person full time.
Involves an 'off the
shelf' exercise
analyzing existing,
accessible data.
Involves collecting and
analyzing existing data
with limited input from
stakeholders.
Involves collecting and
analyzing existing and
new data, including
input from
stakeholders
Involves collecting and
analyzing existing and
new data, including
input from
stakeholders
Provides a broad
overview of potential
health impacts.
Provides a more
detailed overview of
potential health
impacts.
Provides a more
thorough assessment
of potential health
impacts, and more
detail on specific
predicted impacts.
Provides a
comprehensive and
detailed assessment of
potential health
impacts.
Applied when time and
resources are limited.
Applied when time and
resources are limited.
Requires moderate
time and resources.
Requires significant
time and resources.
.
LLJJ I IVIrHL 1 J
Case studies to explore how HIA can be used to engage the public and incorporate EJ concerns and
public health considerations into local environmental decision-making are available at:
https://www.epa.gov/healthresearch/epa-health-impact-assessment-case-studies. The EPA also has
compiled an inventory of HIA resources at: https://www.epa.gov/healthresearch/health-impact-
assessment-hia-resource-and-tool-compilation and synthesized the state of practice at:
https://www.epa.gov/healthresearch/hia-review-svnthesis-report.
Technical Guidance for Assessing Environmentai Justice in Regulatory Analysis - Second Edition
48
-------
Chapter 6: Conducting Regulatory Analyses
to Assess Environmental Justice Concerns
This chapter discusses how to conduct an EJ analysis for a rulemaking.72 In particular, it discusses
methods that may be useful for answering the three analytic questions from Section 3.1 of this
document, which are repeated here:
• Baseline: Are there existing EJ concerns associated with environmental stressors affected by
the regulatory action for population groups of concern?
• Regulatory options: For the regulatory option(s) under consideration, are there potential EJ
concerns associated with environmental stressors that are affected by the regulatory action for
population groups of concern?
• Mitigation or exacerbation of effects: For the regulatory option(s) under consideration, are EJ
concerns exacerbated, mitigated, or unchanged compared to the baseline?
These questions provide the framework for analyzing the effects of a regulatory action on population
groups of concern. The extent to which an analysis can address all three questions will vary due to data
limitations, time and resource constraints, and other technical challenges. These challenges will vary by
media and regulatory context, including the availability of information generated from human health risk
and exposure assessments, or other components of the regulatory analysis. Regardless of the
approach taken, the highest quality and most relevant data should be applied in a manner consistent
with the OMB and EPA data quality guidelines (OMB, 2019; U.S. EPA, 2012a; U.S. EPA, 2002b) and
the Peer Review Handbook (U.S. EPA, 2015e).
In determining whether EJ concerns may arise in the context of a specific regulatory action, some level
of analysis is needed, be it qualitative, quantitative, or some combination of both. For many regulatory
actions, including actions that strengthen environmental protection, it is not possible to rule out EJ
concerns without some level of assessment.
Generally, the EPA prefers quantitative assessments that complement other types of quantitative
regulatory analyses (e.g., benefit-cost analysis, risk assessment) conducted for regulatory actions.
Section 3.3 recommends some level of quantitative analysis, when feasible, to address the three
questions above for regulatory actions where effects or benefits will also be quantified. When
information on exposures, health and environmental outcomes, and other relevant effects by population
groups is available, analysts may be able to quantify exposure in the baseline and likely changes in
exposure for each regulatory option. In cases where such data are unavailable, it may still be possible
to evaluate risk or exposure using other quantitative metrics (e.g., density of regulated facilities as a
function of race, ethnicity, or income).
72 While the focus in this chapter is on population groups mentioned in E.O.s 12898 and 14096, the methods discussed may
be applied to any population group of concern.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
49
-------
When environmental and health effects or benefits cannot be quantified or disaggregated by race,
ethnicity, income, or other relevant demographic or socioeconomic factors, analysts should present
available quantitative and/or qualitative information that sheds light on the concerns and experiences of
communities or population groups relevant to the EJ analysis. Qualitative assessment is particularly
appropriate when high quality and relevant quantitative data are not available.
This chapter is organized as follows: Section 6.1 discusses how a preliminary analysis can be used to
evaluate the feasibility of an in-depth analysis of EJ concerns. Section 6.2 discusses the conceptual
approach to evaluating and comparing the baseline and regulatory options in an in-depth analysis of EJ
concerns. Section 6.3 reviews the data and information needed to assess EJ concerns. Section 6.4
summarizes methods for assessing the three analytic questions for population groups of concern.
Section 6.5 discusses analytic issues, including comparison population groups and geographic issues
when the source of emissions is identifiable and health effects are fairly localized and spatially
distinguishable. Sections 6.6 discusses characterizing and communicating results. Section 6.7
discusses the evaluation of costs and non-health effects. For examples, see the companion document
summarizing several EJ analyses conducted for rulemakings (U.S. EPA, 2024i).
6.1 Preliminary Analysis of EJ Concerns
As discussed in Section 3.2, conducting a preliminary analysis early in the rulemaking process can offer
initial insights that may merit additional investigation regarding whether there are EJ concerns in the
baseline and/or a regulatory action is anticipated to raise EJ concerns.
While there is no single prescribed method for conducting a preliminary analysis, analysts should
review the quality and availability of data, the availability of defensible methods to analyze the data, and
the peer-reviewed literature and public input that may be used to evaluate EJ concerns. In cases where
it has already been determined that a more in-depth EJ analysis is needed, these steps can be directly
incorporated into the early stages of that analysis. Such information may include:
• Proximity of regulated sources near population groups of concern;
• Number and types of sources that may be impacting these populations;
• Types of stressor(s) that may be affecting these populations in the baseline, including from non-
regulated sources;
• Any unique exposure pathways associated with the stressor(s) being regulated;
• Evidence of differential current or historical exposure to or risk from the stressor(s) being
regulated for population groups of concern;
• Evidence of the prevalence of factors or conditions that may modify the effect of the regulated
stressor(s) for population groups of concern;
• Public or community concern(s) about the effects of the potential regulatory action on specific
population groups or communities of concern; and
• History of EJ concerns associated with the stressor(s) or sources(s) being regulated.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
50
-------
A variety of tools and methods are available to support a preliminary analysis.73 For instance, EJScreen
(U.S. EPA, 2024c) allows analysts to quickly examine a variety of demographic and socioeconomic
variables and environmental indicators at the block group resolution for the United States (including
Puerto Rico).74 Demographic and socioeconomic data are drawn from the American Community Survey
five-year summary file.
While EJScreen provides a way to identify areas in the United States where population groups or
communities of concern or environmental indicators are currently at or above a specific relative
percentile, analysts are advised to summarize the demographic and socioeconomic characteristics and
relevant environmental indicators for all communities near affected sources in level terms (e.g., percent
of the population that is low-income), not just for communities with indicator values above a specific
percentile threshold. In addition, it is recommended that analysts avoid using indices that combine
information across multiple variables or indicators, which are hard to interpret, and instead present the
results for each relevant variable separately.75 It is also important to keep in mind that while EJScreen
and other screening tools can shed light on pre-existing differences for the environmental indicators
included in the tools, they typically only provide a snapshot at one point in time and may not include
other important sources of exposure relevant to the regulatory action.
More generally, when evaluating whether a screening tool is sufficient or appropriate for conducting a
preliminary analysis of EJ concerns in the federal rulemaking context, analysts should consider the
purpose for which the tool was originally designed. Certain tools are designed to evaluate
environmental conditions within a certain region or state and, as such, may not contain nationally
representative data. Other tools may be designed to help prioritize funding, permitting, or compliance
and enforcement actions and therefore may define population groups or other key variables in specific
ways, or they may incorporate policy decisions directly into the tool by weighting certain indicators or
characteristics over others.
A preliminary analysis can also act as an input into determining whether it is feasible to conduct an in-
depth assessment. Feasibility is informed by a technical evaluation of available data and methods,
including:76
• Scientific literature that discusses the effects of the stressor(s) being regulated on population
groups of concern;
• Information received via public comments, technical reports, press releases, or other
documentation discussing the environmental and health effects of the stressor(s) being
regulated for population groups of concern, including information on other relevant
environmental or non-environmental stressors;
73 Envirofacts contains information on the location, reported emissions, and compliance history of sources regulated by the
U.S. EPA under various statutes (https://enviro.epa.gov/). EasyRSEI allows analysts to examine the types and amounts of
toxic chemicals reported annually by facilities to the Toxics Release Inventory
(https://edap.epa.aov/public/extensions/EasvRSEI/EasvRSEI.htmn.
74 See the EPA's EJScreen website: www.epa.gov/eiscreen. Other agencies also have EJ screening tools that may prove
useful. For instance, the Department of Health and Human Services has developed an index-based tool at the census tract
level to evaluate the cumulative impacts of environmental burden on human health. See
https://onemap.cdc.goV/portal/apps/sites/#/eii-explorer.
75 For instance, while EJScreen tabulates and reports an aggregate race/ethnicity category, labeled as "people of color",
analysts are encouraged to report results separately for each relevant race and ethnicity category (see Section 2.2.1).
76 Recall that appropriateness is informed by relevant policy, budgetary, and statutory considerations (Section 3.2).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
51
-------
• Availability of spatially disaggregated data for population groups that may live, work, or play in
close proximity to the stressor(s) being regulated, or may otherwise be affected by the
stressor(s); or
• Availability of methods for conducting in-depth analysis (e.g., proximity-based approach, risk- or
exposure-assessment, and mixed methods approach, as discussed below).
If the preliminary analysis reveals that the scientific literature and data are unavailable or of insufficient
quality to pursue an in-depth analysis that characterizes how exposure, risk, or health effects are
distributed across population groups, analyst are expected to explain why additional analysis is not
possible. In particular, analysts are encouraged to discuss relevant evidence, key limitations, and
sources of uncertainty highlighted in the published literature (U.S. EPA, 2024f). Some impacts that
cannot be quantified may still represent important effects that should be considered in the analysis.
6.2 Characterizing the Baseline and Regulatory Options for In-Depth EJ
Analysis
As described in Section 3.2, the five main steps of an in-depth EJ analysis are: (1) identify the sources
being regulated; (2) describe the environmental stressor; (3) characterize affected populations: (4)
compare the affected and comparison groups; and (5) conduct sensitivity analysis. These steps can be
applied to characterize baseline conditions, evaluate the effects of the regulatory options, and make
comparisons between the two to evaluate the three analytic questions posed in Section 3.1, when data
and methods allow.
The OMB (2023b) defines the baseline as "an analytically reasonable forecast of the way the world
would look absent the regulatory action being assessed, including any expected changes to current
conditions over time." It includes the characteristics of current populations and how they are affected by
pollutant(s) prior to the regulatory action under consideration. As the OMB definition implies, however,
the baseline is not a static concept. In particular, the OMB notes that analysts may need to consider the
evolution of the market, compliance with other regulations, and the future effect of current government
programs and policies, as well as other relevant external factors to project future baseline conditions.
How future regulations or policies affect the baseline specification is complex and requires
consideration of many factors. As noted in the EPA's Guidelines for Preparing Economic Analyses
(U.S. EPA, 2024f), ideally all potential influences on baseline conditions would be estimated, but it is
generally not practicable to do so.77 Anticipated changes in baseline demographic composition may
also be relevant in an EJ context. Per the recommendations in Section 3.3, the baseline for the EJ
analysis, including the geographic scope, year of analysis, and health and other effects, should be
consistent with how it is specified in other parts of the regulatory analysis.
Because exposure to other environmental and non-environmental stressors can increase an
individual's responsiveness to negative health effects, it is also important to understand how already-
present risks may interact with the pollutant being regulated (see Chapter 4). Risk and exposure
assessments conducted in support of the regulatory action can offer a valuable source of information in
this regard (see Chapter 5 and Appendix B). While explicit modeling of these interactions is often not
feasible, analysts can evaluate the extent to which there are multiple polluting sources or elevated risks
77 See Section 5.5 of U.S. EPA (2024f) for more detail.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
52
-------
for other key stressors within affected communities, including historic and current social or health
conditions, that affect specific population groups or communities. See Text Box 6.1 for an example.
When data and methods allow, an EJ analysis can also examine the distribution of effects for each
regulatory option - different configurations of the regulatory action being considered. This analysis is
based on a prediction of the state of the world under the regulatory options. For the analysis of EJ
concerns, analysts are encouraged to examine how the exposure, risk of health or environmental
effects, or other outcomes of the regulatory action are distributed across population groupsfor the
regulatory options being considered, where practicable.78
The EJ analysis can then evaluate the change in the exposure or risk of relevant environmental and
health effects for each regulatory option compared to the baseline. In addition to identifying whether the
regulatory action is expected to exacerbate, mitigate, or leave baseline EJ concerns unchanged, the
analysis should shed light on the extent and distribution of these changes.
With these three sets of information - effects in the baseline, effects under the regulatory options, and a
comparison of the two -analysts can characterize the distribution of environmental and health effects
associated with a regulatory action, thus answering all three analytic questions from Section 3.1.
Specifically,
• An assessment of the baseline can inform whether pre-existing differences in environmental
and health effects are associated with the stressor(s) under consideration.
This analysis depicts how the stressor(s) and associated effect(s) are distributed across population
groups prior to any regulatory action. For instance, if emissions or effects are more concentrated in one
population group (e.g., low-income households), the decision-maker may want to take this into
consideration when making decisions about the regulatory action; mechanisms or choices associated
with implementation, for example, may allow the EPA to address pre-existing differences.
• An assessment of the regulatory options being considered can inform how the stressor(s) and
its environmental and health effects are distributed.
It is important to note that analysis of the regulatory options is based on predictions, which may not
always be sufficiently disaggregated across population groups to enable a rigorous EJ analysis. Ideally,
analysts can provide an indication of how the stressor is distributed across population groups for the
options being considered, either quantitatively or qualitatively. There may be some options for which
the distribution of the stressor and its effects across population groups is more equitable than others.
78 The regulatory options in the EJ analysis should be the same as in the other parts of the regulatory analyses (e.g., benefit-
cost analysis) to facilitate comparisons and ensure consistency. Typically, multiple scenarios or options are considered in a
regulatory analysis (OMB, 2023b).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
53
-------
Text Box 6.1 Examples of the Role of Non-Environmental Stressors in EJ Analysis
The Reconsideration of the National Ambient Air Quality Standards for Particulate Matter
(U.S. EPA, 2024m) includes the following discussion of non-environmental stressors:
"EPA has expanded the populations evaluated in this exposure EJ assessment... to evaluate
potential exposure disparities more comprehensively, regardless of any pre-existing biases. Several
populations added to this final action have been included in other rulemaking [regulatory impact
assessments] RIAs evaluating PM2.5 emission changes (e.g., employment status, health insurance
status, and linguistic isolation). Additionally, this RIA assesses exposure in communities with a
legacy of discriminatory land use designations and siting decisions (i.e., historically redlined areas,
using the Home Owners' Loan Corporation (HOLC) gradings A-D, with grade D areas being defined
as redlined). EPA believes all population groups added to this exposure assessment provide
additional insight into community-level vulnerability."
The Multi-Pollutant Emissions Standards for Model Years 2027 and Later Light-Duty and Medium-
Duty Vehicles (U.S. EPA, 2024n) also discusses the role of non-environmental stressors:
"As described [previously], people who live or attend school near major roadways are more likely to
be people of color and/or have a low [socioeconomic status] SES. Additionally, people with low SES
often live in neighborhoods with multiple stressors and health risk factors, including reduced health
insurance coverage rates, higher smoking and drug use rates, limited access to fresh food, visible
neighborhood violence, and elevated rates of obesity and some diseases such as asthma, diabetes,
and ischemic heart disease. Although questions remain, several studies find stronger associations
between air pollution and health in locations with such chronic neighborhood stress, suggesting that
populations in these areas may be more susceptible to the effects of air pollution."
• An assessment of the changes in exposure or risk across population groups - when the
regulatory options are compared to the baseline - can help inform to what extent the regulatory
action will address identified EJ concerns.
It is helpful for analysts to provide information on how effects under each regulatory option compare to
the baseline to show the extent to which an option improves or degrades environmental quality and
health risk across population groups of concern.
Note that a constant reduction in risk or exposure across population groups will likely not mitigate EJ
concerns if there are differences in baseline environmental quality or health risk across population
groups or communities (Maguire and Sheriff, 2011). Conceptually, an EJ concern is only completely
mitigated when there is no difference in the distribution of effects across population groups for the
regulatory options being considered - i.e., everyone is experiencing the same environmental quality or
health risk post-regulation.
6.3 Data and Information to Assess EJ Concerns
In general, the type of analysis that can be conducted depends on the availability and quality of data. In
some cases, spatially resolved, individual-level data may be most appropriate and relevant for an
analysis of EJ concerns. In other cases, distance from a regulated source may be the best available
metric. At times, the best available information may be qualitative, including local knowledge from
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
54
-------
affected communities and Tribes (e.g., Indigenous Knowledge, also referred to as Traditional Ecological
Knowledge). In all cases, analysts should use the highest quality and most relevant data and
information, as discussed below. Text Box 6.2 illustrates how data quality may affect the level of
analysis in an air quality context.
When data are missing or incomplete it is recommended that analysts document what specific types of
data are unavailable or of insufficient quality, including but not limited to cases where the data are
available but not of the desired granularity (spatially or temporally) and/or available for only subsets of
the population.
Recognizing the importance of data quality, data needed to conduct an EJ analysis may include:
• Demographic and socioeconomic characteristics (e.g., race, ethnicity, income);
• Location of pollution sources (e.g., latitude/longitude coordinates, zip code, county);
• Historical, current, and projected emissions or concentrations of stressor(s) relevant to the
regulatory action;
• Prevalence of specific exposure pathways that may increase risk for some population groups;
• Health effects (e.g., hospital and emergency admissions, race and ethnicity-stratified mortality
rates, race and ethnicity-stratified asthma or other morbidity rates);
• Other environmental or non-environmental stressors that may be risk- or effect modifiers (e.g.,
indoor air concentrations, vulnerability to effects of climate change);
• Risk coefficients stratified by population groups of concern (e.g., race, ethnicity, income); and
• Distribution of economic costs, when relevant (see Section 6.7.1).
We briefly describe three types of information frequently used as inputs into an EJ analysis:
demographic and socioeconomic data, emission data for regulated sources, and data on pre-existing
health conditions or other intrinsic or extrinsic factors that may increase an individual's vulnerability
when exposed to releases from regulated sources.
The U.S. Census Bureau is the recommended source for demographic and socioeconomic data in an
EJ analysis. It produces several national-level data products that report these types of data at relatively
fine spatial scales (e.g., census tract, block group), including the decennial Census, the American
Community Survey (ACS), and the American Housing Survey (AHS).
The decennial Census is administered every ten years with the aim of including the total U.S.
population (including U.S. territories). As such, overall margins of error are minimal. Data on race,
ethnicity, age, sex, and owner/renter status and other characteristics are available at the census tract
level. For privacy reasons, statistical noise is deliberately introduced into publicly released Census
results. This has led to larger measurement errors for low-population geographies (Kenny et al., 2024).
Due to these privacy measures and other issues, there is evidence that the Census systematically
undercounts mixed-race populations, Native American individuals, and undocumented immigrants
(Warren, 2022; Kenny et al., 2021; Connolly and Jacobs, 2020).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
55
-------
Text Box 6.2 Data Quality and Spatial Resolution in the Context of Air Regulations
Confidence in analysis (darker = higher level of confidence)
t t
t
t
Examples of Emission changes Source-level
data used: by source type,- dispersion
tract-level modeling,- county-
demographic level baseline health
data data,- block group
demographic data
Regional photochemical
air quality modeling,-
county-level health
data,- PM2.5 and ozone
effect coefficients
stratified by
education/race
Local-scale air quality
modeling,- tract-level
health data,- city-level
effect coefficients,-
projected incidence
rates and demographic
data
Analysts' ability to address how a regulatory action changes the distribution of risk across population
groups depends on the quality and spatial resolution of the data available. Finer-scale air quality,
health, and demographic data allow one to assess the distribution of effects across population
groups and to have greater confidence in the conclusions drawn from these data. When air quality
data are lacking or only available at a coarse level, the ability to assess change in risk across
populations and other conclusions is more limited.
An example in limited data environments: Using race-stratified county-level mortality and morbidity
data, analysts can calculate population-weighted mortality rates by county. Analysts can then use a
highly aggregated baseline air quality modeling projection (e.g., 12 or 36 km) to identify population
groups most exposed to air pollution. Using geospatial tools, it is possible to combine the two
sources of data. The coarse geographic scale of air quality information may inhibit analysts' ability to
detect meaningful differences in effects among and between groups. When risk coefficients are
unavailable, it is not possible to estimate health effects separately for each population group.
An example in data-rich environments: Using finely resolved air quality data, analysts can identify at
a highly disaggregated level (e.g., 1 km) population groups that experience the highest exposure to
air pollution. Analysts can also identify population groups that exhibit the highest baseline incidence
or prevalence rates for air pollution health effects. Using geospatial tools, analysts can spatially
combine the two data sources. Using race-specific or standard risk coefficients analysts can then
estimate health effects for each population group.
Compared to the decennial Census, the ACS contains information on a wider range of demographic
and socioeconomic characteristics, such as income, education, disability status, and employment, at a
greater level of spatial disaggregation - the block group level. Further, the ACS relies on annual
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition 56
-------
population sampling and therefore reflects more timely information, especially for rapidly changing
geographic areas. The five-year ACS is preferable to the one-year ACS because the five-year
estimates contain information on all areas (including those with fewer than 65,000 individuals) and has
estimates of population statistics with lower margins of error. Using more reliable estimates may be
important when investigating specific locations or areas with low population.
Even five-year ACS estimates for low population areas are likely based on smaller samples and have
larger margins of error and reduced accuracy.79 Strategies to improve estimates in areas with low
population include combining data across geographic units or population groups. The U.S. Census
Bureau suppresses estimates with unacceptable levels of statistical reliability, but analysts still need to
carefully consider data quality, especially in rural areas.80
The American Housing Survey is a longitudinal, nationally representative biennial survey at the
housing-unit level that contains information on the quality and cost of housing and resident
characteristics, including demographics. The survey includes owner- and renter-occupied and vacant
housing units but excludes group quarters, businesses, hotels, and motels.81 While national data are
collected in each year in which the AHS is conducted, housing units in major metropolitan areas and
those receiving assistance from specific federal programs are oversampled. Weights are applied to
each housing unit in the national survey to minimize sampling error.82 Statistical techniques, such as
the introduction of noise, are used in the AHS to protect privacy and confidentiality concerns.83
The demographic and socioeconomic data described above are based on where people live. However,
exposure can also occur when individuals are in other locations (e.g., work, school). For this reason,
analysts may want to consider using non-residential demographic and socioeconomic data, such as
spatially resolved data on workers (see Text Box 6.3) and school-aged children.84 85 86
79 Note that the ACS and AHS both use 90% confidence intervals for margins of error.
80 For more information on compensating for low population in rural areas, see Section 2 of Understanding and Using
American Community Survey Data: What Users of Data for Rural Areas Need to Know (U.S. Census Bureau, 2020a) at:
www.census.aov/proarams-survevs/acs/librarv/handbooks/rural.html.
81 It may be possible to incorporate geospatial data on populations omitted from or underrepresented in Census data. For
example, the Homeland Infrastructure Foundation-Level Data contains geospatial prison boundaries on local, state and federal
correctional facilities, and on nursing and assisted care facilities in the United States.
82 The U.S. Census Bureau releases information on sample design, weighting and accuracy of the survey in each year the
AHS is administered. See https://www.census.aov/proarams-survevs/ahs/tech-documentation/def-errors-
changes.2021.html#list-tab-427877428. For information on measurement and sampling error, see
https://www.census.aov/content/dam/Census/proarams-survevs/ahs/publications/h12195-1.pdf
83 For more information on statistical techniques used in the AHS to protect privacy and confidentiality, see
https://www.census.qov/content/dam/Census/proqrams-survevs/ahs/tech-documentation/2015-
later/Disclosure%20Avoidance%202015%20and%20Bevond.pdf
84 Another possible source of data on worker characteristics is the U.S. Census Transportation Planning Products. These data
offer aggregate (but separate) counts of workers employed by two-digit industrial classification and earnings level, two-digit
industrial classification and occupation, race/ethnicity, poverty status, sex, and education level at the Public Use Microdata
Area (PUMA) and, in some cases, county level. See Elliott and Smiley (2019) for an example.
85 For example, Yoo et al. (2023) use the Longitudinal Employer-Households Dynamics Origin-Destination Employment
Statistics to examine disparities in workplace exposure to air pollution by race and ethnicity.
86 The Education Demographic and Geographic Estimates (EDGE) combines national-level data from the National Center for
Education Statistics and the ACS to create socioeconomic and housing indicators for school-aged children and their parents. It
includes spatial information such as school district boundaries, point locations for schools, and rural/urban school classification.
The Bureau of Indian Education (BIE) also produces geospatial data for all schools and education centers overseen by the BIE.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
57
-------
Text Box 6.3 Data for Characterizing Worker Characteristics
Publicly available data linking worker characteristics to specific sources are not generally available.
Nor is this information available at a fine spatial resolution and disaggregated industry classification.
The table below highlights several sources of data on worker characteristics analysts may consider,
along with their main advantages and disadvantages.
Data product
American Community Survey
microdata (U.S. Census
Bureau)
Advantages
Six-digit industry classification
Worker occupation information
Includes household income, race,
ethnicity, sex
Spatial resolution may be finer in urban
areas
Disadvantages
• Spatial resolution may contain multiple
counties outside urban areas
• Small sample sizes at intersection of
spatial unit and industry classification
Longitudinal Employer-
Household Dynamics Origin-
Destination Employment
Statistics, Workplace Area
Characteristics) (U.S. Census
Bureau)
• Very fine spatial resolution (block)
• Aggregate (but separate) annual counts
of workers employed by earnings level,
two-digit industrial classification, sex,
race, ethnicity, and education level
Industrial classification may not be
sufficiently granular in highly
concentrated industrial areas.
No worker occupation information
Statistical modelling necessary to make
claims about worker characteristics or
earnings at two-digit industrial
classification
Quarterly Workforce Indicators
(U.S. Census Bureau)
• Four-digit industry classification
• Quarterly census of private employment
covering 95% of private sector jobs
• Includes race, ethnicity, sex, and age
• County spatial resolution
• No information on occupation/job type
• No information on worker income
• Data suppression common at intersection
of spatial unit, industry classification, and
worker characteristic of interest
EEO-1 Statistics (Equal
Employment Opportunity
Commission)
• National census of private employers with
>50 or >100 workers
• Includes race, ethnicity, and sex
• Statistics for 10 distinct job types
• County spatial resolution
• Three-digit industry classification
• No information on worker income
• Data suppression somewhat common
within spatial unit, industry classification,
and worker characteristic of interest
The EPA can collect worker information from regulated sources under some statutory authorities.
For example, under section 4 of Toxic Substances Control Act, the EPA may issue test orders to
develop information about occupational exposure to a specific chemical substance, chemical
mixture, or categories of chemical substances/mixtures. A carefully crafted test order may be able to
collect exposure information at a facility differentiated by job type and length of service as well as
worker characteristics. Industry associations and trade groups also sometimes publish reports that
provide an overview of worker characteristics within a particular industry based on proprietary
surveys of their members. Access to more granular U.S. Census and Equal Employment
Opportunity Commission data is also a promising avenue to pursue to allow analysts to leverage
more granular information on worker characteristics (Ash and Boyce, 2018; Walker, 2013).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
58
-------
On the emissions side, the EPA collects source-level data in some contexts, which can be used as an
input into fate and transport modeling to project concentrations (see Section 6.4.2).87 Analysts may also
consider nationally representative data from other sources, such as rule-specific data collection
requests to affected sources, datasets available from other Federal agencies, or proprietary datasets.
When nationally representative data are not available, state, local, or Tribal data may support an
illustrative analysis or case study approach. Modeled data - for instance, remote sensing data that
have been combined with a process-based or machine learning model to estimate concentrations -
may also generate spatially- and temporally-resolved information.88 When using geospatial information
on pollution sources, analysts need to be mindful of the limitations associated with how the data are
represented, such as if multi-dimensional sources are represented as points rather than polygons or
how non-point source pollution is characterized.
Absent the ability to explicitly model differences in vulnerability across population groups, it is also
important to gather data on pre-existing conditions or non-chemical stressors that the literature have
shown may contribute to differentiated health risks when individuals are exposed to the pollutant being
regulated.89 Note, however, that data on disease prevalence or consumption patterns differentiated by
demographic or socioeconomic characteristics may be incomplete, only available at a more aggregate
spatial scale, or subject to substantial measurement error.
Both the EPA and the OMB have data quality guidelines that should be followed when evaluating
whether data and information are of sufficient quality for use in regulatory analyses of EJ concerns (see
OMB, 2019; U.S. EPA, 2012a; U.S. EPA, 2002b). When evaluating data quality, it is recommended that
analysts consider whether the methods used to generate or collect the data have been independently
validated and whether the data have been used in peer reviewed studies (U.S. EPA, 2003b). Text Box
3.1 describes the five factors the EPA uses to evaluate the quality and reliability of scientific and
technical information used in support of its actions.
Finally, an EJ analysis should clearly describe all important data sources and references. Unless the
data are confidential business information or some other form of private data, they should be made
publicly available. Providing documentation and access to the data used in an analysis is crucial to the
credibility and reproducibility of the analysis.
6.4 Analytic Methods
A variety of scientifically defensible methods can be used to assess EJ concerns associated with
regulatory actions. The choice of analytic method is most often driven by data availability. Analysts may
also rely on a combination of these methods when analyzing a regulatory action. The conclusions that
can be drawn from the analysis will vary depending on the method used.
87 For instance, the National Emissions Inventory is released every three years based on air emissions data provided by state,
local, and Tribal governments and supplemented by data developed by the EPA. See https://www.epa.gov/air-emissions-
inventories/national-emissions-inventorv-nei for more information.
88 For instance, these types of approaches have been used to estimate ambient air concentration levels (Di et al., 2019, 2016)
and streamflows across the U.S. (Feng et al., 2022).
89 For example, the U.S Census Bureau administers the National Health Interview Survey. The survey collects demographic
and socioeconomic variables from a nationally representative set of households, as well as information on pre-existing
conditions such data on asthma, diabetes, disabilities, mental health status, health insurance access, and food security. For
more information, see https://www.census.qov/proqrams-survevs/nhis.html.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
59
-------
Considerable uncertainty may exist about key relationships and health outcomes, such as how a
reduction in emissions or other types of releases from a given source translates into ambient
environmental quality and how it, in turn, translates into the human health effects of interest. This is
particularly problematic if uncertainties differ across population groups. For instance, if an overexposed
population group is more responsive to exposure (i.e., individuals in the group experience greater
adverse health effects per unit of exposure), then using exposure alone as a proxy will underestimate
the health risk posed by a stressor to that group. On the other hand, if proximity to a pollutant source
does not correlate with exposure, it could overstate potential differences in health risk. Analysts should
select the method that is most appropriate for the available data, recognizing time and resource
constraints. The sections below discuss three methods that are often used for assessing EJ concerns:
proximity-based analysis, use of exposure and risk modeling tools, and qualitative approaches. Note
that these are not the only possible methods that can be used to assess EJ concerns. For each of the
three approaches discussed, we highlight key advantages and limitations.
Regardless of the analytic approach used, the EJ analysis should be presented in a transparent way
and include the following:
• Information about the specific population groups and individuals affected by the regulatory
action;
• Main exposure pathways and expected health and environmental outcomes;
• Evidence for why risk, exposure, or outcomes may vary by population group, including the role
of other relevant environmental and non-environmental stressors;
• Relevant geographic scale;
• Descriptions of the main methods of analysis used;
• Descriptions of key data or modeling assumptions;
• Summary statistics for the baseline and each regulatory option (both the mean and distribution)
by population group;
• An easy-to-understand description of what the summary statistics show;
• Conclusions based on the information available;
• Sensitivity analysis to examine the robustness of results across options presented; and
• Data quality, key sources of uncertainty, and limitations that affect conclusions regarding
potential differential effects.
Analysts should follow best practices appropriate to the questions under consideration (see Text Box
3.2). If it is not feasible to follow a particular best practice, analysts should explain why this is the case.
6.4.1 Proximity-Based Analysis
Proximity to a source of pollution is commonly used when a direct measure of risk or exposure is not
available and the activities or emissions associated with the stressor of concern are likely localized
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
60
-------
(Maantay et al., 2022; Banzhaf et al., 2019a; Cameron et al., 2012; Wolverton, 2009; Baden and
Coursey, 2002).90 Generally speaking, a proximity-based approach compares the demographic and
socioeconomic characteristics of population groups affected by the sources of pollution to the
demographic and socioeconomic characteristics of population groups unaffected by these sources.
Note that this approach typically cannot differentiate between sources based on the magnitude of
emissions, concentrations, exposure, or risk of health effects.
It is important to note that a proximity-based approach is not recommended when risk of exposure to a
specific stressor is not correlated with the location of its source. For example, exposure to pollutants
found in drinking water systems can be more dependent on distribution system characteristics than
proximity to a pollutant source or the treatment plant. Likewise, exposure to specific chemicals may
occur at home or in the workplace, such as through use of cleaning or personal care products or
because of lack of access to personal protective equipment.
For practical reasons, the boundary of an affected area is usually based on a Census-defined
geographic area (e.g., census tract or block group) or a distance-based buffer (e.g., a specified radius
around a site). It is critical to use accurate spatial information when mapping the location of sources of
pollution. Analysts must also decide what distance from the facility most accurately reflects the
community's exposure to a stressor; no single distance is appropriate for all analyses. The buffer
distance around sources of pollution can be chosen to approximate actual risk and exposure, although
distance should be the same around each source. In some cases, it may be possible to use a
dispersion model to select a buffer that approximates the effect of atmospheric conditions (for instance,
wind direction and weather patterns) on exposure; these types of models are more data intensive.91
Using a proximity-based approach to evaluate the implications of a regulatory action to improve surface
water quality deserves special mention, as proximity is along a river network or water feature affected
by the regulation. In these cases, analysts should select a distance upstream or downstream that is
likely related to the stressor(s). Pollutant transport and dilution models can also provide insight on the
relevant distance. Some pollutants may degrade after a relatively short distance traveling in water,
while others (e.g., bromide) maintain their physical properties for hundreds of kilometers. In addition to
choosing the up or downstream distance, it is appropriate to incorporate a buffer around the water
segment to define the affected population. Depending on the regulation's effects, the buffer could be
relatively narrow (e.g., 1, 3, or 5 miles) or wide (e.g., up to 25 miles). Multiple buffers around affected
water segments may also be appropriate to reflect uncertainty as well as distinct exposure pathways,
such as via drinking water or subsistence fishing.
Regardless of how the size or extent of the affected area is selected, a proximity-based approach
typically assumes that the effects of the stressor(s) occur only within the designated boundary (i.e.,
people located outside the boundary do not suffer ill effects) and that all individuals residing within the
90 Even when risk or exposure modeling is available, proximity-based analysis with a relatively small distance buffer (e.g., one
or three miles) may offer insights into who is impacted by potential harms from changes in economic activity (e.g., noise,
odors, traffic, leaks).
91 It may also be possible to use more continuous measures of distance such as distance to the nearest polluting site or, when
additional information is available, an emission-weighted distance measure. For an overview of proximity-based analysis,
including a discussion of various spatial analysis techniques used in the literature, see Chakraborty et al (2011), Chakraborty
and Maantay (2011), and Mohai and Saha (2007).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
61
-------
boundary are equally exposed.92 As such, a proximity-based analysis is not able to determine which
populations within the boundary may face higher risks or adverse health effects. The results of a
proximity-based analysis may also vary with the geographic unit of analysis (e.g., Mascarenhas et al.,
2021; Mohai and Saha, 2007; Ringquist, 2005). For this reason, analysts should explore alternative
geographic units or distances when defining proximity to a source and describe the choices and
assumptions that are used in selecting specific buffers. When conducting a proximity-based analysis,
the results can also vary with the method used for assigning populations from Census-designated
geographic units (e.g., a tract or block group) to a specific distance buffer (see Text Box 6.4).
The two groups - individuals located near and far from the source, as determined by the selected buffer
- can be compared based on simple statistical or regression estimation techniques. Statistical tests on
summary data can be used to identify whether, on average, statistically discernible differences exist in
the characteristics of the two groups. Regression techniques, such as a binary logit, can formalize this
comparison, where the dependent variable takes on the value of 1 for areas where one or more
sources are located, and 0 indicates areas with no sources of the stressor. The independent variables
are demographic or socioeconomic variables, such as share of the population that is low-income in that
area. A statistically significant coefficient on an independent variable indicates a measurable difference
in the demographic or socioeconomic variable across geographic areas with and without stressor
sources. Note, however, that the coefficient estimates from such a regression cannot be interpreted as
identifying the causal effects of race, income, or other characteristics on source location.93 See Section
6.6.3 for a discussion of statistical significance.
Advantages of Proximity-Based Analysis
• Provides a quantitative analysis of the characteristics of communities in nearby locations;
• Can be a statistically rigorous approach if supported by data;
• Accounts for where individuals reside, providing a proxy for exposure when other more detailed
information is unavailable; and
• Can be used to identify potential hot spots.
Disadvantages of Proximity-Based Analysis
• Requires accurate information on locations of sources;
• Less useful when risk of exposure to the stressor is not correlated with source location;
• Cannot distinguish between sources based on the level of exposure, risk, or health effects for
the population within the boundary; and
• Exposure is often defined as a binary indicator instead of a continuous measure.
92 Chakraborty and Maantay (2011) address how to account for areas with more than one pollutant source, which are typically
treated the same as those with only one source. Each pollutant source is treated as identical with regard to its effect on the
nearby community. In reality, sources may vary widely in size, age, and production techniques resulting in differing amounts of
emissions.
93 See Banzhaf et al. (2019b) for an overview and empirical evidence of the possible causal mechanisms between pollution
and demographics.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
62
-------
6.4.2 Exposure and Risk Modeling
Exposure and risk modeling generally falls under the domain of HHRA, where a set of common
methods and tools are typically employed to inform risk-based regulatory decisions, as described in
Chapter 5 and Appendix B. These tools can be useful, by themselves or in combination with other tools
and data, for evaluating EJ concerns.
When data are available, analysts may choose to combine either a direct measure of or proxy for
exposure with fate and transport modeling to examine distributional effects at a disaggregated level.
For instance, the EPA's Air Toxics Screening Assessment tool uses fate and transport modeling to
estimate hazardous air pollutants concentrations and respiratory and cancer risks from point, non-point,
and mobile sources at the census tract level. The Risk-Screening Environmental Indicators (RSEI)
model combines a fate and transport model and information on chemical toxicity to estimate the
dispersion of toxic chemicals reported by specific facilities to the Toxics Release Inventory (TRI). Many
EPA offices also maintain fate and transport models specific to a particular regulatory context.
Information from these types of modeling exercises can be combined with demographic and
socioeconomic data to generate baseline and regulatory distributions of pollutants by population groups
of concern. (See Appendix B for examples.)
Likewise, direct measures of surface water quality or modeled water quality data can be combined with
hydrological modeling to estimate pollutant concentrations in waterways. This allows analysts to
consider exposure to pollutants such as lead and mercury via fish consumption. Drinking water quality
sampling results or violations of the Safe Drinking Water Act can also be used along with water system
service area boundaries to estimate exposure to pollutants through drinking water and assess
distributional effects across population groups.94
In cases where disaggregated information is available on differences in exposure across population
groups and there is a credible and scientifically defensible approach for translating exposures into a
response (e.g., evidence from the literature or HHRA on how a given level of exposure manifests as a
change in the risk of a specific health effect), it may be possible to characterize differences in health
effects due to the regulatory action. See U.S. EPA (2024f), U.S. EPA (2019a), and U.S. EPA (2014b)
for specific guidance on risk assessment, exposure assessment, and quantifying benefits, respectively.
In some cases, it also may be possible to combine exposure data with information on differential
responses across population groups.
Advantages of Exposure and Risk Modeling Methods
• Represent the most detailed and rigorous type of analysis; and
• Provide the most direct source of information on exposures or other outcomes.
Disadvantages of Exposure and Risk Modeling Methods
• Require detailed data at a fine geographic scale;
94 When evaluating exposure to specific drinking water pollutants, it is important to know who is served by a specific water
system. However, drinking water service areas rarely correspond to Census-defined or other standardized geographic
designations. See https://www.epa.qov/qround-water-and-drinkinq-water/communitv-water-svstem-service-area-boundaries.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
63
-------
• Are more complex to implement; and
• Provide results in a form that may be more challenging to communicate to the public.
6.4.3 Combining Qualitative and Quantitative Approaches
While the EPA often uses quantitative data and analysis to support the regulatory process, available
data are often not sufficiently disaggregated to allow quantification of distributional effects in the
baseline and across regulatory options. Other times, only partial information may be available.95 In
either case, the use of qualitative information or methods may be an appropriate supplement to
quantitative approaches. Qualitative methods may be particularly useful for offering insight into people's
values, behaviors, motivations, cultures, and local lived experiences, or when providing context for the
cumulative effects of exposure to multiple stressors on human health and well-being, which are often
omitted from quantitative assessment. Qualitative approaches may also be useful at the preliminary
stages of analysis.
Qualitative approaches systematically collect data via a series of open-ended questions and then
evaluate them to identify emerging patterns or themes (Tashakkori et al., 2020). Sample size is typically
small but allows for practitioners to dive more deeply into specific questions of interest. Qualitative
approaches share one or more of the following characteristics:
• Employ a variety of empirical materials, such as case study, narrative or personal experience,
introspection, life story, interview, observational, historical, interactive, and visual text;
• Gather empirical materials using some form of observation or interviewing method;
• May be iterative, with initial results informing later choices; and
• May rely on primary or secondary data sources, or a combination of the two.
Qualitative data can also be evaluated using a variety of approaches that range from identifying
interconnections within narrative data to more formal coding and statistical analysis of emerging themes
(referred to as content analysis).
Qualitative and quantitative approaches can also be used in combination (referred to as mixed
methods). They can be used iteratively or sequentially to inform a subsequent analytic stage (e.g.,
qualitative information may be collected via focus group to inform subsequent quantitative analysis of a
larger sample). Qualitative and quantitative approaches can also be used concurrently to shed light on
different aspects of the same phenomenon or to identify areas where themes or patterns converge or
diverge across the two data sets. It is also possible to combine qualitative and quantitative data into a
single analysis (Taherdoost, 2022).
Analysts should use their best judgment when evaluating the appropriate use of quantitative and
qualitative approaches for the analysis of EJ concerns. Similar to quantitative data, qualitative data can
be evaluated for both internal validity (e.g., are the findings credible?) and external validity (e.g., are the
95 U.S. EPA (2024f) discusses how to consider qualitative information in benefit-cost analysis (Chapters 7 and 11).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
64
-------
findings transferable?), and are subject to EPA guidance on the quality of scientific and technical
information used in support of Agency decisions (See Text Box 3.1).96
Note that approaches that directly engage local communities or Tribes can be long, time-intensive
processes and are subject to Paperwork Reduction Act requirements and Tribal consultation policies.
Special thought should also be given to addressing barriers to meaningful engagement when using
qualitative approaches, including those related to disability, language access, and lack of technical
assistance, tools and resources (See Sections 2.3 and 5.3.1.2, and Text Box 5.1).
Advantages of Qualitative Approaches
• Useful supplement when data are unavailable or incomplete for conducting a quantitative
analysis; and
• Allow analysts to incorporate hard-to-quantify information, such as cultural factors, pre-existing
vulnerabilities, public narratives, and community lived experiences.
Disadvantages of Qualitative Approaches
• Can be difficult to summarize results succinctly;
• Results can be uncertain, and the degree of uncertainty can be difficult to characterize; and
• Can be difficult to compare to quantitative information, (e.g., from a benefit-cost analysis or risk
assessment).
6.5 Analytic Considerations
Regardless of the analytic approach taken, analysts make a number of key decisions that can have a
substantial effect on the results of the analysis, including: the geographic and temporal scope of the
analysis; how to specify the comparison population group; how to spatially identify and aggregate
effects across affected and unaffected populations; whether to conduct analysis from a community
and/or facility perspective; and how to evaluate underlying variability, including the potential for
hotspots.
An important general strategy in analyzing EJ concerns is the use of sensitivity analysis. Due to the
uncertainties associated with the analytic decisions discussed below, sensitivity analysis around key
assumptions is often critical for clearly communicating results to the public.
6.5.1 Geographic and Temporal Scope
The geographic scope of analysis for an EPA regulatory action is often the entire United States since
requirements typically apply nationwide. However, in some cases the effects of a regulatory action are
expected to be concentrated in specific regions or states. In such cases, it may make sense to analyze
and present differences in health and environmental outcomes across population groups at both a
national and a sub-national level. Because the geographic scope can affect the results of the analysis
96 SeeTesch (2013) for a discussion of different types of qualitative analyses. Tashakkori etal. (2020), Seawright (2016), and
Lieberman (2005) provide approaches and best practices for combining qualitative and quantitative research. Mohai and Saha
(2015) also provide a literature review of mixed method and qualitative studies in the EJ context.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
65
-------
(Baden et al., 2007), analysts should make certain that the scope is relevant for the regulatory action
under consideration. In addition, it is important to keep in mind that differences in health and
environmental outcomes in one region or state may not necessarily hold in other regions or states.
It may be important to evaluate regulatory action effects on both shorter and longer time horizons. For
instance, while a regulatory action may result in near-term reductions in emissions, changes in health
and other risks may occur on a longer timeframe. In some cases, effects may even be felt
intergenerationally (e.g., climate change) and the analysis may accordingly extend beyond the current
generation to include a robust discussion of far-future health effects and costs. In general, the period of
time over which the analysis is conducted should also be consistent with other parts of the regulatory
analysis.
The scope of the analysis should generally match the scope used in other parts of the regulatory
analysis (e.g., benefit-cost analysis). However, in some situations, using a different time horizon or
spatial scale may be appropriate when considering EJ. For example, phasing in of regulatory
requirements or relocation of polluting activities in response to the regulatory action may result in EJ
concerns due to effects that occur on a time horizon or spatial scale that differs from other effects
considered in the regulatory analysis. If such situations arise, analysts should clearly articulate the
reasons for considering an alternative time horizon.
Another aspect of characterizing temporal scope is adequately anticipating the long run dynamic effects
of a regulatory action (Cain et al., 2024). The literature uses spatial sorting models to examine how
regulations may affect residential location choice but typically focuses on a specific city or region (e.g.,
Redding and Rossi-Hansberg, 2017; Kuminoff et al., 2015).97 98 Spatial sorting can occur when
improved environmental quality is capitalized into housing values, attracting higher-income households
and shifting renters and lower-income households to less expensive neighborhoods with lower
environmental quality (Melstrom and Mohammadi, 2022). On the other hand, some residents may be
more likely to move into high-risk zones due to differences in housing prices (Bakkensen and Ma,
2020). Given the challenges of modeling these types of effects on a national scale, it is recommended
that analysts qualitatively discuss possible household responses based on the available literature, while
acknowledging the limitations of the analysis.
6.5.2 Comparison Population Group
To evaluate differences in effects for population groups of concern, results need to be presented
relative to another group, typically referred to as a comparison population group. How the comparison
population group is selected has important implications for evaluating differences in health, risk, or
exposure effects across population groups of concern. It is possible to define the comparison
population group as individuals with similar socioeconomic characteristics in areas of the state, region,
or nation unaffected by the regulatory action (i.e., within-group comparison) or as individuals with
different socioeconomic characteristics within the affected areas (i.e., across-group comparison).
97 One exception is Fan et al. (2018). They link spatial sorting and economy-wide models of the United States to explore where
people migrate in response to increased risk of extreme temperatures, while accounting for wage and housing price
feedbacks.
98 Likewise, while hedonic price methods may be useful for demonstrating how changes in environmental quality factor into
housing prices, predicting the effect of such price changes on household migration by race or income may be infeasible.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
66
-------
Analysts should aim to define the comparison population group for an across-group comparison as
similar as possible to the population group of concern, but without the socioeconomic characteristics
defining the group of concern. For example, analysts could compare the proportion of low-income
households within areas affected by the regulatory action to the proportion of non-low-income
households within the same affected areas. If analysts have fate-and-transport information on
emissions, they can compare the average concentrations faced by low-income households within the
affected areas to those faced by non-low-income households living in the same areas. Thus, the results
from an across-group comparison indicate how the likelihood of risk or exposure within the affected
areas varies with demographic and socioeconomic characteristics.
A within-group comparison compares the likelihood of risk or exposure for a specific demographic or
socioeconomic group in affected areas to the likelihood of risk or exposure for that same demographic
or socioeconomic group elsewhere. Again, analysts should aim for the comparison group to be as
similar as possible to the population group of concern but without the risk or exposure of interest. For
example, analysts can compare the proportion of low-income households within areas affected by the
regulatory action to the proportion of low-income households in unaffected areas. Similarly, if analysts
have information on the fate and transport of emissions, they can compare the average concentrations
faced by low-income households within the affected areas to those faced by low-income households
living in areas unaffected by the regulatory action.
If a regulatory action is expected to differentially affect populations within a given area (e.g.,
communities living near regulated facilities or in a specific region), then a combination of within- and
across-group comparisons can demonstrate whether there are differences between specific population
groups of concern and the general population. Across-group comparisons are also informative in
instances where it is difficult to identify a comparison population group because a large share of the
U.S. population will be affected by the regulatory action. In these instances, it may not be appropriate to
use a comparison population group that includes the population of concern (such as the entire U.S.
population), as this will reduce the differences between them by construction.
It is unlikely that the same comparison population group will be appropriate in every instance. Analysts
should carefully document the criteria used to select the comparison population group for a particular
regulatory action, why it provides a useful counterfactual to the population group of concern, how the
two groups differ, and whether and why those differences might be important to the interpretation of
results within the EJ analysis.
In some contexts, it may make sense to define the comparison population group at a sub-national level
to reflect differences in socioeconomic composition across geographic regions. For instance, because
larger populations are concentrated in urban areas, the results of the analysis are often dominated by
effects in these areas. If a regulatory action primarily affects rural areas," inclusion of urban areas in
the comparison population group may obscure underlying differences.
Using a sub-national comparison population group may also be more defensible when there is a great
deal of heterogeneity in industrial development and economic growth and/or inherent differences in
demographic and socioeconomic composition across geographic regions (e.g., relatively more
99 See the Federal Committee on Statistical Methodology's discussion of the different definitions the federal government uses
to classify rural populations, as well as the advantages and disadvantages associated with the different definitions:
https://nces.ed.gov/fcsm/edt/rural definitions.html
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
67
-------
Hispanics reside in the Southwest) (e.g., Bowen, 2001). In these instances, analysts may consider
comparing the populations of concern near an affected source to populations living just outside of the
affected area. By selecting a comparison group that is geographically proximal and of a relatively
similar size to the affected population, analysts can better control for other confounding factors. See
Text Box 6.4 for examples.
Note, however, that placing restrictions on the comparison population group may "reduce the power of
statistical tests by reducing sample sizes" (Ringquist, 2005) or bias results against finding differences in
adverse human health and environmental effects because such restrictions reduce variation in
demographic or socioeconomic variables of interest. In selecting a comparison population group,
analysts should therefore evaluate how different comparison population groups affect the way
information is conveyed. When appropriate and practicable, analysts should conduct sensitivity analysis
using alternate definitions of the comparison population group to provide a more complete depiction of
potential effects.
6.5.3 Spatial Identification and Aggregating Effects
The spatial distribution of health and welfare outcomes is a relevant consideration for some regulatory
actions, such as those that reduce emissions from point sources that have fairly localized effects or
when there is a differential distribution of associated health or environmental effects. In other cases, the
regulatory action's effects may be more widespread, and spatial distribution is less relevant (e.g., when
exposure to a chemical substance depends on its purchase, use, transport, or disposal).
When exposures, risks, or human health effects are spatially distributed, analysts need to determine
how to spatially identify affected and unaffected populations. The nature of the stressor(s) should guide
analysts' choices of the geographic area of analysis. Some air pollutants, for example, may be emitted
out of tall stacks and travel long distances, affecting individuals hundreds of miles away from the
sources and thereby making it appropriate to choose a relatively large geographic area. In contrast,
water pollutants or waste facilities may have more localized effects, making it appropriate to select
relatively small areas for analysis. Likewise, an assessment of local effects from point sources -
including possible traffic, odors, and noise implications from changes in production - may call for more
spatially resolved data than those that affect regional air quality.
Complications can arise when the spatial resolution of the analysis is either too refined or too coarse.
For example, small geographic areas of analysis (e.g., less than one mile from the source location) may
not be sufficiently outside of an emitting source's fenceline to capture potential effects on nearby
populations. A geographic unit of analysis that is too large may begin to resemble state or national
averages and can be more difficult to interpret due to the influences of many, multiple sources of risk
and exposure.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
68
-------
Text Box 6.4 Selecting a Comparison Population Group - Recent Examples
For the final Phasedown of Hydrofluorocarbons rule (U.S. EPA, 2021a), analysts examined the
socioeconomic characteristics of communities living within a specific distance of a production facility
subject to the rule both in aggregate and for each individual facility. The comparison population
group for the facility-by-facility analysis was selected to reflect whether a facility is in a rural or urban
area (i.e., overall or rural average) and presented at both the state and national levels.
For the final Control of Air Pollution from Heavy Duty Engine and Vehicle rule (U.S. EPA, 2022f),
analysts used a proximity-based approach to examine the proportion of the population living within
different distances of major truck freight routes by race, ethnicity, and income. The analysis then
compared the characteristics of those living within 1,000 meters to those living beyond 1,000 meters.
Results were further delineated based on U.S. Department of Agriculture rural-urban continuum
designations and by region.
For the final Mercury and Air Toxics Standards rule (U.S. EPA, 2011c), analysts examined mortality
risk associated with fine particulate matter (PM) by race, income, and poverty level for people living
in high-risk counties (i.e., in the counties identified as experiencing the top five percent of risks from
exposure). The comparison population group was defined as people living in counties not facing a
high mortality risk.
For the final Lead Emissions from Aircraft Engines Endangerment finding (U.S. EPA, 2023g),
analysts examined the characteristics of residents within 1 kilometer of airport runways in the U.S.
(where piston-engine aircraft may use leaded aviation fuel). Resident characteristics included race,
ethnicity, poverty level, and children under age five. The comparison group was defined as residents
that live within 1 to 5 kilometers of those locations.
Another challenge that arises when using geospatial data is when data are spatially autocorrelated -
that is, locations in closer proximity are more highly correlated with the variable of interest than those
further away. When this occurs, then the assumption that error terms are independently distributed is
violated (Chakraborty and Maantay, 2011), which can violate the assumptions of some statistical
techniques.100 When considering whether and by what method to address spatial autocorrelation
(Mennis and Heckert, 2017; Chakraborty, 2011), it is important to consider the goal of the analysis
(e.g., statistical association, prediction, significance testing of an estimated effect, or distinguishing
correlation from causation).101
The quality and type of data available also affect the spatial resolution of the analysis (see Section 6.3).
For instance, in rural areas Census-based geographies can be large, introducing a higher level of
uncertainty regarding where specific populations groups live within that boundary. Thus, more than one
geographic area of analysis to examine the robustness of results may be useful since effects are
100 For more discussion on methods to address spatial autocorrelation, see Chapter 8 in Haining (2003). For examples of
regression techniques that have been employed to address spatial autocorrelation, see Kuminoff et al. (2010).
101 Bayesian regression techniques may be appropriate in some circumstances, for instance when combining data across
multiple data sets where there are complex interdependencies across space and time, and it is difficult to quantify uncertainty
using traditional approaches. See Chun et al. (2012) and Carlin and Xia (1999) for examples of how these approaches have
been applied in the EJ context.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
69
-------
unlikely to be neatly contained within geographic boundaries and results may be sensitive to the choice
of the geographic area of analysis (Baden et al., 2007; Mohai and Bryant, 1992).
Census-based geographic delineations and definitions often align with topographical or infrastructure
features such as rivers, highways, and railroads. As a result, they may exclude a portion of the affected
population that experiences the same adverse effects from a stressor, even if they are on the other side
of the physical feature. While Census-based definitions are easily accessible and offer many options
regarding geographic scale, use of a spatial-defined distance can potentially allow for a more flexible
approach. Using geospatial tools, analysts can define spatial buffers around emissions sources that are
more uniform in size and easier to customize to reflect the appropriate scale and characteristics of the
emissions being analyzed (e.g., fate and transport).
Buffers can be created and combined with Census data in many ways, including selecting the Census
units (e.g., tracts or block groups) that intersect the buffer circle, selecting tracts with centroids within
the buffer circle, spatially intersecting the buffer circle with the tract polygon, and transferring the
attributes from tracts to the buffer area using area- or population-weighting (see Text Box 6.5). Mohai
and Saha (2007, 2006) show that using distance from a facility instead of a buffer-based approach may
provide a more complete comparison of effects (see also Mascarenhas et al., 2021; Mohai and Saha,
2015).
Analysts should be aware of the potential for the "modifiable areal unit problem" when aggregating
geospatial data. The modifiable areal unit problem refers to the fact that results can be sensitive to the
level of spatial aggregation used in the analysis, such as how the results may differ if the data are
aggregated to the census tract versus county level (see Shadbegian and Wolverton, 2015; Baden et al.,
2007; Mohai and Bryant, 1992). Additionally, analysts should exercise caution when extrapolating the
results to more spatially resolved units. For example, relationships estimated using county level data
may mask neighborhood level associations between pollution exposure and demographic and
socioeconomic characteristics (see Banzhaf et al., 2019a).
When selecting a unit of analysis, it is also important to weigh any potential tradeoffs between
completeness (fully capturing the populations at risk) and variability in risk (for instance, possibly
masking information about those most at risk by including populations that are much less affected).
Analysts are encouraged to discuss the approach used to create buffers and aggregate geospatial
data, as what is most appropriate will vary with the stressor(s) affected and data used in the analysis,
and to provide a transparent justification of their choice. In some cases, it may be helpful to consider
multiple buffers to evaluate the effects of a regulatory action, for instance, because of uncertainty
regarding fate and transport of a specific environmental stressor or because the regulatory action
affects environmental stressors that travel different distances.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
70
-------
Text Box 6.5 Centroid-Based vs. Areal Apportionment Methods
Demographic and socioeconomic data are typically available at an aggregate level within the
boundaries of a Census geographic unit (e.g., tract or block group). Analysts must then decide how
to map the available data to circular buffers around an emissions source. There are two main
approaches.
The centroid-based method assigns the population from a Census geographic unit to the circular
buffer when the centroid of the unit lies within it. In the figure below, the centroid method excludes
the population in Census unit Ci because its centroid lies outside the circular buffer but includes the
entire populations of Census units C2, C3, and C4.
An alternative method is areal apportionment, which assigns populations to a circular buffer based
on the extent to which a Census geographic unit overlaps with that buffer. In the example from
above, this approach includes only the portions of C1 - C4 that overlap with the circular buffer. It
apportions the population living within C1 - C4 to the circular buffer by assuming they are uniformly
distributed within each Census geographic unit.
For a given spatial resolution, areal apportionment can yield more accurate estimates than the
centroid-based approach. When the centroid-based approach is applied to Census geographic units
with large spatial extents (e.g., rural areas), analysts may find that the centroids of some units are
not located within the circular buffer. This leads to the exclusion of geographic areas from the
analysis and can skew results. Note, however, that the centroid-based method is less
computationally intensive and therefore may have the advantage when Census geographic units are
small.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
71
-------
6.5.4 Facility vs Community-Based Perspectives
Exposure to other environmental and non-environmental stressors can increase the vulnerability of
individuals or population groups to negative health effects from exposure to a specific environmental
hazard. While explicit modeling of these interactions is often not feasible, analysts can shed light on this
issue by evaluating and presenting results using not only a facility, but also a community-based
perspective.
An analysis with a facility-based perspective primarily considers who may be exposed to sources
regulated by the specific action under consideration. For example, such an analysis would examine
proximity, emissions, concentrations, or risk associated with each regulated source in conjunction with
the demographic and socioeconomic characteristics of those most likely to be exposed.
However, communities may be affected by multiple sources of pollution relevant to characterizing risk
for a specific regulatory action. An analysis that takes a community-based perspective considers
proximity, emissions, concentrations, or risk to a given community from multiple nearby sources of
pollution to which individuals are exposed, accounting for the possibility that certain communities face
increased vulnerability due to a greater number of nearby pollution sources.
Examples of analyses that take a community-based approach include:
• The total number of polluting facilities within a specific distance from the regulated source.102
• Measures of other sources of exposure to the same environmental stressor being regulated
(e.g., age of home and proximity to a major highway are positively associated with the likelihood
of homes having lead in paint, service lines, and soil).
• Measures of exposure or risk to a broader array of contaminants in communities near the
regulated source (e.g., cancer risks from exposure to air toxics).103
• Explicitly accounting for non-regulated or other routes of exposure when modeling risk faced by
the affected communities.104
6.5.5 Evaluating Underlying Variability and Identifying Potential Hot Spots
In addition to presenting aggregate results for population groups of concern affected by the regulatory
action, it is important to understand the extent to which there are heterogeneous effects, both within
specific population groups as well as across communities, given that communities often vary widely in
the risks they face from the affected sources as well as from other environmental and non-
environmental stressors. When data allow, analysts should characterize the distribution of risks,
exposures, or outcomes within each population group of concern, not just average effects, with
particular attention paid to the characteristics of populations at higher risk of exposure. When relying on
proximity-based analysis, differentiating results by key facility characteristics that may be correlated
with risk (e.g., plant age, capacity, production levels, accident history, types of chemicals stored on site)
can be useful.
102 For example, see the EJ analyses in U.S. EPA (2024j) and U.S. EPA (2022g).
103 For cancer risks associated with air toxics, see https://www.epa.qov/AirToxScreen.
104 See the EJ analysis in U.S. EPA (2022h), in which cancer risk is modeled using hazardous air pollutant concentrations from
both regulated and non-regulated sources within a specific distance of regulated sources.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
72
-------
It is also important to evaluate the potential for hot spots, with particular attention paid to the
communities in the upper end of the distribution of exposure or risk. Hot spots refer to geographic areas
with higher levels of localized concentrations of emissions from one or more sources within a larger
geographic area with more "normal" environmental quality. Hot spots may result from baseline
conditions, such as exposure to other pre-existing stressors within the community. For example, the
siting of polluting facilities near populations of concern may be reflective of historic discriminatory land
use policy and practices (Estien et al., 2024; Gonzalez et al., 2023). Studies have demonstrated that
population groups of concern experience higher concentrations of air toxics due to proximity to nearby
industrial facilities (Bouvier, 2014; Zwickl et al., 2014; Ash et al., 2013; Pastor et al., 2002). Extractive
industries are also associated with localized environmental concerns such as contamination of
groundwater in rural areas and Tribal lands from abandoned uranium mines, resulting in high
incidences of kidney cancer and other adverse health effects (Ingram et al., 2020; Lewis et al., 2017;
Corlin et al., 2016). See Text Box 6.6 for a specific example.
It is also possible that hot spots may be created, exacerbated, or mitigated following a regulatory
action. Relevant issues to consider may include proximity to multiple sources of pollution, specific
exposure pathways, and other drivers of increased vulnerability. Qualitative or other sources of data
may also help to identify specific population groups or communities where a more detailed analysis is
warranted.
A preliminary analysis early in the analytic process may help identify the potential for hot spots. In
addition, information received via public comments can yield insights into the potential for a regulation
to create or exacerbate hot spots.105 More sophisticated modeling or econometric approaches may also
facilitate systematic identification of potential hotspots (e.g., fate and transport modeling, hydrologic
modeling, hedonic analysis). If a relatively small number of potential hot spots are identified, case
studies or in-depth qualitative analysis may be useful.106
105 For instance, the public has expressed concern that cap-and-trade policies designed to reduce carbon dioxide emissions
may lead to increases in criteria air pollutants in already overburdened neighborhoods, further exacerbating existing health
issues. In the context of the California trading program, research has not reached consensus on the degree to which the policy
has exacerbated existing emissions in communities of color or low-income communities (e.g., Mansurand Sheriff, 2021;
Hernandez-Cortes and Meng, 2020; Cushing et al., 2018; Grainger and Ruangmas, 2017; Fowlie et al., 2012).
106 See Arcury et al. (2021), Rocha et al. (2017), Mohai and Saha (2015), Schwartz et al. (2015), and Grineski (2009) for
examples of qualitative discussions of hot spots.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
73
-------
Text Box 6.6 Hot Spots Example: Abandoned Uranium Mines on Navajo Nation Land
In the western United States, many abandoned hardrock mines are located near or on Tribal lands.
Over 600,000 Native Americans live within 10 kilometers of an abandoned mine (Lewis et al, 2017).
Of particular concern are the over 4,000 abandoned uranium mines, 75% of which are within 80
kilometers of an Indian reservation. Over 500 abandoned uranium mines are located on or near the
Navajo Nation in the Four Corners region alone.
Documented adverse health outcomes for Navajo Nation members exposed to uranium mine waste
during the mining era (1940s - 1980s) include congenital anomalies at birth (Shields et al., 1992)
and later life kidney disease (Lewis et al., 2017). Navajo men who were underground uranium
miners had high excess risks of lung cancer (Gilliland et al., 2000). One report found that Navajo
residents were more than seven times more likely to die from gallbladder cancer, more than four
times more likely to die from stomach cancer, more than two times more likely to die from kidney
cancer, and almost two times more likely to die from liver cancer compared to non-Hispanic White
people (Navajo Epidemiology Center, 2023).
In addition, these mines' legacy contamination of Tribal groundwater poses a drinking water concern
to residents, given limited public water infrastructure in the region and widespread reliance on
private wells. Studies testing drinking water in the Navajo Nation have detected uranium above the
national regulatory limit (Ingram et al., 2020; Corlin et al., 2016).
Northern AIM Region
North Central Al M Region
Wnlern Al M Region
[C'enlr!il AHM,Region,
Cameron
S" Kastcrn Al'M Region
Southern AI M Region
rwt ircrcrrcr! PCorp . 'itSCO USGS TAQ NP8
•datter NL Survey, fcvi Mb " I. tw Cum
OpySiiwiMipcyaife/yi sndVxOiSUw»cocvimn Nation AUM Regions
Map of the Navajo Nation with abandoned mine sites represented by Xs and the Navajo Agencies and
Chapters outlined within the map (OpenStreetMap on Garmin May 13, 2019). Retrieved from:
https://wiki.openstreetmap.orq/w/index.php?title=OSM Map On Garmin&oidid=1851574
Technical Guidance for Assessing Environmentai Justice in Regulatory Analysis - Second Edition
74
-------
6.6 Characterizing Analytic Results
Once an EJ analysis has been conducted, analysts face choices about how to characterize and
communicate the results. Analysts need to present summary metrics for relevant population groups of
concern and the comparison population group and characterize the differences between them. This
section discusses the way in which information from the analysis can be summarized and presented,
including the choice of summary metrics, ways of displaying the results in tables, maps or other visual
displays, and the distinction between statistical and policy significance when interpreting results.
6.6.1 Choice of Summary Metrics
Simple summary measures can be used to characterize the distribution of health and environmental
effects in the baseline and for regulatory options relative to appropriate comparison population groups.
Metrics commonly used to summarize information on effects across population groups of concern
relative to a comparison population group include:
• Levels, stratified by demographic or socioeconomic characteristic (e.g., average PM2.5
concentrations for Black residents within three miles of regulated sources compared to average
PM2.5 concentrations nationally).
• Percents, stratified by demographic or socioeconomic characteristic (e.g., percent low-income
within three miles of regulated sources compared to percent low-income nationally).
• Differences in means (e.g., the difference in mean cancer risk for low-income vs. for non-low-
income residents affected by the regulatory action).
• Percentile rankings (e.g., how affected areas rank, in percentile terms, relative to the national,
regional, or state average).
• Relative ratios (e.g., the ratio of average risk or exposure for one group to the average risk or
exposure for another group within the same geographic area).107
• Inequality indexes, such as the Gini coefficient (Muller et al., 2018), measure the extent to which
an outcome (e.g., emissions or exposure) is distributed unequally across society or
communities.108 Lorenz curves are closely related to Gini coefficients and can be used to
visualize inequality across the distribution of pollution exposures (Andarge et al., 2024).
107 A relative ratio of one indicates that the specific group has the same risk or exposure as the comparison group. The higher
the ratio is above one, the higher the risk or exposure relative to the comparison group. If the ratio falls below one, this
indicates that the specific group has a lower risk or exposure than the comparison group.
108 Alternatives to the Gini coefficient include the Atkinson index (Levy et al., 2006) and the Kolm-Pollak equally distributed
equivalent measures (Sheriff and Maguire, 2020). There are important interpretational differences between inequality
measures as applied to income or wealth (a good) vs. pollution exposure (a bad). For example, the Atkinson index places
greatest weight on the worst off when applied to income, but with respect to environmental harms it more prominently weights
those with the lowest pollution burden. Separately, inequality indexes may obscure potentially important information about
absolute levels of exposure.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
75
-------
Analysts should consider characterizing results of the EJ analysis using more than one type of
summary metric to provide a richer picture of potential effects. For instance, relative ratios can facilitate
comparisons across groups or locations because all ratios are in common units. However, without
presenting information on the absolute levels of risk or exposure, it is not possible to determine if either
group is at risk of experiencing a potential health effect. Analysts should also present information that
communicates underlying heterogeneity in the data, such as the degree of spread in the data relative to
the mean (i.e., standard deviation).
Counts of the number of sources or geographic areas where the percent of a specific population group
living nearby exceeds a particular threshold (e.g., the state/national average or a specific percentile)
are not recommended. Counts are hard to interpret because they do not account for differences in
population size or density across geographic areas. It is more informative to display metrics that
characterize the full population or risk distribution to understand the extent to which affected
communities differ from the comparison group.
To the extent that the underlying data allow, analysts should disaggregate the summary information so
that the public can discern how risk, exposure, and/or health effects vary for different types of
individuals within a population group. For instance, exposure or health effects can be presented for
income quantiles, in addition to presenting this information for those above or below a particular income
threshold. Likewise, analysts might describe the average demographic or socioeconomic
characteristics of workers differentiated by the affected industry and/or geographic areas relative to a
comparable but unaffected population of workers.
In many cases, the distribution of environmental exposure or risk across individuals or communities
follows a non-normal distribution with a long right tail. In other words, while many individuals may be
exposed to low or moderate levels of environmental stressors, a handful of them may face differentially
high risk. In these cases, demonstrating that the average of the affected group is similar to the national
average is not very informative. Rather, it would be useful to characterize the full distribution as well as
the exposure or risk for individuals or communities that are in the high-end of exposure, such as those
in the 95th or 99th percentile (Gochfeld and Burger, 2011). Information on risk, exposure, and/or health
effects can be presented for the average-exposed individual as well as a maximally-exposed individual
in each population group. If specific communities are substantially affected, analysts can present
summary statistics for those specific communities in addition to presenting aggregate summary
statistics for all communities affected by the regulatory action. (Also see Section 6.5.5.)
6.6.2 Displaying Results Visually
Tables, maps, and other visual displays help communicate a large amount of information in an
organized way to facilitate comparisons, convey results, and support discussion. Careful thought should
go into how information is presented, particularly when there are:
• Multiple comparison groups (e.g., state, U.S., rural areas);
• Different types of effects (e.g., pollutants, health effects, or other environmental metrics);
• Multiple categories of regulated facilities or types of sources;
• Many individual sources;
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
76
-------
• Clustering of sources in specific geographic areas;
• Multiple scenarios (e.g., baseline, multiple regulatory options); or
• Sensitivity analysis around key analytic assumptions (e.g., buffer distance).
Analysts need to clearly explain how to interpret the information presented in tables, maps, or figures to
properly contextualize results and guard against erroneous conclusions (e.g., a large percentage
change from a small baseline value may not be a large change in absolute terms).
Often more than one table is needed to present results. In addition, holding or shading specific cells
can ease navigation of a dense table of results. Table 6.1 illustrates how results for multiple types of
sources and several distance buffers can be presented within a single table. This example also uses
shading to indicate values above the national average.
Table 6.1 Example Summary Table for Proximity-Based Analysis Results
Sociodemographic
Category
Population
within 1 Mile of
Sites with
Legacy CCR Sis
Population
within 3 Miles of
Sites with
Legacy CCR Sis
Population
within 1 Mile of
Sites with
CCRMUs
Population
within 3 Miles of
Sites with
CCRMUs
U.S.
Population
Race
Asian
10.36%
4.66%
2.37%
2.82%
5.64%
Black or African American
13.47%
17.03%
8.37%
13.73%
12.53%
Native Hawaiian/Pacific
Islander
.03%
.08%
.06%
.07%
.18%
Native American or
Alaskan Native
.79%
.93%
.86%
.78%
.82%
Other
16.65%
15.82%
20.63%
18.94%
12.82%
White
58.47%
61.47%
67.71%
63.67%
68.01%
Ethnicity
Hispanic (any race)
26.27%
22.0%
32.61%
27.02%
19.24%
People of color
People of Color
52.42%
46.31%
46.70%
46.59%
41.14%
Poverty Level
Households below the
poverty level
16.21%
16.11%
14.94%
14.9%
12.71%
Other Sociodemographic
Indicators
Linguistically isolated
households
9.23%
5.42%
9.38%
6.05%
4.84%
Less than a high school
diploma
17.60%
14.02%
17.11%
15.57%
11.24%
Person with disability
15.53%
15.61%
14.66%
15.23%
12.70%
Source: Table 6-9. Estimated Percent of Key Sociodemographic Indicators Near Legacy CCR Surface Impoundment (SI) and
CCR Management Unit (MU) Sites (U.S. EPA, 2024k).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
77
-------
Visually displaying information in maps or figures can also help demonstrate how sources, risks, and
exposures are geographically distributed across population groups, including baseline conditions and
spatial clustering of sources (see Figure 6.1 for an example). Choropleth maps, which use color or
shading to represent the range of a variable's values, can be used to visualize the levels of
environmental, demographic and socioeconomic variables across space.
Note that it can be difficult to visually discern differences between baseline and regulatory options in
maps or figures unless differences are large.109 However, differences not discernible on a map may still
be important. Additionally, it is important to consider how visual indicators can be used to characterize
data uncertainty, how geographic boundaries relevant to the analytic context (e.g., watershed, state, or
Tribal lands) are identified on a map, and how to select appropriate intervals for visually representing
the distribution of the data. For this reason, visual displays are only suggestive of potential effects and
should be accompanied by tables or other graphics that allow the reader to access the underlying
statistical information.
Figure 6.1 Example Map of Relative Cancer Risks from Air Toxics within 1 Mile of
Regulated Sources
f nQofina Chemical Inc - 0^97;
^ J VA
(Chemours Louisville Works
tBaikin America Inc. -1.2)
(Chemours El Dorado -1.5) 1 J
J \ ¦-* /— A
TX [Mexichem Fltgfcinc. - 3.7>—\
/-~-v (Honeywell InternationiSWfic - Geismar Complex - 2.7)
U- — / \ FL
.Chemours - Corpus Christi Plant - 0.51 J \
Cancer R«k Relative
to State Rural Average
Source: U.S. EPA (2Q22g).
109 For an overview of general mapping best practices to communicate EJ concerns, such as selecting a projection, avoiding
unintentional misrepresentation, and choosing a color scale to represent values, see Stieb etal (2019).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
78
-------
6.6.3 Statistical Significance and Other Considerations
Tests of statistical significance can be used to examine whether the difference between the mean
values of two groups is due to factors other than chance.110 This can be done for a pairwise
comparison, which does not control for other factors, or via a regression approach, which allows
analysts to assess the relationship between two variables while controlling for other factors.
It may also be useful to examine parts of the distribution further from the mean (e.g., quantile
approaches) or to use approaches that can account for outliers, skewness or heteroscedasticity
(varying levels of spread) in the data. Note that the ability to test statistical significance is predicated on
having a sufficiently large sample size and, for parametric approaches, an assumed distribution (e.g.,
normality).111
It is important to understand that a statistical difference does not necessarily indicate that the difference
is meaningful from a policy perspective. For instance, analysts may find that low-income households
are more likely to be located near a pollution source than wealthier households, and that this effect is
statistically significant (i.e., the effect is statistically distinguishable from zero and not due to sampling
error).112 However, the difference in likelihood between these types of households could still be quite
small in magnitude. Analysts need to examine what the difference implies (e.g., how different poverty is
across geographic areas), and summarize those differences in a manner appropriate for policy
relevance.
When using multi-variate regression analysis, analysts need to be aware that many demographic and
socioeconomic characteristics are highly correlated with each other, making it difficult to interpret the
meaning of a coefficient on any given variable. In cases where standard errors may be correlated
across groups of observations, analysts can cluster standard errors to account for correlation (Abadie
et al., 2023; Cameron and Miller, 2015). These clusters can be at the geographic unit level, such as by
county, and/or at the temporal level, such as by year. Finally, analysts can consider other factors aside
from demographic and socioeconomic characteristics that may have influenced the location of sources
(e.g., labor and land costs, land use restrictions, proximity to highways, railroads, or ports). Regression
techniques can partially control for these factors; the use of statistical tests on summary data cannot.
See Kim and Chun (2018) and Gilbert and Chakraborty (2011) for examples of how researchers have
approached these issues.
It is also important for analysts to be aware of and discuss the biases and limitations introduced when
proximity or distance is used instead of risk and exposure modeling (see Mohai and Saha, 2015;
Chakraborty and Maantay, 2011). Given the analytic challenges associated with proximity-based
110 Tests of statistical significance could include a t-test or a chi-square test. Statistical significance is often reported in terms of
a p-value. A low p-value (e.g., <0.05) indicates that the likelihood that changes in the predictor variable are not related to
changes in the response variable is very low (i.e., the relationship between the predictor and response variable is statistically
distinguishable from zero).
111 See U.S. EPA (2006) for more information on different types of correlation coefficients, comparisons of means across
populations including nonparametric tests, treatment of outliers, tests of distributional assumptions, and further discussion of
statistical significance. The EPA also maintains a website on different analytic approaches and methods for conducting a
statistical data assessment. See https://www.epa.aov/caddis/caddis-volume-4-data-analvsis for more information.
112 When using U.S. Census Bureau ACS data, analysts can use the Census' Statistical Testing Tool to conduct statistical
analyses. See https://www.census.qov/proqrams-survevs/acs/quidance/statistical-testinq-tool.html for more information and to
download the tool. Also see Chapter 7 in U.S. Census Bureau (2020b).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
79
-------
analysis, it may only be possible to draw limited conclusions regarding differences across populations
groups (see Section 6.4.1).
Finally, it is important to address and characterize uncertainty. Point estimates alone do not provide
information about whether estimates are robust to alternate assumptions, nor can they convey the full
range of potential outcomes. When statistical analysis is used, information such as confidence intervals
and variance should be presented. Sensitivity analysis can also play a role in understanding the
robustness of outcomes to key assumptions. Where the analysis is sensitive to the choice of model or
method used, this uncertainty should also be described.113 Uncertainty can also be discussed by
highlighting limitations in the literature, identifying caveats associated with results, or highlighting gaps
in the data.
6.7 Assessing the Distribution of Costs and Other Effects
This section addresses when it may be appropriate to evaluate how economic costs or challenges are
distributed across population groups, how compliance and enforcement may vary across regulatory
options under consideration, and the evaluation of non-health effects. We specifically refer to costs as
defined in U.S. EPA (2024f).114
6.7.1 Distribution of Economic Costs
This EJ Technical Guidance mainly focuses on approaches to assess the potential for differential
exposure, risk, or health effects associated with regulatory actions on population groups of concern.
However, certain directives (e.g., E.O. 13175, E.O. 14008, and OMB Circular A-4) identify the
distribution of economic costs or challenges as an important consideration in developing policy
alternatives and for regulatory analysis. The economics literature also typically considers both costs
and benefits when evaluating distributional consequences of an environmental policy to understand its
net effects. Fullerton (2011) discusses six possible types of distributional effects that may result from an
environmental policy: costs to consumers via changes in relative product prices; costs to producers or
factors of production via changes in the relative returns to capital and labor; the distribution of scarcity
rents (i.e., excess benefits due to restricted nature of a good, such as pollution permits); the distribution
of environmental quality improvements; temporary costs of adjustment and transition (e.g., for capital
and labor); and the capitalization of environmental improvements into asset prices (e.g., land or
housing values).
In the context of EJ, the distribution of health or environment effects alone might convey an incomplete
- and potentially biased - picture of the overall burden faced by population groups of concern. For
instance, if costs are unevenly distributed such that low-income households bear a larger relative
share, it is possible that they may experience net costs even after accounting for environmental
improvements. That said, the consideration of economic costs in an EJ context may be challenging,
given a lack of data and methods in many instances.
113 Model uncertainty may also be addressed through techniques such as model averaging. See Steel (2020) for an overview
of model averaging applications in economics, for example. Chapter 5 of the Guidelines for Preparing Economic Analyses
(U.S. EPA, 2024f) also contains a more detailed discussion of statistical and model uncertainty.
114 Private costs are the costs that the purchaser of a good or service pays the seller. Social cost is the total burden a
regulation imposes on the economy. The bearers of social costs can be either specific individuals or society at large.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
80
-------
Whether to undertake an analysis of economic costs as it pertains to EJ is a case-by-case
determination. It will depend on the relevance of the information for the regulatory decision at hand, the
likelihood that economic costs of the regulatory action will be concentrated among particular types of
households, and the availability of data and methods to conduct the analysis.115 Analysts should
coordinate with economists from the Office of Policy when evaluating the potential relevance of
economic costs for EJ and the degree to which they can be discussed or analyzed.
In many cases, analysis of economic costs from an EJ perspective will not substantially alter the
assessment of distributional effects for population groups of concern. For instance, often the costs of
regulatory action are passed onto consumers as changes in prices or wages that are spread fairly
evenly across many households. When these price changes are small, the effect on an individual
household also will likely be relatively small. In this case, further analysis is unlikely to yield additional
insights.
However, in some circumstances further exploration of the distribution of economic costs may offer
substantial insight because costs are expected to differentially affect specific population groups. For
example, further analysis may be warranted when costs to comply with the regulatory action represent
a noticeably higher proportion of income for some population groups; when some population groups are
less able to adapt to or substitute away from goods or services with now higher prices; when changes
in environmental quality or health and costs are likely to accrue to the same set of individuals; when
costs are concentrated on some types of households (e.g., renters) or in specific geographic areas;
when there are identifiable plant closures in or relocation of facilities away from or into communities in
which population groups of concern reside and work; or when behavioral changes in response to the
costs of the regulatory action leave population groups of concern less protected than other groups.
While the Agency continues to investigate ways to improve incorporation of economic costs into an
analysis of EJ concerns, it recognizes that, even in cases where the information is relevant, data or
methods may not exist for full examination of the distributional implications of costs. For example, the
EPA may expect pollution control costs to be passed on to electricity consumers in the form of higher
prices that differentially affect budget-constrained households in particular regions more than others. To
evaluate the effects of the regulatory action properly, analysts need to understand how costs are
passed through to customers (which may differ by state); how these changes are broken down between
residential and commercial customers; what assistance is available for low-income consumers; how
consumption patterns differ by race, ethnicity, and income; and how these consumption patterns may
change in response to electricity price changes. Likewise, if environmental improvements associated
with the regulatory action are unevenly distributed, demand for housing in some neighborhoods may
affect rental prices for housing. This, in turn, may result in households moving to other locations that
have a different risk and exposure profile.
In addition to methodological limitations, incomplete data may limit the ability of analysts to fully
characterize the distribution of costs across population groups of concern. Specifically, available data
may only shed light on baseline distributions, without anticipating the distribution of costs in cases
where the regulatory action is expected to result in indirect behavioral changes through changes in
115 Note that there may be other effects of a regulatory action (e.g., employment) beyond direct compliance and social costs
but understanding how all effects vary across population groups may not be feasible. For example, data on the distribution of
changes in employment across low-income households may be difficult to assess.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
81
-------
price.116 Due to method and data limitations, it might not be possible to predict the total effect of a
regulatory action on different population groups. In these instances, the issue can be qualitatively
discussed and the limitations and assumptions associated with characterizing costs explained.
When analyzing the distribution of costs, other considerations include the time frame and use of partial
versus general equilibrium approaches for the analysis. For instance, most consumers may experience
similar price changes due to a regulatory action, but in the short run budget-constrained households
have more difficulty accommodating higher prices. In contrast, higher automobile prices due to a
regulatory action will initially affect higher income households who purchase new cars more frequently;
over a longer period of time, these higher prices may also affect lower-income households due to
changes in the prices for used cars. More extensive analysis could consider the use of dynamic general
equilibrium analysis to examine first and second-order costs and their implications for changes in
wages and prices across households over time. However, such analyses are typically resource- and
time-intensive, usually only utilized in cases where sectors are expected to experience significant
effects as the result of a regulatory action, and generally focused on medium- to long-run effects (U.S.
EPA, 2024f).
6.7.2 Considering Compliance and Enforcement
Evidence suggests that compliance with environmental regulations can vary widely across sources in
ways that exacerbate pre-existing disparities (e.g., Fedinick et al., 2019; Allaire et al., 2018; McDonald
and Jones, 2018; Balazs et al., 2012;). Analysts may want to consider whether regulated sources have
a history of significant non-compliance or enforcement actions taken against them under various
statutes or how capacity for monitoring and enforcement may differ across communities, including
those on Tribal lands. Past compliance issues may indicate pre-existing EJ concerns that warrant
further investigation.117
Analysts are encouraged to consider differences in compliance and ease of enforcement across
regulatory options in the EJ analysis. When there are pre-existing differences in risk or exposure,
options consistent with applicable law that improve monitoring coverage or encourage compliance can
reduce exposure in communities with EJ concerns (e.g., enhanced reporting requirements for higher
risk sources). Collecting, processing, and making publicly available real-time monitoring or remotely
sensed data may also be effective for enhancing public awareness and participation (U.S. EPA,
2021d).118
116 Data on household consumption patterns are available from the Consumer Expenditure Survey. The baseline distribution of
electricity and energy prices by household type is available from the Energy Information Administration. In addition, industry-
specific data on household consumption patterns may be available for specific products or services related to the regulatory
action. When such disaggregated data are available, they are less likely to differentiate by race and ethnicity, though they may
be available by income.
117 There is also a literature that explores whether the intensity of enforcement activities for environmental regulations varies
with demographics such as race and income (Konisky et al., 2021; Shadbegian and Gray, 2012).
118 For example, the oil and gas sector climate review rule (U.S. EPA, 2023h) allows certified third parties to report methane
leaks from oil and gas sources using advanced detection technologies, which the EPA makes publicly available. Likewise, the
2015 Petroleum Refinery Sector Risk and Technology Review and New Source Performance Standards created a dashboard
for benzene concentrations from refinery fenceline monitoring. See Textbox 2.2 and:
https://awsedap.epa.gov/public/extensions/Fenceline Monitorinq/Fenceline Monitorinq.html?sheet=MonitorinqDashboard
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
82
-------
6.7.3 Other Effects and Considerations
While this technical guidance mainly focuses on tools that analysts may use to evaluate differences in
health effects across population groups of concern, the distribution of non-health effects associated
with environmental stressors affected by the regulatory action may also be important to consider. For
instance, certain population groups may place a higher value on a cultural resource (e.g., spiritual or
sacred sites). If a regulatory option affects those resources, then the groups with a higher value will
experience a different effect than groups that do not place a value on the cultural resource. Likewise,
some regulatory options may differentially affect access to specific recreational activities for some
population groups.
Quantifying changes in non-health effects may be challenging. Often, data on the distribution of
baseline conditions for non-health effects are not easily available or are difficult to quantify, and/or are
not suitable for analyzing the effects of a regulatory action. For instance, data on some ecosystem
services (e.g., cultural uses of specific ecosystems) in the United States are quite limited in availability
compared to baseline health data, such as mortality incidence. Likewise, data and models to assess
how various regulatory options affect non-health related endpoints may not be available.
When the distribution of non-health effects is difficult to quantify, a case study approach may more
easily accommodates qualitative sources of information or allow for enhanced meaningful engagement
to gain better understanding of how these endpoints may be affected by the regulatory action (see
Section 5.3.1.2). For example, analysts may note any non-health effects associated with specific
cultural practices for population groups of concern, discuss how they are distributed across population
groups in the baseline, and describe how they may be affected by the regulatory action under
consideration when feasible.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
83
-------
Chapter 7: Research Priorities to Fill Key
Data and Methodological Gaps
High quality, scientific peer-reviewed data, methods, tools, and findings are necessary to support the
conclusions drawn from prospective analyses of EJ concerns, including information about population
group characteristics, environmental exposures and contaminants, routes of exposure, health
outcomes, and current and past experiences and concerns within communities. Keeping up with current
data and state-of-the-art methodologies to inform analyses of EJ concerns is of critical importance to
the EPA.119
For the purposes of identifying research priorities related to the intersection of EJ and regulatory
actions, data gaps include cases where data (including environmental and demographic or
socioeconomic data) are missing all together or limited in scope. For example, the spatial or temporal
resolution of the data may be insufficient, data may be unavailable or inaccessible due to privacy
issues, or no data have been collected to date. Methodological gaps occur when the current peer-
reviewed literature does not point to an established method or existing published methods are
insufficient to evaluate a specific question. The EPA program offices conducting EJ analyses are
interested in addressing data and methodological gaps to improve their analyses and advance
equitable outcomes. The intensity of effort and amount of time needed to address important research
gaps depend upon factors such as the complexity of the issue being studied and the extent to which
conventional or new research techniques are required. Note that the specific nature of some data and
method gaps will also vary across programs and regulatory contexts.
Developing research plans and strategies to address the breadth of EJ issues is an iterative process
requiring multiple levels of public engagement. This section provides a summary of data and
methodological gaps identified through brainstorming sessions (i.e., virtual focus groups) with EPA
program office management and staff who write or inform the development of regulatory actions and
through input from the SAB, the public, and Tribes on research needs and priorities relevant to the
analysis of EJ concerns for regulatory actions (See Text Box 7.1). Together, recommendations from
these groups will help the EPA understand and identify research priorities related to data, methods,
tools, and information for assessing EJ concerns in regulatory analysis.120
119 For more information on EPA efforts to use rigorous and relevant data and methods to support policy and decision making
as part of the Agency's implementation of the Foundations for Evidence-Based Policymaking Act of 2018, refer to the EPA's
Evaluation and Evidence-Building Policy: https://www.epa.qov/evaluate/evidence-act.
120 In developing research plans and strategies responsive to the data and methods gaps identified by the public, Tribes, and
the SAB for the conduct of EJ analysis for rulemakings, the broad set of science, data, and research gaps identified by the
Environmental Justice Subcommittee of the National Science and Technology Council to help advance the EJ goals is also
relevant (NSTC, 2024).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
84
-------
Text Box 7.1 Main Categories of Identified Data and Methodological Needs
1. Building consistency in terminology and definitions
2. Enhancing consideration of EJ in human health risk assessments
3. Cross-cutting analytic issues
4. Meaningful engagement and risk communication
5. Incorporating EJ into regulatory analysis
7.1 Building Consistency in Terminology and Definitions
While the EPA has operationalized consistent, agency-wide definitions for many key EJ-related terms,
EPA program office participants expressed a need for guidance on how to interpret terminology when
research areas or peer-reviewed studies use differing definitions. Participants also identified common
terminology and definitions for risk communication as an important need. Precise terminology and
definitions underpin the EPA's risk analyses and provides an important basis for ensuring that clear and
consistent terminology is used in incorporating EJ concerns into regulatory analysis.
7.2 Enhancing Consideration of EJ in Human Health Risk Assessment
The following sections describe identified research priorities relevant to different aspects of HHRA:
planning and scoping and problem formulation, effects assessment, exposure assessment, and risk
characterization.
7.2.1 Planning and Scoping and Problem Formulation
Problem formulation is a process for generating and evaluating preliminary hypotheses about why
health effects may be associated with specific stressors. Program office participants highlighted a need
for research to evaluate the impact of adopting various assumptions and definitions when considering
EJ at these early planning stages. For example, it would be valuable to understand the implications of
using different income thresholds to identify low-income populations on variability in establishing
reference populations for the risk assessment.
7.2.2 Effects Assessment
Effects assessment includes both hazard identification and dose-response assessment. Hazard
identification is the process of identifying the type of hazard to human health (e.g., cancer, birth defects)
posed by the exposure of interest. In an EJ context, one can ask, "What adverse health outcomes may
be caused by the pollutant(s) and how might this response vary by population group?" Dose-response
assessment addresses the relationship between the exposure or dose of a contaminant and the
occurrence of specific health effects. In an EJ context, analysts can ask, "What adverse health
outcomes exist at different exposures, and do they vary by type or incidence in population groups of
concern?"
Identified data and methodological needs for effects assessment focus on better understanding the
causal relationships between demographic and socioeconomic characteristics or other relevant factors
and adverse health outcomes associated with environmental stressors, and how these might vary
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
85
-------
across population groups. Program office participants and members of the public also identified the
need to develop methods that integrate community characteristics, social conditions, and cultural
influences with environmental stressors into risk assessments. For example, current data indicate that
communities with EJ concerns may be exposed to a greater number and quantity of environmental
stressors based on their proximity to waste sites, landfills, congested roadways, and manufacturing
facilities. Such communities may experience co-exposure to multiple contaminants that may contribute
to variability in individual responses. Program office participants also identified a need to incorporate
lifestages and pre-existing conditions, including allostatic load (cumulative stress effects) and changes
to DNA structure that can result from exposure to environmental stressors, into risk assessment.
Risk assessment uses a variety of dose-response models and tools to estimate the dose or
concentration relationships for adverse health outcomes. Program office participants highlighted the
need to ensure that dose-response modeling accounts for differences in vulnerability associated with
specific population groups, especially for infants and children. An important first step would be to
produce a comprehensive review of each relevant dose-response function that includes an analysis of
baseline risk variation across different population groups. This information would enable risk analysts to
consider the range of population-specific risk distributions along the dose-response curve.
7.2.3 Exposure Assessment
EPA program office participants noted a need to better understand actual exposures rather than relying
on standard models of fixed behavior. A critical area for research is improving cumulative risk
assessment and cumulative impact assessment methods for evaluating exposure to multiple
environmental and non-environmental stressors. Analysts also identified the need to delineate more
clearly between measures of cumulative impacts, measures of exposure, and/or indicators of risk to
communities experiencing multiple and layered stressors. In addition, program office participants and
Tribes emphasized continuing to develop ways in which community input regarding its values and
traditions can inform a cumulative assessment.
Tribes, the SAB, and program office participants highlighted the need for higher quality land use and
environmental exposure data, especially:
• Exposures people face when they are outside the home (e.g., the workplace, especially for
outdoor workers; commuting; schools; recreational activities; or senior centers) via activity diary
datasets or household surveys;
• More spatially resolved and comprehensive water pollution and drinking water quality data,
including from private wells;
• Real-time monitoring data, including measurements of routine emissions and emergency events
that impact air and water quality, especially in non- or under-monitored areas, such as on Tribal
lands; and
• Longer-term chronic, intergenerational, and legacy exposures.
7.2.4 Risk Characterization
The final step in risk assessment is characterizing risk. Risk characterization strives to provide a clear
discussion of the overall findings, key areas of uncertainty, overall data quality, and data limitations that
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
86
-------
may affect methodology development and the overall conclusion. EPA program office participants and
members of the public identified several data and methodological needs in this area:
• How baseline risk varies by demographic or socioeconomic group and lifestage;
• How activity patterns change over time as individuals age or move to inform future-year
projections for exposure and risk;
• Better incorporation of epidemiological data when assessing the health risks of pesticides; and
• Demographic and socioeconomic data for workers in specific locations (including access to
personal protective equipment), small business owners, and owners of regulated facilities.
7.3 Cross-Cutting Analytic Needs
Beyond characterizing the risk paradigm, EPA program office participants, Tribes, and members of the
public identified several other cross-cutting analytic needs that have bearing on EJ-related analyses.
They emphasized the importance of filling data gaps on topics including biomonitoring data, drinking
water service boundaries, and fish and game consumption data. Program office participants, Tribes,
and the SAB also highlighted the need for more spatially and temporally resolved human health
(including on intrinsic and extrinsic factors) and environmental sampling (e.g., water and soil) data.
The following are cross-cutting considerations that have bearing on filling data gaps, especially for
more spatially resolved data:
• While race, ethnicity, and income characteristics are often used to characterize communities of
concern, program office participants asked for research into ways to characterize vulnerable
communities more broadly, both in the context of potential effects from climate change and from
other environmental stressors, at a spatially disaggregated scale.
• Thoughtfully standardized approaches to data collection and analytic methods for more spatially
resolved data may require coordination across EPA program offices and with other federal
agencies, external researchers and practitioners, and other relevant stakeholders to limit
duplication and ensure that the data serve the needs of as many users as possible.
• There may also be benefits to clear and consistent definitions of key spatial features and their
representation in the spatial datasets used to assess EJ concerns. For example, differences
arise when representing ports as points (i.e., a location on a map represented as a zero-
dimensional point) vs. polygons (i.e., an area on a map represented as a two-dimensional
shape, such as a site footprint).
• Another identified data need for conducting risk evaluations and EJ analyses is improved
understanding of how population demographics change over time. One aspect is how the U.S.
population will geographically shift overtime, including migration patterns due to climate
change, and how this will impact health outcomes at both the national and regional level. A
second aspect is how population and/or exposure dynamics change across various time frames
and lifestages (e.g., due to aging).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
87
-------
• EPA program office participants raised the question of how to balance the need for improved
spatial resolution to adequately consider variability in exposure and health effects (e.g.,
biomarker data) with maintaining anonymity and protecting individuals from accidental
disclosure of personally identifiable information.
• Generating more spatially resolved estimates requires running models with higher-resolution
data, which can result in computational challenges (e.g., requires significant computer memory,
data storage, expertise, or computation time) that are sometimes costly to address.
• It would be useful to establish mechanisms for data standardization and transparency and
provide guidance on alternative data sources or methodologies when comprehensive data are
not available. Additionally, members of the public asked that the EPA create a centralized data
repository for comprehensive and standardized data.
• As emerging methods for combining remotely sensed data with process-based models become
more widely available, there is a need to better understand their data quality implications when
incorporating them into EJ analysis.
7.4 Meaningful Engagement and Risk Communication
As emphasized in Chapters 2 and 5, meaningful engagement and community outreach throughout the
policy process are integral to the consideration of EJ concerns. EPA program office participants and
Tribes pointed to the need for research on appropriate ways to collect and use community-generated
information in the EPA's regulatory analyses, including data collected by community groups, Indigenous
Knowledge (also referred to as Traditional Ecological Knowledge), and other information that reflects
the lived experiences of communities. These data are typically community-generated and more
qualitative in nature. Members of the public also emphasized the importance of developing guidance on
how to incorporate Indigenous Knowledge into analysis in collaboration with Indigenous and Tribal
communities.
Program office participants also expressed a desire for more assistance on effective public
engagement and communication about EJ concerns throughout the rulemaking process (e.g., how to
best communicate about individual regulatory actions as part of a larger and more dynamic set of policy
goals and how to measure the effectiveness of communication approaches on a continuous basis).
Both Tribes and the SAB identified the need for better communication of the risks and effects
associated with pollution exposure and non-environmental stressors to population groups of concern.
7.5 Incorporating EJ into Regulatory Analysis
Data and methodological needs for incorporating EJ concerns into regulatory analysis are discussed in
this section.
7.5.1 Data Gaps for Evaluating EJ Concerns
Often, data most relevant to EJ analysis are not sufficiently disaggregated by race, ethnicity, income, or
other demographic or socioeconomic characteristics of interest or available at a more spatially granular
level, which is necessary to better understand the distributional effects of a particular regulatory action.
EPA program office participants and Tribes identified several specific data gaps:
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
88
-------
• Finer resolution air quality data and alternative ways to collect them;
• Spatially granular household information on access to income-based government programs that
may offset some distributional effects (e.g., Supplemental Nutrition Assistance Program
assistance, utilities assistance);
• Delineation of water and other utility service areas that may affect the type or amount of
assistance received (e.g., healthy homes inspections and modifications, such as radon
remediation);
• How product use varies with demographic and socioeconomic characteristics; and
• Data on subsistence fishers, where they live, and their fish consumption behavior.
Collaboration with other federal agencies and other stakeholders to facilitate the sharing and access to
data sources was also identified as a need. Currently, access to data collected by other federal
agencies, Tribes, states, or local governments, universities, and non-government organizations varies.
7.5.2 Methodological Gaps for Evaluating EJ Concerns in the Baseline and for
Regulatory Options
Program office participants and members of the public indicated a need to continue to improve
characterization of relevant pre-existing conditions into the baseline for EJ analysis. Program office
participants also pointed to a lack of methodological tools to account for behavioral responses to a
regulatory action when analyzing their distributional effects. This research gap is likely broader than just
analyses of EJ concerns but can be particularly important for understanding who is ultimately affected
by the regulatory action. For example, program office participants identified the need to better
understand the adaptive behavior of lower-income households (e.g., purchase of air filters or bottled
water) and improved approaches for evaluating affected worker and employment effects. Program
office participants also identified improved modeling of the incidence of price changes faced by
households that vary with respect to key demographic and socioeconomic characteristics as a need for
evaluating the effects of a regulatory action.
To better evaluate the effects a regulatory action, program office participants expressed a need for
dose-response curves that vary by demographic and socioeconomic characteristics; information on
how to consider exposures during critical lifestages, such as childhood; and the link between genetic
factors or behaviors that could give rise to greater vulnerability. Another frequently noted
methodological gap was how to incorporate non-environmental stressors into the analysis and
consideration of the cumulative effects of exposure to multiple stressors on human health and well-
being.
EPA program office participants and members of the public also noted a need for research into
methods to capture EJ concerns in specific regulatory contexts. For example, it is not always clear how
to analyze EJ concerns for global pollutants; mobile sources; or ubiquitous chemicals where it is difficult
to characterize exposures due to their wide commercial use, persistence, and accumulation of these
chemicals overtime. In addition, program office participants identified the need for improved
approaches for analyzing the EJ implications of regulations that indirectly affect health by affecting
which chemicals are used in manufacturing and production or how information is provided.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
89
-------
Program office participants also identified as a methodological gap how to leverage qualitative analysis
in assessing environmental justice in regulatory analysis. Mixed methods approaches and qualitative
case studies can be time- and resource-intensive and often focus on narrow applications in a specific
setting or place. Additional research is needed to understand how these methods may be leveraged for
regional and national scale applications.
7.5.4 Other Analytic Considerations
Program office participants expressed a need for more spatially granular Census data to better align
with the spatial resolution of certain environmental sampling data. Given the aggregate nature of the
publicly available U.S. Census data (e.g., the ACS), analyses do not report information on how
demographic and socioeconomic characteristics vary within a given administrative unit (e.g., within a
block group).
Program office participants also identified the need to investigate downstream chemical effects relevant
to evaluating EJ concerns and potential risk mitigation options. Finally, program office participants
noted the importance of better understanding the implications of low probability, but high consequence,
events (e.g., accidental releases).
7.6 Other General Needs
EPA program office participants and members of the public identified the need for enhanced cross-
agency coordination and information sharing. Specifically, they expressed interest in sharing best
practices across EPA offices through a methods or resource library to address common EJ analytic
questions, such as:
• Stratification of effects by demographic and socioeconomic characteristics;
• Best practices for proximity-based and other types of quantitative analysis;
• Expanded use of additional surrogate or proxy variables for identifying populations of interest
(e.g., pre-existing health conditions, critical service gaps);
• Understanding population-level risk; and
• Measuring and modeling cumulative impacts/exposure/risk to communities.
Program office participants also expressed a need for in-house technical support and/or training for
addressing specific EJ methodological gaps in their work.
7.7 Next Steps
The EPA is a science-based agency. As such, it is committed to the pursuit of research related to
advancing EJ analysis for regulatory actions to better meet the needs of Agency analysts, decision-
makers, and the public in support of scientifically sound regulatory decisions that protect the health of
all communities.
Each EPA program office engages in research to address specific needs and concerns, including those
related to EJ concerns. In addition, the Office of Policy works jointly with program offices to craft, review
and promulgate regulations that protect human health and the environment and ensure they comply
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
90
-------
with key statutes and Executive Orders. It also helps to promote analytic consistency and rigor across
the Agency's regulatory portfolio and conducts research to develop improved methods for measuring
the effects of regulation, including those related to EJ concerns.
The EPA's Office of Environmental Justice and External Civil Rights was launched in October 2022 to:
• Support communities by providing grants and technical assistance;
• Ensure equity, EJ, and civil rights are incorporated into EPA's policies and programs;
• Ensure compliance and enforcement of federal civil rights laws; and
• Provide conflict prevention and resolution on environmental issues.
The Office of Environmental Justice and External Civil Rights serves as a resource for EPA program
offices to strengthen consideration of environmental justice during the rulemaking process. In addition,
it develops and improves tools, such as EJScreen, and works along-side partners to assess the
community-level impacts of the agency's activities and identify communities with environmental justice
concerns.
The EPA's Office of Research and Development (ORD) also actively pursues and supports
environmental justice research. It developed an EJ Research Roadmap to highlight the role of ORD
science in addressing EJ concerns (U.S. EPA, 2016d). ORD EJ research is publicly available at
https://www.epa.gov/ei-research.
Building on this roadmap, the ORD also developed a series of actionable recommendations for
cumulative impacts research, including: establishing the decision context for cumulative impact
assessment with meaningful public engagement, addressing scientific considerations for meeting
community needs with holistic and fit-for-purpose approaches with the support of different screening
and decision tools, empowering local decisions and actions through participatory science, supporting
science translation and delivery to meet community needs, and providing research management and
support (e.g., coordinating with other EPA offices and external partners) (U.S. EPA, 2022c). The
document also includes a summary of current research areas and an overview of research gaps and
barriers identified by ORD (many of which are consistent with the topic areas described in this chapter)
that informed research conducted under the fiscal year 2023-2026 Strategic Research Action Plans.121
121 The ORD Fiscal Year 2023-2026 Strategic Research Action Plans are available at: https://www.epa.gov/research/strateaic-
research-action-plans-fiscal-vears-2023-2026.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
91
-------
References
Abadie, A., S. Athey, G. W. Imbens, and J. M. Wooldridge. 2023. When should you adjust standard
errors for clustering? The Quarterly Journal of Economics 138(1): 1-35.
Arcury, T., H. Chen, T. Arnold, S. Quandt, K. Anderson, R. Scott, J. Talton, and S. Daniel. 2021.
Pesticide exposure among Latinx child farmworkers in North Carolina. American Journal of
Industrial Medicine 64: 602—619.
Aker A., R. McConnell, R. Loch-Caruso, S. Park, B. Mukherjee, and others. 2020. Interactions between
chemicals and non-chemical stressors: The modifying effect of life events on the association
between triclocarban, phenols and parabens with gestational length in a Puerto Rican cohort.
Science of the Total Environment 708: 134719.
Akinbami, L., A.E. Simon, and L.M. Rossen. 2016. Changing trends in asthma prevalence among
children. Pediatrics 137(1): 1-7.
Allaire, M., H. Wu, and U. Lall. 2018. National trends in drinking water quality violations. Proceedings of
the National Academy of Sciences 115(9): 2078-2083.
Americans with Disabilities Act of 1990, 42 U.S.C. § 12101 (1990). Retrieved from:
https://www.ada.gov/pubs/adastatute08.htm.
Andarge, T., Y. Ji, B. Keeler, D. Keiser, and C. McKenzie. 2024. Environmental justice and the Clean
Water Act: Implications for economic analyses of clean water regulations. Environmental and
Energy Policy and the Economy 5: 70-126.
Ash, M., J.K. Boyce, J.G. Chang, and H. Scharber. 2013. Is environmental justice good for White folks?
Industrial air toxics exposure in urban America. Social Science Quarterly 94(3): 616-636.
Ash, M., and J.K. Boyce. 2018. Racial disparities in pollution exposure and employment at US industrial
facilities. PNAS 115(42): 10636-10641
Backhaus T., and M. Faust. 2012. Predictive environmental risk assessment of chemical mixtures: A
conceptual framework. Environmental Science & Technology 46: 2564-2573.
Baden, B.M., and D. Coursey. 2002. The locality of waste sites within the city of Chicago: A
demographic, social, and economic analysis. Resource and Energy Economics 24: 53-93.
Baden, B.M., D.S. Noonan, and R.M. Turaga. 2007. Scales of justice: Is there a geographic bias in
environmental equity analysis? Journal of Environmental Planning and Management 50(2):
163-85.
Bakkensen, L., and L. Ma. 2020. Sorting over flood risk and implications for policy reform. Journal of
Environmental Economics and Management 104: 102362.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
92
-------
Bakkensen, L., Ma, L. Muehlenbachs, and L. Benitez. 2024. Cumulative impacts in environmental
justice: Insights from economics and policy. Regional Science and Urban Economics 107:
103993
Balazs, C. L., R. Morello-Frosch, A.E Hubbard and I. Ray. 2012. Environmental justice implications of
arsenic contamination in California's San Joaquin Valley: a cross-sectional, cluster-design
examining exposure and compliance in community drinking water systems. Environmental
Health 11(1): 1-12.
Banzhaf, S., L. Ma, and C. Timmins. 2019a. Environmental justice: The economics of race, place, and
pollution. Journal of Economic Perspectives 33(1): 185-208.
Banzhaf, S., L. Ma, and C. Timmins. 2019b. Environmental justice: Establishing causal relationships.
Annual Review of Resource Economics 11(1): 377-398.
Barzyk, T., S. Wilson, and A. Wilson. 2015. Community, state, and federal approaches to cumulative
risk assessment: Challenges and opportunities for integration. International Journal of
Environmental Research and Public Health 12(5): 4546-4571.
Bekkar, B., S. Pacheco, R. Basu, and N. DeNicola. 2020. Association of air pollution and heat exposure
with preterm birth, low birth weight, and stillbirth in the US: A systematic review. JAMA Network
Open 3(6): e208243.
Bouvier, R. 2014. Distribution of income and toxic emissions in Maine, United States: Inequality in two
dimensions. Ecological Economics 102: 39-47.
Bowen, W. 2001. Environmental Justice through Research-Based Decision-Making. New York:
Garland.
Braverman, P.A., E. Arkin, D. Proctor, T. Kauh, and N. Holm. 2022. Systemic and structural racism:
definitions, examples, health damages, and approaches to dismantling. Health Affairs 41(2):
171-178.
Brender, J.D., J.A. Maantay, and J. Chakraborty. 2011. Residential proximity to environmental hazards
and adverse health outcomes. American Journal of Public Health 101 (S1): S37-S52.
Buchanan, S., L. Targos, K. Nagy, K. Kearney, and M. Turyk. 2015. Fish consumption and hair mercury
among Asians in Chicago. Journal of Occupational and Environmental Medicine 57(12): 1325-
30.
Bullard, R.D., P. Mohai, R. Saha, and B. Wright. 2008. Toxic Wastes and Race at Twenty: 1987-2007
Grassroots Struggles to Dismantle Environmental Racism in the United States. Cleveland, OH:
United Church of Christ Justice and Wtness Ministries.
Burger, J., and M. Gochfeld. 2011. Conceptual environmental justice model for evaluating chemical
pathways of exposure in low-income, minority, Native American, and other unique exposure
populations. American Journal of Public Health 101 (S1): S64-73.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
93
-------
Cain, L., D. Hernandez-Cortes, C. Timmins, and P. Weber. 2024. Recent findings and methodologies in
economics research in environmental justice. Review of Environmental Economics and Policy
18(10): 116-142.
California Environmental Protection Agency. 2015. Air Toxics Hot Spots Program - Risk Assessment
Guidelines, Sacramento, CA: California Environmental Protection Agency, Office of
Environmental Health Hazard Assessment. Retrieved from:
http://oehha.ca.gov/air/hot spots/hotspots2015.html.
Cameron, A. C., and D.L. Miller. 2015. A practitioner's guide to cluster-robust inference. Journal of
Human Resources 50(2): 317-372.
Cameron, T., G. Crawford, and I. McConnaha. 2012. Superfund taint and neighborhood change:
Ethnicity, age distributions, and household structure. In Banzhaf, S. (Ed.), The Political
Economy of Environmental Justice. Stanford University Press, pp. 137-169.
Carlin, B.P., and H. Xia. 1999. Assessing environmental justice using Bayesian hierarchical models:
two case studies. Journal of Exposure Analysis and Environmental Epidemiology 9(1): 66-78.
Carroll S, I. Garba, O. Figueroa-Rodriguez, J. Holbrook, R. Lovett, S. Materechera, M. Parsons, K.
Raseroka, D. Rodriguez-Lonebear, R. Rowe, and R. Sara. 2020. The CARE principles for
Indigenous data governance. Data Science Journal 19: 1-12.
Cattaneo, I., A. Kalian, M. Di Nicola, B. Dujardin, S. Levorato, and others. 2023. Risk assessment of
combined exposure to multiple chemicals at the European Food Safety Authority: Principles,
guidance documents, applications and future challenges. Toxins 15(1): 40.
Centers for Disease Control and Prevention (CDC). 2022. National Report on Human Exposure to
Environmental Chemicals. Updated. Atlanta, GA: Centers for Disease Control and Prevention,
U.S. Department of Health and Human Services. Retrieved from:
https://www.cdc.gov/exposurereport/
Chakraborty, J. 2020. Unequal proximity to environmental pollution: An intersectional analysis of people
with disabilities in Harris County, Texas. Professional Geographer 72(4): 521-534.
Chakraborty J., T.W. Collins, and S.E. Grineski. 2016. Environmental justice research: Contemporary
issues and emerging topics. International Journal of Environmental Research and Public Health.
13(11): 1072.
Chakraborty, J. 2011. Revisiting Tobler's first law of geography: Spatial regression models for
assessing environmental justice and health risk disparities. In: Maantay, J. McLafferty, S. (Eds.),
Geospatial Analysis of Environmental Health. Geotechnologies and the Environment, vol 4.
Springer-Dordrecht, pp. 337-356.
Chakraborty, J., and J. Maantay. 2011. Proximity analysis for exposure assessment in environmental
health justice research. In J. Maantay and S. McLafferty (Eds.), Geospatial Analysis of
Environmental Health. Geotechnologies and the Environment, vol 4. Springer-Verlag, pp. 111—
138.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
94
-------
Chakraborty, J., J. A. Maantay, and J. Brender. 2011. Disproportionate proximity to environmental
health hazards: Methods, models, and measurement. American Journal of Public Health
101 (S1): S27-36.
Christensen, G., Z. Li, J. Pearce, M. Marcus, J. Lah, L. Waller, S. Ebelt, and A. Huls. 2022. The
complex relationship of air pollution and neighborhood socioeconomic status and their
association with cognitive decline. Environment International 167: 107416.
Chun,Y., Y. Kim, and H. Campbell. 2012. Using Bayesian methods to control for spatial autocorrelation
in environmental justice research: An illustration using Toxics Release Inventory data for a
sunbelt county. Journal of Urban Affairs 34(4): 419-439.
Citro, C.F., and R.T. Michael (Eds.). 1995. Measuring Poverty: A New Approach. Washington, DC:
National Academy Press. Retrieved from: http://www.nap.edu/catalog/4759/measuring-poverty-
a-new-approach.
Clougherty, J., J. Shmool, and L. Kubzansky. 2014. The Role of non-chemical stressors in mediating
socioeconomic susceptibility to environmental chemicals. Current Environmental Health Reports
1: 302-313.
Clougherty J. and C. Rider. 2020. Integration of psychosocial and chemical stressors in risk
assessment. Current Opinion in Toxicology 22: 25-29.
Colmer, J, I. Hardman, J. Shimshack, and J. Voorheis. 2020. Disparities in PM2.5 air pollution in the
United States. Science 369: 575-578.
Congressional Research Service. 2022. The Supplemental Poverty Measure: Its Core Concepts,
Development, and Use. Report R45031. Retrieved from:
https://crsreports.congress.goV/product/pdf/R/R45031.
Connolly, M. and B. Jacobs, 2020. Counting Indigenous American Indians and Alaska Natives in the
US Census. Statistical Journal of the IAOS 36(1): 201-210.
Corburn, J. 2002. Environmental justice, local knowledge, and risk: The discourse of a community-
based cumulative exposure assessment. Environmental Management 29(4): 451-466.
Corlin, L., T. Rock, J. Cordova, M. Woodin, J. L. Durant, D. Gute, J. Ingram, and D. Brugge. 2016.
Health effects and environmental justice concerns of exposure to uranium in drinking water.
Current Environmental Health Reports 3: 434-442.
Couch, S. R., and C.J. Coles. 2011. Community stress, psychosocial hazards, and EPA decision-
making in communities impacted by chronic technological disasters. American Journal of Public
Health 101(S1): S140-S148.
Council on Environmental Quality (CEQ). 1997. Environmental Justice: Guidance Under the National
Environmental Policy Act. Washington, D.C.: Executive Office of the President. Retrieved from:
https://www.epa.gov/environmentaliustice/ceg-environmental-iustice-guidance-under-national-
environmental-policy-act.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
95
-------
Currie, J., J. Graff Zivin, K. Meckel, M. Neidell, and W. Schlenker. 2013. Something in the water:
Contaminated drinking water and infant health. Canadian Journal of Economics 46(3): 791-810.
Cushing, L., D. Blaustein-Rejto, M. Wander, M. Pastor, J. Sadd, A. Zhu, and R. Morello-Frosch. 2018.
Carbon trading, co-pollutants, and environmental equity: Evidence from California's cap-and-
trade program (2011-2015). PLoS Medicine15(7): e1002604.
Deere, B.P., and K.C. Ferdinand. 2020. Hypertension and race/ethnicity. Current Opinion in Cardiology
35(4): 342-350.
Di, Q., I. Kloog, P. Koutrakis, A. Lyapustin, Y. Wang, and J. Schwartz. 2016. Assessing PM2.5
exposures with high spatiotemporal resolution across the continental United States.
Environmental Science & Technology 50(9): 4712-4721.
Di, Q., Y. Wang, A. Zanobetti, Y. Wang, P. Koutrakis, C. Choirat, F. Dominici, and J.D. Schwartz. 2017.
Air pollution and mortality in the Medicare population. New England Journal of Medicine
376(26): 2513-2522.
Di, Q., H. Amini, L. Shi, I. Kloog, R. Silvern, and others. 2019. An ensemble-based model of PM2.5
concentrations across the contiguous United States with high spatiotemporal resolution.
Environment International 130: 104909.
Di Fonzo, D., A. Fabri, and R. Pasetto. 2022. Distributive justice in environmental health hazards from
industrial contamination: A systematic review of national and near-national assessments of
social inequalities. Social Science & Medicine 297: 114834.
Ding, N., C. A. Karvonene-Gutierrez, A.R. Zota, B. Mukherjee, S.D. Harlow, and S.K. Park. 2023. The
role of exposure to per- and polyfluoroalkyl substances in racial/ethnic disparities in
hypertension: Results from the study of women's health across the nation. Environmental
Research 227(15): 115813.
Dosemagen, S., and E. Williams. 2022. Data usability: The forgotten segment of environmental data
workflows. Frontiers in Climate 4: 785269.
Elliott, J. R., and K. T. Smiley. 2019. Place, space, and racially unequal exposures to pollution at home
and work. Social Currents 6(1): 32-50.
Erickson, M., C. Brown, E. Tomaszewski, J. Ayotte, J. Bohlke, D. Kent and S. Qi. 2024. Prioritizing
water availability study settings to address geogenic contaminants and related societal factors.
Environmental Monitoring and Assessment 196: 303.
Estien, C., C. Wlkinson, R. Morello-Frosch, and C. Schell. 2024. Historic redlining is associated with
disparities in environmental quality across California. Environmental Science & Technology
Letters 11(2): 54-59.
Executive Order (E.O.)12866: Regulatory Planning and Review. Federal Register 58:190 (4 October
1993) p. 51735. Retrieved from: http://www.archives.gov/federal-reqister/executive-
orders/pdf/12866. pdf.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
96
-------
E.O. 12898: Federal Actions to Address Environmental Justice in Minority Populations and Low-Income
Populations. Federal Register 59:32 (16 February 1994) p. 7629. Retrieved from:
http://www.archives.gov/federal-reqister/executive-orders/pdf/12898.pdf.
E.O. 13175: Consultation and Coordination with Indian Tribal Governments. Federal Register 65:218 (9
November 2000) p. 67249. Retrieved from: https://federalreqister.qov/a/00-29003.
E.O. 14008: Tackling the Climate Crisis at Home and Abroad. Federal Register 2021-02177 (27
January 2021). Retrieved from: https://www.federalregister.gov/documents/2021/02/01/2Q21-
02177/tackling-the-climate-crisis-at-home-and-abroad.
E.O. 14091. Further Advancing Racial Equity and Support for Underserved Communities Through the
Federal Government. Federal Register 2023-03779 (16 February 2023). Retrieved from:
https://www.federalregister.gov/documents/2023/02/22/2023-Q3779/further-advancing-racial-
eguity-and-support-for-underserved-communities-through-the-federal.
E.O. 14094: Modernizing Regulatory Review 88:69 (6 April 2023). p. 21879. Retrieved from:
https://www.govinfo.gov/content/pkg/FR-2023-04-11/pdf/2023-0776Q.pdf.
E.O. 14096: Revitalizing Our Nation's Commitment to Environmental Justice for All. Federal Register
88:80 (21 April 2023). p. 23231. Retrieved from: https://www.govinfo.gov/content/pkg/FR-2023-
04-26/pdf/2023-08955.pdf.
Faber, D.R., and E.J. Krieg. 2002. Unequal exposure to ecological hazards: Environmental injustices in
the Commonwealth of Massachusetts. Environmental Health Perspectives 110(S2): 277-288.
Faber, D.R., and E.J. Krieg. 2005. Unequal Exposure to Ecological Hazards 2005: Environmental
Injustices in the Commonwealth of Massachusetts: A Report by the Philanthropy and
Environmental Justice Research Project. Boston, MA: Northeastern University. Retrieved from:
https://www.issuelab.org/resources/2980/2980.pdf.
Fan, Q., K. Fisher-Vanden, and H. Klaiber. 2018. Climate change, migration, and regional economic
impacts in the U.S. Journal of the Association of Environmental and Resource Economists 5(3):
643-671.
Fann, N., H. A. Roman, C.M. Fulcher, M.A. Gentile, B.J. Hubbell, K. Wesson, J.I. Levy. 2011.
Maximizing health benefits and minimizing inequality: Incorporating local-scale data in the
design and evaluation of air quality policies. Risk Analysis 31(6): 908-922.
Fedinick, K.P., S. Taylor, and M. Roberts. 2019. Watered Down Justice: Communities of Color and the
SDWA. Natural Resources Defense Council Report R-19-09-A.
Feng, D., J. Liu, K. Lawson, and C. Shen. 2022. Differentiable, learnable, regionalized process-based
models with multiphysical outputs can approach state-of-the-art hydrologic prediction accuracy.
Water Resources Research 58(10): e2022WR032404.
Fowlie, M., S. Holland, and E. Mansur. 2012. What do emissions markets deliver and to whom?
Evidence from Southern California's NOx trading program. American Economic Review 102(2):
965-993.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
97
-------
Fox, L. and K. Burns. 2021. The Supplemental Poverty Measure: 2020. Current Population Reports,
No. P60-275. Retrieved from:
https://www.census.gov/content/dam/Census/librarv/publications/2021/demo/p60-275.pdf.
Freudenberg, N., M. Pastor, and B. Israel. 2011. Strengthening community capacity to participate in
making decisions to reduce disproportionate environmental exposures. American Journal of
Public Health 101 (S1): S123-130.
Fullerton, D. 2011. Six distributional effects of environmental policy. Risk Analysis 31(6): 923-929.
Gee, G.C., and D.C. Payne-Sturges. 2004. Environmental health disparities: A framework integrating
psychosocial and environmental concepts. Environmental Health Perspectives 112(17): 1645-
1653.
Gilbert, A., and J. Chakraborty. 2011. Using geographically weighted regression for environmental
justice analysis: Cumulative cancer risks from air toxics in Florida. Social Science Research
40(1): 273-286.
Gilliland, F., W. Hunt, M. Pardilla, and C. Key, 2000. Uranium mining and lung cancer among Navajo
men in New Mexico and Arizona, 1969 to 1993. Journal of Occupational and Environmental
Medicine 42(3): 278-283.
Gochfeld, M., and J. Burger. 2011. Disproportionate exposures in environmental justice and other
populations: The importance of outliers. American Journal of Public Health 101 (S1): S53-63.
Gonzalez, D., Nardone, A., Nguyen, A., Morello-Frosch, R., and Casey, J. 2023. Historic redlining and
the siting of oil and gas wells in the United States. Journal of Exposure Science &
Environmental Epidemiology 33(1): 76-83.
Grainger, C. and T. Ruangmas. 2018. Who wins from emissions trading? Evidence from California.
Environmental and Resource Economics 71: 703-727.
Grineski, S.E. 2009. Human-environment interactions and environmental justice: How do diverse
parents of asthmatic children minimize hazards? Society and Natural Resources 22(8): 727-
743.
Grineski, S.E., and T.W. Collins. 2018. Geographic and social disparities in exposure to air
neurotoxicants at U.S. public schools. Environmental Research 161: 580-587.
Grove, M., L. Ogden, S. Pickett, C. Boone, G. Buckley, D. Locke, C. Lord, and B. Hall. 2017. The
legacy effect: Understanding how segregation and environmental injustice unfold overtime in
Baltimore. Annals of the American Association of Geographers 108(2): 524-537.
Haining, R. P. Spatial Data Analysis: Theory and Practice. Cambridge University Press, 2003.
Helm J., M. Nishioka, J. Brody, R. Rudel, and R. Dodson. 2018. Measurement of endocrine disrupting
and asthma-associated chemicals in hair products used by Black women. Environmental
Research 165: 448-458.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
98
-------
Hernandez-Cortes, D. and K. Meng. 2023. Do environmental markets cause environmental injustice?
Evidence from California's carbon market. Journal of Public Economics 217: 104786.
Hoyt, D., and L. Raun. 2015. Measured and estimated benzene and volatile organic carbon (VOC)
emissions at a major US refinery/chemical plant: Comparison and prioritization. Journal of the
Air& Waste Management Association 65(8): 1020-1031.
Hyland, C., and O. Laribi. 2017. Review of take-home pesticide exposure pathway in children living in
agricultural areas. Environmental Research 156: 559-570.
Iceland, J. 2003. Dynamics of Economic Weil-Being: Poverty 1996-1999. Current Population Reports,
pp. 70-91. Washington, DC: U.S. Census Bureau.
Ingram, J., L. Jones, J. Credo, and T. Rock. 2020. Uranium and arsenic unregulated water issues on
Navajo lands. Journal of Vacuum Science & Technology A 38(3): 031003.
Intergovernmental Panel on Climate Change (IPCC). 2014. Impacts, Adaptation, and Vulnerability. Part
A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge University Press,
Cambridge.
IPCC. 2018. Global Warming of 1.5°C. Cambridge University Press, Cambridge.
Jbaily, A., X. Zhou, J. Liu, T. Lee, L. Kamareddine, S. Verguet and F. Dominici. 2022. Air pollution
exposure disparities across US population and income groups. Nature 60: 228-233.
Jimenez, M., N. DeVille, E. Elliott, J. Schiff, G. Wit, J. Hart, and P. James. 2021. Associations between
nature exposure and health: A review of the evidence. International Journal of Environmental
Research and Public Health 18: 4790.
Josey, K., S. Delaney, X. Wu, R. Nethery, P. DeSouza, D. Braun, and F. Dominici. 2023. Air pollution
and mortality at the intersection of race and social class. New England Journal of Medicine
388(15): 1396-1404.
Kalweit, A., R. Herrick, M. Flynn, J. Spengler, J. Berko Jr, J. Levy, and D. Ceballos. 2020. Eliminating
take-home exposures: recognizing the role of occupational health and safety in broader
community health. Annals of Work Exposures and Health 64(3): 236-249.
Kenny, C., S. Kuriwaki, C. McCartan, E. Rosenman, T. Simko, and K. Imai. 2021. The use of differential
privacy for census data and its impact on redistricting: The case of the 2020 US Census.
Science Advances 7(41): eabk3283.
Kenny, C., C. McCartan, S. Kuriwaki, T. Simko, and K. Imai. 2024. Evaluating bias and noise induced
by the US Census Bureau's privacy protection methods. Science Advances 10(18): eadl2524.
Kim, Y., and Y. Chun. 2018. Revisiting environmental inequity in Southern California: Does
environmental risk increase in ethnically homogeneous or mixed communities? Urban Studies
56(9): 1748-1767.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
99
-------
Knapp, E.A., A.M. Kress, C.B. Parker, G.P. Page, K. McArthur, and others. 2023. The environmental
influences on child health outcomes (ECHO)-wide cohort. American Journal of Epidemiology
192(8): 1249-1263.
Konisky, D. M., C. Reenock, and S. Conley. 2021. Environmental injustice in Clean Water Act
enforcement: Racial and income disparities in inspection time. Environmental Research
Letters 16(8): 084020.
Kuminoff, N., C. Parmeter, and J. Pope. 2010. Which hedonic models can we trust to recover the
marginal willingness to pay for environmental amenities? Journal of Environmental Economics
and Management 60(3): 145-160.
Kuminoff, N., T. Schoellman, and C. Timmins. 2015. Environmental regulations and the welfare effects
of job layoffs in the United States: A spatial approach. Review of Environmental Economics and
Policy 9(2): 198-218.
Lall, U., T. Johnson, P. Colohan, A. Aghakouchak, C. Brown, G. McCabe, R. Pulwarty, and A.
Sankarasubramanian. 2018. Water, In Impacts, Risks, and Adaptation in the United States:
Fourth National Climate Assessment, Volume II. U.S. Global Change Research Program,
Washington, DC. pp. 145-173.
Lane, H.M., R. Morello-Frosch, J.D. Marshall, and J.S. Apte. 2022. Historic redlining is associated with
present-day air pollution disparities in U.S. cities. Environmental Science & Technology Letters
9(4): 345-350.
Letellier, N., S. Zamora, C. Spoon, J. Yang, M. Mortamais, G. Escobar, D. Sears, M. Jankowska, and
T. Benmarhnia. 2022. Air pollution and metabolic disorders: Dynamic versus static measures of
exposure among Hispanics/Latinos and non-Hispanics. Environmental Research 209: 112846.
Levy, J., S. Chemerynski, and J. Tuchmann. 2006. Incorporating concepts of inequality and inequity into
health benefits analysis. International Journal for Equity in Health 5(2).
Lewis, J., J. Hoover, and D. MacKenzie. 2017. Mining and environmental health disparities in Native
American communities. Current Environmental Health Reports 4: 130-141.
Lieberman, E.S. 2005. Nested analysis as a mixed d-method strategy for comparative research.
American Political Science Review 99(3): 435-452.
Liu, Y., and H. Eicher-Miller. 2021. Food insecurity and cardiovascular disease. Current Atherosclerosis
Reports 23(6): 24.
Maantay, J., A. Winner, and A. Maroko. 2022. Geospatial analysis of the urban health environment, in
Geospatial Technology for Human Weil-Being and Health. Cham: Springer International
Publishing, 151-183.
Maantay, J. 2001. Zoning, equity, and public health. American Journal of Public Health 91(7): 1033-
1041.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
100
-------
Maguire, K., and G. Sheriff. 2011. Comparing distributions of environmental outcomes for regulatory
environmental justice analysis. International Journal of Environmental Research and Public
Health 8: 1707-1726.
Mani A., S. Mullainathan, E. Shafir, and J. Zhao. 2013. Poverty impedes cognitive function. Science
341(6149): 976-80.
Mansur E., and G. Sheriff. 2021. On the measurement of environmental inequality: ranking emissions
distributions generated by different policy instruments. Journal of the Association of
Environmental and Resource Economists (8)4: 655-861.
Manuck, T.A. 2017. Racial and ethnic differences in preterm birth: A complex, multifactorial problem.
Semin Perinatol 41(8): 511-518.
Martinez-Morata, I., B. Bostick, O. Conroy-Ben, D. Duncan, M. Jones, M. Spaur, K. Patterson, S. Prins,
A. Navas-Acien and A. Nigra. 2022. Nationwide geospatial analysis of county racial and ethnic
composition and public drinking water arsenic and uranium. Nature Communications 13(1):
7461.
Mascarenhas, M., R. Grattet, and K. Mege. 2021. Toxic waste and race in twenty-first century America
Environment and Society 12(1): 108-126.
McDonald, Y.J., and N.E. Jones. 2018. Drinking water violations and environmental justice in the
United States, 2011-2015. American Journal of Public Health 108(10): 1401-1407.
McEwen, B.S., and P. Tucker. 2011. Critical biological pathways for chronic psychosocial stress and
research opportunities to advance the consideration of stress in chemical risk assessment.
American Journal of Public Health 101 (S1): S131-S139.
McHale C., G. Osborne, R. Morello-Frosch A.G. Salmon, M.S. Sandy, G. Solomon, L. Zhang, M.T.
Smith, and L. Zeise. 2018. Assessing health risks from multiple environmental stressors: Moving
from G x E to I x E. Mutation Research/Reviews in Mutation Research 775 (Jan.-March): 11-
20.
McPartland, J., R. Shaffer, M. Fox, K. Nachman, T. Burke, and R.A Denison. 2022. Charting a path
forward: Assessing the science of chemical risk evaluations under the Toxic Substances Control
Act in the context of recent National Academies recommendations. Environmental Health
Perspectives 130(2): 25003.
Melstrom, R., and R. Mohammadi. 2022. Residential mobility, brownfield remediation, and
environmental gentrification in Chicago. Land Economics 98(1): 62-77.
Mennis, J. and M. Heckert. 2017. Application of spatial statistical techniques. The Routledge Handbook
of Environmental Justice. Routledge, pp. 207-221.
Mohai, P., and B. Bryant. 1992. Race and the Incidence of Environmental Hazards. Westview Press,
Boulder, CO.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
101
-------
Mohai, P., Lantz, P. M., Morenoff, J., House, J. S., Mero, R. P. 2009. Racial and socioeconomic
disparities in residential proximity to polluting industrial facilities: Evidence from the Americans'
Changing Lives Study. American Journal of Public Health 99(3): S649-S656.
Mohai, P. and R. Saha. 2006. Reassessing racial and socioeconomic disparities in environmental
justice research. Demography 43(2): 383-399.
Mohai, P., and R. Saha. 2007. Racial inequality in the distribution of hazardous waste: A national level
reassessment. Social Problems 54(3): 343-370.
Mohai, P., and R. Saha. 2015. Which came first, people or pollution? A review of theory and evidence
from longitudinal environmental justice studies. Environmental Research Letters 10(12):
125011.
Morello-Frosch, R. and B.M. Jesdale. 2006. Separate and unequal: residential segregation and
estimated cancer risks associated with ambient air toxics in US metropolitan areas.
Environmental Health Perspectives 114(3): 386-393.
Morello-Frosch, R., and O.K. Obasogie. 2023. The climate gap and the color line—Racial health
inequities and climate change. New England Journal of Medicine 388(10): 943-949
Morello-Frosch, R., M. Zuk, M. Jerrett, B. Shamasunder, and A.D. Kyle. 2011.Understanding the
cumulative impacts of inequalities in environmental health: Implications for policy. Health Affairs
30(5): 879-887.
Mueller, J.T., and S. Gasteyer. 2021. The widespread and unjust drinking water and clean water crisis
in the United States. Nature Communications 12(1): 3544.
Muller, N., P. Matthews, and V. Wiltshire-Gordon. 2018. The distribution of income is worse than you
think: Including pollution impacts into measures of income inequality. Public Library of Science
One 13(3): e0192461.
Munoz-Pizza, D., M. Villada-Canela, M. Reyna, J. Texcalac-Sangrador, and A. Osornio-Vargas. 2020.
Air pollution and children's respiratory health: a scoping review of socioeconomic status as an
effect modifier. International Journal of Public Health 65: 649-660.
National Academies of Sciences, Engineering, and Medicine (NASEM). 2016. A Framework for
Educating Health Professionals to Address the Social Determinants of Health. Washington,
D.C.: National Academy Press.
NASEM. 2017. Communities in Action: Pathways to Health Equity. Washington, D.C.: The National
Academies Press.
NASEM. 2023. Building Confidence in New Evidence Streams for Human Health Risk Assessment:
Lessons Learned from Laboratory Mammalian Toxicity Tests. Washington, D.C.: National
Academy Press.
National Research Council (NRC). 1983. Risk Assessment in the Federal Government: Managing the
Process. Washington, D.C.: National Academy Press.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
102
-------
NRC. 1991. Environmental Epidemiology, Volume 1: Public Health and Hazardous Wastes.
Washington, D.C.: National Academy Press.
NRC. 2009. Science and Decisions: Advancing Risk Assessment. Washington, D.C.: National
Academy Press.
NRC. 2011a. Improving Health in the United States: The Role of Health Impact Assessment.
Washington, D.C.: National Academy Press.
NRC. 2011b. America's Climate Choices. Washington, D.C.: The National Academies Press.
National Science and Technology Council (NSTC). 2024. Environmental Justice Science, Data, and
Research Plan. Retrieved from: https://www.whitehouse.gov/wp-
content/uploads/2024/07/NSTC-EJ-Research-Plan-Julv-2024.pdf.
Navajo Epidemiology Center. 2023. Cancer Among the Navajo 2014 - 2018. Navajo Cancer
Workgroup, Navajo Department of Health, Navajo Nation. Retrieved from: https://nec.navajo-
nsn.gov/Portals/0/Reports/NavaioCancerReport%2013Nov2023.pdf.
Nolte, C.G., P. Dolwick, N. Fann, L. Horowitz, V. Naik, R. Pinder, T. Spero, D. Winner, and L. Ziska.
2018. Air quality, In Impacts, Risks, and Adaptation in the United States: Fourth National
Climate Assessment, Volume II. U.S. Global Change Research Program, Washington, D.C.
Nweke, O., D. Payne-Sturges, L. Garcia, C. Lee, H. Zenick, P. Grevatt, W. Sanders, H. Case, and I.
Dankwa-Mullan. 2011. Symposium on integrating the science of environmental justice into
decision-making at the Environmental Protection Agency: An overview. American Journal of
Public Health 101(S1): S19-26.
Office of Management and Budget (OMB). 1978. Statistical Policy Directive No. 14, Definition of
Poverty for Statistical Purposes. Federal Register 43:87 (4 May 1978) p. 19269.
OMB. 2019. Improving Implementation of the Information Quality Act. Retrieved from:
https://www.whitehouse.gov/wp-content/uploads/2019/04/M-19-15.pdf.
OMB. 2023a. Broadening Public Participation and Community Engagement in the Regulatory Process.
July 19, 2023. Memorandum. Retrieved from: https://www.whitehouse.gov/wp-
content/uploads/2023/07/Broadeninq-Public-Participation-and-Communitv-Enqaqement-in-the-
Regulatorv-Process.pdf.
OMB. 2023b. Circular A-4, Regulatory Analysis. Retrieved from: https://www.whitehouse.gov/wp-
content/uploads/2023/11/CircularA-4.pdf.
OMB. 2024. Revisions to OMB's Statistical Policy Directive No. 15: Standards for Maintaining,
Collecting, and Presenting Federal Data on Race and Ethnicity Federal Register 89 FR 22182
(28 March).
Office of Science and Technology Policy (OSTP) and CEQ. 2022. Guidance for Federal Departments
and Agencies on Indigenous Knowledge. November 30. Retrieved from:
https://www.whitehouse.gov/wp-content/uploads/2022/12/QSTP-CEQ-IK-Guidance.pdf.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
103
-------
Pace, C., C. Balazs, K. Bangia, N. Depsky, A. Renteria, R. Morello-Frosch, L.J. Cushing. 2022.
Inequities in drinking water quality among domestic well communities and community water
systems, California, 2011-2019. American Journal of Public Health 112(1): 88-97.
Padula, A.M., Z. Rivera-Nunez, and E.S. Barrett. 2020. Combined impacts of prenatal environmental
exposures and psychosocial stress on offspring health: Air pollution and metals. Current
Environmental Health Report 7: 89-100.
Pastor, Jr, M., Sadd, J. L., and Morello-Frosch, R. 2002. Who's minding the kids? Pollution, public
schools, and environmental justice in Los Angeles. Social Science Quarterly 83(1): 263-280.
Payne-Sturges, D. 2011. Humanizing science at the U.S. Environmental Protection Agency. American
Journal of Public Health 101 (S1): S8-S12.
Payne-Sturges D., M. Scammell, J. Levy, D. Cory-Slechta, E. Symanski, J. CarrShmool, R. Laumbach,
S. Under, J. Clougherty. 2018. Methods for evaluating the combined effects of chemical and
non-chemical exposures for cumulative environmental health risk assessment. International
Journal of Environmental Research and Public Health 15(12): 2797.
Pelacho, M., H. Rodriguez, F. Broncano, R. Kubus, F. Sanz Garcia, B. Gavete, and A. Lafuente. 2021.
Science as a commons: Improving the governance of knowledge through citizen science, In K.
Vohland et al. (Eds.) The Science of Citizen Science. Springer Press. Switzerland, pp. 57-78.
Redding, S., and E. Rossi-Hansberg. 2017. Quantitative spatial economics. Annual Review of
Economics 9: 21-58.
Ringquist, E. 2005. Assessing evidence of environmental inequities: A meta-analysis. Journal of Policy
Analysis and Management 24(2): 223-247.
Rowangould, G. 2013. A census of the near-roadway population: public health and environmental
justice considerations. Transportation Research Part D 25:59-67.
Rocha, J.I., S.E. Grineski, and T.W. Collins. 2017. A qualitative examination of factors shaping high and
low exposures to hazardous air pollutants among Hispanic households in Miami. Local
Environment 22(10): 1252-1267.
Sandman P. 1989. Hazard versus outrage in the public perception of risk. In Covello, V., McCallum, D.,
Pavlova, M. (Eds.). Effective Risk Communication: The Role and Responsibility of Government
and Non-Government Organizations, vol 4. Springer New York, N.Y. pp. 45-49.
Schwartz, J., D. Bellinger, and T. Glass. 2011a. Exploring potential sources of differential vulnerability
and susceptibility in risk from environmental hazards to expand the scope of risk assessment.
American Journal of Public Health 101: S94-S101.
Schwartz, J., D. Bellinger, and T. Glass, 2011b. Expanding the scope of risk assessment: Methods of
studying differential vulnerability and susceptibility. American Journal of Public Health 101:
S102-S109.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
104
-------
Schwartz, N., C. von Glascoe, V. Torres, L. Ramos, and C. Soria-Delgado. 2015. Where they (live,
work and) spray: Pesticide exposure, childhood asthma and environmental justice among
Mexican-American farmworkers. Health and Place 32: 83-92.
Science Advisory Board (SAB). 2015. SAB Review of the EPA's Draft Technical Guidance for
Assessing Environmental Justice in Regulatory Analysis, U.S. EPA. April 23, 2015.
Schultz A., N. Stanton, R. Pomazal, M. Lange, R. Irving, J. Meiman, B. Shelton, and K.C. Malecki.
2023. Biomonitoring of perflouroalkyl and polyflouroalkyl substances (PFAS) from the Survey of
the Health of Wisconsin (SHOW) 2014-2016 and comparison with the National Health and
Nutrition Examination Survey (NHANES). Journal of Exposure Science and Environmental
Epidemiology 33(5): 766-777.
Seawright, J. 2016. Multi-Method Social Science: Combining Qualitative and Quantitative Tools.
Cambridge: Cambridge University Press.
Sexton, K., and S. Under. 2010. The Role of cumulative risk assessment in decisions about
environmental justice. International Journal of Environmental Research and Public Health 7(11):
4037-4049.
Sexton, K. 2013. Evolution of public participation in the assessment and management of environmental
health risks: A brief history of developments in the United States. Journal of Public Health
Research 2(2): e18.
Shadbegian, R. J., and W. B. Gray. 2012. Spatial patterns in regulatory enforcement. In Banzhaf, S.
(Ed.). The Political Economy of Environmental Justice, pp. 225-248.
Shadbegian, R. and A. Wolverton. 2015. Evaluating environmental justice: Analytic lessons from the
academic literature and in practice. In Konisky, D. (Ed.). A False Promise of Justice? Evaluating
the Federal Government's Response to Environmental Inequity. MIT Press.
Shao, S., L. Liu, and Z. Tian. 2022. Does the environmental inequality matter? A literature
review. Environmental Geochemistry and Health 44: 3133-3156.
Sheriff, G., and K. Maguire. 2020. Health risk, inequality indexes, and environmental justice. Risk
Analysis 40(12): 2661-2674.
Shields, L.M., W.H. Wiese, B.J. Skipper, B. Charley, and L. Banally. 1992. Navajo birth outcomes in the
Shiprock uranium mining area. Health Physics 63(5): 542-551.
Shkembi, A., L. Smith, and R. Neitzel. 2022; Linking environmental injustices in Detroit, Ml to
institutional racial segregation through historical federal redlining. Journal of Exposure Science
& Environmental Epidemiology 34: 389-398.
Short, K. 2012. The research Supplemental Poverty Measure: 2011. Current Population Reports, pp.
60-244. Washington, DC: U.S. Census Bureau.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
105
-------
Solar O., and A. Irwin. 2010. A Conceptual Framework for Action on the Social Determinants of Health:
Social Determinants of Health Discussion Paper 2. Geneva: World Health Organization.
Retrieved from:
https://iris.who.int/bitstream/handle/10665/44489/9789241500852 enq.pdf?sequence=1.
Steel, M. 2020. Model averaging and its use in economics. Journal of Economic Literature 58(3): 644-
719.
Stephens, B. 2016. What have we learned about the microbiomes of indoor environments? Systems
1(4): e00083-16.
Stieb, D., A. Huang, R. Hocking, D. Crouse, A. Osornio-Vargas, and P. Villeneuve. 2019. Using maps
to communicate environmental exposures and health risks: Review and best-practice
recommendations. Environmental Research 176: 108518.
Switzer, D. 2019. Citizen partisanship, local government, and environmental policy
implementation. Urban Affairs Review 55(3): 675-702.
Swope, C.B., D. Hernandez, and L.J. Cushing. 2022. The relationship of historical redlining with
present-day neighborhood environmental and health outcomes: A scoping review and
conceptual model. Journal of Urban Health 99: 959-983.
Taherdoost, H. 2022. What are different research approaches? Comprehensive review of qualitative,
quantitative, and mixed method research, their applications, types, and limitations. Journal of
Management Science and Engineering Research 05(01): 53-63.
Tashakkori, A., R. Johnson, and C. Teddlie. 2020. Foundations of Mixed Methods Research:
Integrating Quantitative and Qualitative Approaches in the Social and Behavioral Sciences.
SAGE Publications.
Teodoro, M.P., M. Haider, and D. Switzer. 2018. US environmental policy implementation on Tribal
lands: trust, neglect, and justice. Policy Studies Journal 46(1): 37-59.
Terry, A., K. Herrick, J. Afful, and N. Ahluwalia. 2018. Seafood consumption in the United States, 2013-
2016. NCHS Data Brief No. 32.
Tesch, R. 2013. Qualitative Research. Routledge Falmer. New York, New York.
Tessum, C.W., D.A. Paolella, S.E. Chambliss, J.S. Apte, J.D. Hill, J.D. Marshall. 2021. PM2.5 polluters
disproportionately and systemically affect people of color in the United States. Science
Advances 7(18): eabf4491.
Triantafyllidou, S., J. Burkhardt, J. Tully, K. Cahalan, M. DeSantis, D. Lytle, and M. Schock. 2021.
Variability and sampling of lead (Pb) in drinking water: Assessing potential human exposure
depends on the sampling protocol. Environment International 146: 106259.
Tulve, N., J. Ruiz, K. Lichtveld, S. Darney, and J. Quackenboss. 2016. Development of a conceptual
framework depicting a child's total (built, natural, social) environment in order to optimize health
and well-being. Journal of Environmental and Health Sciences 2(2): 1-8.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
106
-------
U.S. Census Bureau. 2020a. Understanding and Using American Community Survey Data: What Users
of Data for Rural Areas Need to Know. Retrieved from: www.census.gov/programs-
s u rve vs/acs/l i b rarv/h an d boo ks/ru ra I. htm I.
U.S. Census Bureau. 2020b. Understanding and Using American Community Survey Data: What All
Data Users Need to Know. Retrieved from:
https://www.census.gov/content/dam/Census/librarv/publications/2020/acs/acs general handbo
ok 2020.pdf.
U.S. Census Bureau. 2022. How the Census Bureau Measures Poverty. Retrieved from:
https://www.census.gov/topics/income-povertv/povertv/guidance/povertv-measures.html
U.S. Census Bureau. 2024. Ancestry. Retrieved from:
https://www.census.gov/topics/population/ancestrv.html
U.S. EPA. 1986. Guidelines for the Health Risk Assessment of Chemical Mixtures. Retrieved from:
http://cfpub.epa. gov/ncea/risk/recordisplav.cfm?deid=22567.
U.S. EPA 1990. Technical Support Document on Risk Assessment of Chemical Mixtures (EPA/600/8-
90/064). Washington, DC: U.S. EPA, Office of Research and Development. Retrieved from:
http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid=35770.
U.S. EPA. 1997. Guidance on Cumulative Risk Assessment: Planning and Scoping. Retrieved from:
https://www.epa.gov/sites/default/files/2015-01/documents/cumrisk2 O.pdf.
U.S. EPA. 2000a. Supplementary Guidance for Conducting Health Risk Assessment of Chemical
Mixtures. Retrieved from: https://cfpub.epa.gov/ncea/raf/chem mix.htm.
U.S. EPA. 2000b. Risk Characterization Handbook (EPA 100-B-00-002). Washington, D.C.: U.S. EPA,
Office of Research and Development. Retrieved from:
http://www.epa.gov/sites/production/files/2015-
10/documents/osp risk characterization handbook 2000.pdf.
U.S. EPA. 2001. General Principles for Performing Aggregate Exposure and Risk Assessments.
Washington, D.C.: U.S. EPA, Office of Pesticide Programs. Retrieved from:
http://www.epa.gov/sites/production/files/2015-07/documents/aggregate.pdf.
U.S. EPA. 2002a. Guidance on Cumulative Risk Assessment of Pesticide Chemicals that have a
Common Mechanism of Toxicity. Washington, D.C.: U.S. EPA, Office of Pesticide Programs.
Retrieved from: http://www.epa.gov/sites/production/files/2015-
07/documents/guidance on common mechanism.pdf.
U.S. EPA. 2002b. Guidelines for Ensuring and Maximizing the Quality, Objectivity, Utility, and Integrity
of Information Disseminated by the Environmental Protection Agency (EPA/260R-02-008).
Washington, D.C.: U.S. EPA, Office of Environmental Information. Retrieved from:
http://www.epa.gov/sites/production/files/2015-08/documents/epa-info-gualitv-guidelines.pdf.
U.S. EPA. 2003a. Framework for Cumulative Risk Assessment (EPA-630-P-02-001F). Washington,
D.C.: U.S. EPA, Risk Assessment Forum. Retrieved from:
http://www.epa.gov/sites/production/files/2014-11/documents/frmwrk cum risk assmnt.pdf.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
107
-------
U.S. EPA. 2003b. A Summary of General Assessment Factors for Evaluating the Quality of Scientific
and Technical Information (EPA-100-B-03-001). Washington, D.C.: U.S. EPA, Science Policy
Council. Retrieved from: https://www.epa.gov/sites/default/files/2015-01/documents/assess2.pdf
U.S. EPA. 2005a. Guidelines for Carcinogen Risk Assessment (EPA/630/P-03/001F). Washington,
D.C.: U.S. EPA, Risk Assessment Forum. Retrieved from:
https://www.epa.gov/sites/default/files/2013-09/documents/cancer guidelines final 3-25-05.pdf
U.S. EPA. 2005b. Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to
Carcinogens (EPA/630/R-03/003F). Washington, D.C.: U.S. EPA, Risk Assessment Forum.
Retrieved from: https://www.epa.gov/sites/default/files/2013-
09/documents/childrens supplement final.pdf.
U.S. EPA. 2006. Framework for Assessing Health Risk of Environmental Exposures to Children
(EPA/600/R-05/093F). Washington, D.C.: U.S. EPA. Retrieved from:
https://assessments.epa.gov/risk/document/&deid=158363.
U.S. EPA. 2007a. Considerations for Developing a Dosimetry-Based Cumulative Risk Assessment
Approach for Mixtures of Environmental Contaminants (Final Report). Retrieved from:
http://cfpub.epa.gov/ncea/risk/recordisplav.cfm?deid=172725.
U.S. EPA. 2007b. Concepts, Methods, and Data Sources for Cumulative Health Risk Assessment of
Multiple Chemicals, Exposures and Effects: A Resource Document (Final Report) (EPA/600/R-
06/013F). Washington, D.C.: U.S. EPA, Office of Research and Development. Retrieved from:
https://assessments.epa.gov/risk/document/&deid=190187.
U.S. EPA. 2010a. Responses to Significant Comments on the 2009 Proposed Rule on the Primary
National Ambient Air Quality Standards for Nitrogen Dioxide (OAR-2006-0922). Retrieved from:
http://www3.epa.gov/ttn/naags/standards/nox/data/20100122rtc.pdf.
U.S. EPA. 2010b. Waste and Cleanup Risk Assessment Glossary. Office of Land and Emergency
Management. Retrieved from:
https://sor.epa.gov/sor internet/registry/termreg/searchandretrieve/glossariesandkeywordlists/s
earch.do?details=&vocabName=Waste%20and%20Cleanup%20Risk%20Assess.
U.S. EPA. 2011a. Exposure Factors Handbook: 2011 Edition (EPA-600-R-090-052F). Washington,
D.C.: U.S. EPA, Office of Research and Development. Retrieved from:
http://www.epa.gov/expobox/exposure-factors-handbook-2011-edition.
U.S. EPA. 2011 b. An Update on Ongoing and Future EPA Actions to Empower Communities and
Advance the Integration of Environmental Justice in Decision-Making and Research. Accessed
at https://www.epa.gov/sites/default/files/2015-02/documents/100-dav-challenge-2011-1Q-
26.pdf.
U.S. EPA. 2011 c. Regulatory Impact Analysis for the Final Mercury and Air Toxics Standards (EPA-
452-R-11-011). Research Triangle Park: U.S. EPA, Office of Air Quality Planning and
Standards. Retrieved from: https://www3.epa.gov/ttnecas 1 /regdata/RIAs/matsriafinal.pdf.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
108
-------
U.S. EPA. 2012a. Guidance for Evaluating and Documenting the Quality of Existing Scientific and
Technical Information. Washington, D.C.: U.S. EPA, Science and Technology Policy Council.
Retrieved from: http://www.epa.qov/sites/production/files/2015-05/documents/assess3.pdf.
U.S. EPA. 2012b. Microbial Risk Assessment Guideline: Pathogenic Microorganisms with Focus on
Food and Water (EPA/100/J-12/001). Retrieved from:
https://www.epa.gov/sites/default/files/2013-09/documents/mra-guideline-final.pdf
U.S. EPA. 2012c. Chlorpyrifos - Evaluation of the Potential Risks from Spray Drift and the Impact of
Potential Risk Reduction Measures. Memorandum. Washington, D.C.: U.S. EPA, Office of
Chemical Safety and Pollution Prevention. Retrieved from:
https://www.regulations.gov/document/EPA-HQ-OPP-2008-0850-01Q5
U.S. EPA. 2013a. America's Children and the Environment, Third Edition (EPA-240-R-13-001).
Washington, D.C.: U.S. EPA. Retrieved from: http://www.epa.gov/ace.
U.S. EPA. 2013b. Model Guidelines for Public Participation: An Update to the 1996 NEJAC Model Plan
for Public Participation. Retrieved from: https://www.epa.gov/environmentaliustice/model-
guidelines-public-participation.
U.S. EPA. 2013c. Residential Exposure Assessment Standard Operating Procedures. Addenda 1:
Consideration of Spray Drift. Office of Pesticide Programs. EPA-HQ-OPP-2013-0676.
U.S. EPA. 2014a. Policy on Environmental Justice for Working with Federally Recognized Tribes and
Indigenous Peoples. Washington, D.C.: U.S. EPA. Retrieved from:
https://www.epa.gov/sites/production/files/2015-02/documents/ei-indigenous-policy.pdf.
U.S. EPA. 2014b. Framework for Human Health Risk Assessment to Inform Decision-Making
(EPA/100/R-14/001). Washington, D.C.: U.S. EPA, Office of the Science Advisor, Risk
Assessment Forum. Retrieved from: https://www.epa.gov/sites/production/files/2014-
12/d ocu m e nts/h h ra-f ra m ewo rk-f i na I-2 014. pdf.
U.S. EPA 2015a. Guidance on Considering Environmental Justice During the Development of an
Action. Retrieved from: https://www.epa.gov/environmentaliustice/guidance-considering-
environmental-iustice-during-development-action.
U.S. EPA. 2015b. Better Decisions Through Consultation and Collaboration. U.S. EPA: Conflict
Prevention and Resolution Center. Retrieved from: https://www.epa.gov/sites/default/files/2015-
09/documents/better decisions.pdf.
U.S. EPA. 2015c. Risk and Technology Review - Petroleum Refineries Fact Sheet for Communities.
Retrieved from: https://www.epa.gov/sites/default/files/2016-06/documents/201Q-
0682 factsheet communities 09292015.pdf.
U.S. EPA. 2015d. Disposal of Coal Combustion Residuals from Electric Utilities Final Rule.
Washington, D.C.; U.S. EPA, Office of Resource Conservation and Recovery. Retrieved from:
https://www.federalregister.gov/documents/2015/04/17/2015-0Q257/hazardous-and-solid-waste-
management-system-disposal-of-coal-combustion-residuals-from-electric.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
109
-------
U.S. EPA 2015e. Peer Review Handbook. Science and Technology Policy Council. 4th Edition
(EPA/100/B-15/001). Retrieved from: http://www.epa.gov/osa/peer-review-handbook-4th-
edition-2015.
U.S. EPA. 2016a. Guidance for Conducting Fish Consumption Surveys (823B16002). U.S. EPA. Office
of Water. Retrieved from: https://www.epa.gov/sites/default/files/2016-12/documents/guidance-
fish-consumption-surveys.pdf.
U.S. EPA. 2016b. Petroleum Refinery Fenceiine Monitoring Stakeholder Engagement: Webinar for EJ
Groups, Fenceiine Communities and the Public. June 22. Retrieved from:
https://www.epa.gov/sites/default/files/2016-06/documents/communitv webinar iune 2016.pdf.
US EPA. 2016c. Integrated Review Plan for the National Ambient Air Quality Standards for Particulate
Matter. EPA-452/R-16-005. https://www3.epa.gov/ttn/naags/standards/pm/data/201612-final-
integrated-review-plan.pdf
U.S. EPA. 2016d. Environmental Justice Research Road Map (EPA 601/R-16/006). December.
Retrieved from: https://www.epa.gov/sites/default/files/2017-
01/documents/researchroadmap environmentaliustice 508 compliant.pdf.
U.S. EPA. 2019a. Guidelines for Human Exposure Assessment (EPA/100/B-19/001). Washington, D.C:
U.S. EPA. Risk Assessment Forum. Retrieved from:
https://www.epa.gov/sites/default/files/2020-
01/documents/guidelines for human exposure assessment final2019.pdf.
U.S. EPA. 2019b. Handbook for Citizen Science Quality Assurance and Documentation (EPA-206-B-
18-001). Retrieved from: https://www.epa.gov/sites/default/files/2019-
03/documents/508 csgapphandbook 3 5 19 mmedits.pdf
U.S. EPA. 2019c. Integrated Science Assessment (ISA) for Particulate Matter. Final Report. December.
Retrieved from: https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=347534#tab-3.
U.S. EPA. 2021a. Phase Down of Hydrofluorocarbons. Final Rule. Retrieved from:
https://www.epa.gov/climate-hfcs-reduction/final-rule-phasedown-hvdrofluorocarbons-
establishing-allowance-allocation.
U.S. EPA. 2021b. Indoor Air Quality. Retrieved from: https://www.epa.gov/report-environment/indoor-
air-gualitv.
U.S. EPA. 2021c. Climate Change and Social Vulnerability in the United States: A Focus on Six
Impacts. Retrieved from: https://www.epa.gov/system/files/documents/2021-09/climate-
vulnerability september-2021 508.pdf.
U.S. EPA. 2021 d. Using All Appropriate Injunctive Relief Tools in Civil Enforcement Settlements.
Memo. April 26. Retrieved from: https://www.epa.gov/sites/default/files/2021-
04/documents/usingallappropri ateiniunctiverelieftoolsincivilenforcementsettlement0426.pdf.
U.S. EPA. 2021e. Report on the Environment. Glossary. Retrieved from: https://www.epa.gov/report-
environment/roe-glossarv.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
110
-------
U.S. EPA. 2022a. EPA Legal Tools to Advance Environmental Justice. May. Retrieved from:
https://www.epa.gov/oqc/epa-leqal-tools-advance-environmental-iustice.
U.S. EPA. 2022b. Exposure Assessment Tools by Lifestages and Populations - Highly Exposed or
Other Susceptible Population Groups. Retrieved from: https://www.epa.gov/expobox/exposure-
assessment-tools-lifestaqes-and-populations-hiqhlv-exposed-or-other-susceptible.
U.S. EPA. 2022c. Cumulative Impacts Research: Recommendations forEPA's Office of Research and
Development (EPA 600/R-22/014a). September. Retrieved from:
https://www.epa.gov/svstem/files/documents/2022-
09/Cumulative%20lmpacts%20Research%20Final%20Report FINAL-EPA%20600-R-22-
014a.pdf.
U.S. EPA. 2022d. Supplement to 2019 Integrated Science Assessment (ISA) for Particulate Matter.
Retrieved from: https://cfpub.epa.gov/ncea/isa/recordisplay.cfm?deid=354490#tab-3.
U.S. EPA. 2022e. Using Participatory Science at EPA: Vision and Principles. June. Retrieved from:
https://www.epa.gov/svstem/files/documents/2022-
06/EPA%20Vision%20for%20Participatorv%20Science%206.23.22.pdf.
U.S. EPA. 2022f. Regulatory Impact Analysis for Control of Air Pollution from New Motor Vehicles:
Heavy-Duty Engine and Vehicle Standards. Retrieved from:
https://nepis. epa.gov/Exe/ZyPDF. cgi?Dockey=P1016A9N.pdf.
U.S. EPA. 2022g. Regulatory Impact Analysis for Phasing Down Production and Consumption of
Hydrofluorocarbons. Retrieved from: https://www.epa.gov/svstem/files/documents/2022-
07/RIA%20for%20Phasing%20Down%20Production%20and%20Consumption%20of%20Hydrof
luorocarbons%20%28HFCs%29.pdf.
U.S. EPA. 2022h. Analysis of Demographic Factors for Populations Living Near Hazardous Organic
NESHAP (HON) Facilities. Retrieved from: https://www.regulations.gov/document/EPA-HQ-
OAR-2022-0730-0082.
U.S. EPA. 2022i. Strategy to Reduce Lead Exposures and Disparities in U.S. Communities. Retrieved
from: https://www.epa.gov/lead/final-strategy-reduce-lead-exposures-and-disparities-us-
communities.
U.S. EPA. 2023a. EPA Legal Tools to Advance Environmental Justice: Cumulative Impacts Addendum.
Retrieved from: https://www.epa.gov/ogc/epa-legal-tools-advance-environmental-iustice.
U.S. EPA. 2023b. EPA Legal Tools to Advance Environmental Justice: Executive Order 14096
Addendum. Retrieved from: https://www.epa.gov/ogc/epa-legal-tools-advance-environmental-
iustice.
U.S. EPA. 2023c. Supplemental Effluent Limitations Guidelines and Standards for the Steam Electric
Power Generating Point Source Category. Final rule.
U.S. EPA. 2023d. Exposure Assessment Tools by Media - Aquatic Biota. Retrieved from:
https://www.epa.gov/expobox/exposure-assessment-tools-media-aguatic-biota.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
111
-------
U.S. EPA. 2023e. Principles and Best Practices for Oversight of State Implementation and Enforcement
of Federal Environmental Laws. Memo from the Administrator. February 17. Retrieved from:
https://www.epa.gov/system/files/documents/2023-03/principles-and-best-practices-oversight-
federal-environmental-proqrams-2023.pdf.
U.S. EPA. 2023f. EPA Policy on Consultation with Indian Tribes: Guidance for Discussing Tribal Treaty
or Similar Rights. Retrieved from: https://www.epa.gov/svstem/files/documents/2023-12/epa-
guidance-for-discussing-tribal-treaty-or-similar-rights-2023.pdf.
U.S. EPA. 2023g. Finding That Lead Emissions from Aircraft Engines That Operate on Leaded Fuel
Cause or Contribute to Air Pollution That May Reasonably Be Anticipated to Endanger Public
Health and Welfare. Retrieved from: https://www.govinfo.gov/content/pkg/FR-2023-10-
20/pdf/2023-23247. pdf
U.S. EPA. 2023h. Standards of Performance for New, Reconstructed, and Modified Sources and
Emissions Guidelines for Existing Sources: Oil and Natural Gas Sector Climate Review. Final
Rule. Retrieved from: https://www.epa.gov/controlling-air-pollution-oil-and-natural-gas-
operations/epas-final-rule-oil-and-natural-gas
U.S. EPA. 2024a. National Emission Standards for Hazardous Air Pollutants: Ethylene Oxide Emission
Standards for Sterilization Facilities Residual Risk and Technology Review. Final rule.
U.S. EPA. 2024b. Regulatory Impact Analysis for Per-and Polyfluoroalkyl Substances (PFAS) National
Primary Drinking Water Standard. Final rule.
U.S. EPA. 2024c. EJScreen Technical Documentation. Version 2.3. Washington, D.C.: U.S. EPA.
Retrieved from: https://www.epa.gov/system/files/documents/2024-07/eiscreen-tech-doc-
version-2-3.pdf.
U.S. EPA. 2024d. Achieving Health and Environmental Protection Through EPA's Meaningful
Engagement Policy. Retrieved from: https://www.epa.gov/environmentaliustice/epas-
meaningful-engagement-policy .
U.S. EPA. 2024e. Supplemental Effluent Limitations Guidelines and Standards for the Steam Electric
Power Generating Point Source Category. Final rule.
U.S. EPA. 2024f. Guidelines for Preparing Economic Analyses, third edition (EPA-240-R-24-001).
Washington, DC: U.S. EPA, National Center for Environmental Economics. Retrieved from:
https://www.epa.gov/environmental-economics/guidelines-preparing-economic-analyses.
U.S. EPA. 2024g. Economic Analysis of the Final Reconsideration of the Dust-Lead Hazard Standards
and Post-Abatement Clearance Levels. Final rule.
U.S. EPA. 2024h. Interim Framework to Advance Consideration of Cumulative Impacts. November.
U.S. EPA. 2024i. Compendium of Examples from Recent EJ Analyses Conducted for EPA
Rulemakings. Retrieved from: https://www.epa.gov/environmentaliustice/technical-guidance-
assessing-environmental-iustice-regulatory-analysis.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
112
-------
U.S. EPA. 2024j. Economic Analysis of the TSCA Section 6 Final Rule for Asbestos Risk Management,
Part 1. Retrieved from: https://www.requlations.qov/docket/EPA-HQ-OPPT-2021-0Q57.
U.S. EPA. 2024k. Regulatory Impact Analysis for Hazardous and Solid Waste Management System:
Disposal of Coal Combustion Residuals from Electric Utilities; Legacy Surface Impoundments.
Final rule. Retrieved from: https://www.regulations.gov/document/EPA-HQ-QLEM-2020-0107-
1067.
U.S. EPA. 2024I. Integrated Risk Information System (IRIS) Online Glossary. Retrieved from:
https://www.epa.gov/iris/iris-glossary.
U.S. EPA. 2024m. Reconsideration of the National Ambient Air Quality Standards for Particulate Matter
(PM). Final rule.
U.S. EPA. 2024n. Multi-Pollutant Emissions Standards for Model Years 2021 and Later Light-Duty and
Medium-Duty Vehicle. Final rule.
U.S. Federal Communications Commission (FCC). 2020. Broadband Deployment Report. June 8.
Retrieved from: https://www.fcc.gov/reports-research/reports/broadband-progress-reports/2020-
broadband-deplovment-report.
U.S. Global Change Research Program (USGCRP). 2016. The Impacts of Climate Change on Human
Health in the United States: A Scientific Assessment. Washington, D.C.
USGCRP. 2018. Impacts, Risks, and Adaptation in the United States: Fourth National Climate
Assessment, Volume II. Washington, D.C.
USGCRP. 2023. Fifth National Climate Assessment. Crimmins, A.R., C.W. Avery, D.R. Easterling, K.E.
Kunkel, B.C. Stewart, and T.K. Maycock, Eds. Washington, D.C.
Varshavsky, J.R., S.D. Rayasam, J.B. Sass, D.A. Axelrad, C.F. Cranor, D. Hattis, R. Hauser, P.D.
Koman, E.C. Marquez, R. Morello-Frosch, C. Oksas, S. Patton, J.F. Robinson, S.
Sathyanarayana, P.M. Shepard, and T.J. Woodruff. 2023. Current practice and
recommendations for advancing how human variability and susceptibility are considered in
chemical risk assessment. Environmental Health 21 (Suppl 1): 133.
Walker, W. 2013. The transitional costs of sectoral reallocation: Evidence from the Clean Air Act and
the workforce. The Quarterly Journal of Economics 128(4): 1787-1835.
Warren, R. 2022. Undercount of undocumented residents in the 2020 American Community Survey and
estimates and trends in the undocumented population from 2010 to 2020, by US state and
country of origin. Journal on Migration and Human Security 10(4): 228-237.
Webler, T., and S. Tuler. 2021. Four decades of public participation in risk decision making. Risk
Analysis 41(3): 503-518.
White, K., F. Khan, C. Peck, and M. Corbin. 2013. Guidance on modeling offsite deposition of
pesticides via spray drift for ecological and drinking water assessments. Draft. EPA-HQ-OPP-
2013-0676.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
113
-------
Wilson, S.M., F. Howell, S. Wing, and M. Sobsey. 2002. Environmental injustice and the Mississippi
hog industry. Environmental Health Perspectives 110(S2): 195-201.
Wlson, S.M., K. Burwell-Naney, C. Jiang, H. Zhang, A. Samantapudi, R. Murray, L. Dalemarre, L. Rice,
and E. Wlliams 2015. Assessment of sociodemographic and geographic disparities in cancer
risk from air toxics in South Carolina. Environmental Research 140: 562-568.
Wolverton, A. 2009. Effects of socio-economic and input-related factors on polluting plants' location
decisions. Berkeley Electronic Journal of Economic Analysis and Policy Advances 9(1): Article
14.
Wolverton, A. 2023. Environmental justice analysis for EPA rulemakings: Opportunities and challenges.
Review of Environmental Economics and Policy 17(2): 346-353.
Woodburn, A. 2017. Investigating neighborhood change in airport-adjacent communities in multiairport
regions, 1970-2010. Transportation Research Record 2626(1): 1-8.
World Health Organization (WHO). 2006. Principles for evaluating health risks in children.
Environmental Health Criteria 237.
WHO. 2023. Air Quality and Health. Retrieved from: https://www.who.int/teams/environment-climate-
chanqe-and-health/air-qualitv-and-health/health-impacts .
WHO. 2024. Operational Framework for Monitoring Social Determinants of Health Equity. Retrieved
from: https://iris.who.int/bitstream/handle/10665/375732/9789240088320-enq.pdf?sequence=1
Yoo, E-H. Cooke, and Y. Eum. 2023. Examining the geographical distribution of air pollution disparities
across different racial and ethnic groups: Incorporating workplace addresses. Health & Place
84: 103112.
Zwickl, K., M. Ash, and J. Boyce. 2014. Regional variation in environmental inequality: Industrial air
toxics exposure in US cities. Ecological Economics 107: 494-509.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
114
-------
Appendix A: Select EPA Guidance and
Other Documents
This appendix contains a list of EPA technical guidance and other documents that may be helpful to
analysts when evaluating EJ concerns for regulatory actions. Table A1 lists relevant technical guidance,
guidelines, and framework documents. Table A2 lists other EPA documents that may prove useful.
Table A.1 Relevant EPA Guidance Documents
1 Topic Area
Title
Publication Year
Web Link 1
Economics
Guidelines for Preparing Economic
Analyses
2024f
https://www.epa.aov/sites/default/files/2017-
08/documents/ee-0568-50.pdf
Human Health Risk
Framework
Framework for Human Health Risk
Assessment to Inform Decision-
Making
2014b
https://www.epa.aov/sites/production/files/20
14-12/documents/hhra-framework-final-
2014.pdf
Human Health Risk
Framework
Framework for Assessing Health
Risk of Environmental Exposures
to Children
2006
https://assessments.epa.aov/risk/document/
&deid=158363
Other Health Risk
Guidance
Microbial Risk Assessment
Guideline: Pathogenic
Microorganisms with Focus on
Food and Water
2012b
http://www.epa.aov/sites/production/files/201
3-09/documents/mra-auideline-final.pdf
Other Health Risk
Guidance
Guidelines for Carcinogen Risk
Assessment
2005a
https://www.epa.aov/sites/default/files/2013-
09/documents/cancer auidelines final 3-
25-05.pdf
Other Health Risk
Guidance
Supplemental Guidance for
Assessing Susceptibility from
Early-Life Exposure to Carcinogens
2005b
https://www.epa.aov/sites/default/files/2013-
09/documents/childrens supplement final.p
df
Other Health Risk
Guidance
Supplementary Guidance for
Conducting Health Risk
Assessment of Chemical Mixtures
2000a
http://ofmpub.epa.aov/eims/eimscomm.aetfil
e?p download id=4486
Other Health Risk
Guidance
Technical Support Document on
Risk Assessment of Chemical
Mixtures
1990
http://cfpub.epa.aov/ncea/cfm/recordisplav.c
fm?deid=35770
Other Health Risk
Guidance
Guidelines for the Health Risk
Assessment of Chemical Mixtures
1986
http://cfpub.epa.aov/ncea/risk/recordisplav.cf
m?deid=22567
Exposure
Assessment
Guidelines for Human Exposure
Assessment
2019a
https://www.epa.aov/risk/auidelines-human-
exposure-assessment
Exposure
Assessment
Guidance for Conducting Fish
Consumption Surveys
2016a
https://www.epa.aov/sites/default/files/2016-
12/documents/auidance-fish-consumption-
survevs.pdf
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition A-1
-------
Topic Area
Title
Publication Year
Web Link 1
Exposure
Assessment
Exposure Factors Handbook
2011a122
https://www.epa.qov/expobox/about-
exposure-factors-handbook
Risk
Characterization
Risk Characterization Handbook
2000b
http://www.epa.qov/risk/risk-
characterization-handbook
Cumulative Risk
Assessment
Considerations for Developing a
Dosimetry-Based Cumulative Risk
Assessment Approach for Mixtures
of Environmental Contaminants
(Final Report)
2007a
http://cfpub.epa.aov/ncea/risk/recordisplav.cf
m?deid=172725
Cumulative Risk
Assessment
Concepts, Methods, and Data
Sources for Cumulative Health Risk
Assessment of Multiple Chemicals,
Exposures, and Effects: A
Resource Document (Final Report)
2007b
https://assessments.epa.aov/risk/document/
&deid=190187
Cumulative Risk
Assessment
Framework for Cumulative Risk
Assessment
2003a
http://www.epa.aov/sites/production/files/201
L
11/documents/frmwrk cum risk assmnt.pdf
Cumulative Risk
Assessment
Guidance on Cumulative Risk
Assessment of Pesticide
Chemicals that have a Common
2002a
http://www.epa.aov/sites/production/files/201
5;
07/documents/auidance on common mech
Mechanism of Toxicity
anism.pdf
Cumulative Risk
Assessment
General Principles for Performing
Aggregate Exposure and Risk
Assessments
2001
http://www.epa.aov/pesticide-science-and-
assessina-pesticide-risks/aeneral-principles-
performina-aaareaate-exposure-and
Cumulative Risk Guidance on Cumulative Risk https://www.epa.gov/sites/default/files/2015-
Assessment Assessment: Planning and Scoping 01/documents/cumrisk2 O.pdf
122 While the latest edition of the Exposure Factors Handbook was published in 2011, the EPA has updated several individual
chapters more recently: Soil and Dust Ingestion in 2017; Intake of Fruits and Vegetables, Intake of Meat, Dairy Products, and
Fats, Intake of Grain Products, and Building Characteristics in 2018; and Ingestion of Water and Other Select Liquids was
updated in 2019.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
A-2
-------
Table A.2 Other Relevant EPA Documents
Topic Area
Title
Publication Year
Web Link
Meaningful
Engagement
Achieving Health and Environmental
Protection Through EPA's
Meaningful Engagement Policy
2024d
https://www.epa.aov/environmentaliustice/ep
as-meaninaful-enaaaement-policv
Meaningful
Engagement
Better Decisions Through
Consultation and Collaboration
2015b
https://www.epa.aov/sites/default/files/2015-
09/documents/better decisions.pdf
Meaningful
Engagement
Policy on Environmental Justice for
Working with Federally Recognized
Tribes and Indigenous People
2014a
https://www.epa.aov/sites/production/files/201
5-02/documents/ei-indiaenous-policv.pdf
Meaningful
Engagement
Model Guidelines for Public
Participation: An Update to the1996
NEJAC Model Plan for Public
Participation
2013b
https://www.epa.aov/environmentaliustice/mo
del-auidelines-public-participation
Meaningful
Engagement
An Update on Ongoing and Future
EPA Actions to Empower
Communities and Advance the
Integration of Environmental Justice
in Decision-Making and Research
2011b
https://www.epa.aov/sites/default/files/2015-
02/documents/100-dav-challenae-2011-10-
26.pdf.
Data Quality
Handbook for Citizen Science
Quality Assurance and
Documentation
2019b
https://www.epa.aov/sites/default/files/2019-
03/documents/508 csaapphandbook 3 5 19
mmedits.pdf
Data Quality
Peer Review Handbook
2015e
http://www.epa.aov/osa/peer-review-
handbook-4th-edition-2015
Data Quality
Guidance for Evaluating and
Documenting the Quality of Existing
Scientific and Technical Information.
2012a
http://www.epa.aov/sites/production/files/201
5-05/documents/assess3.pdf
Data Quality
A Summary of General Assessment
Factors for Evaluating the Quality of
Scientific and Technical Information
2003b
https://www.epa.aov/sites/default/files/2015-
01 /documents/assess2. pdf
Data Quality
Guidelines for Ensuring and
Maximizing the Quality, Objectivity,
Utility, and Integrity of Information
Disseminated by the Environmental
Protection Agency
2002b
http://www.epa.aov/sites/production/files/201
5-08/documents/epa-info-aualitv-
auidelines.pdf
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
A-3
-------
Appendix B: Incorporating EJ Concerns
into Exposure and Effects Assessments
The planning, scoping, and problem formulation phases provide a key opportunity to ensure that EJ
concerns are incorporated into HHRA. This appendix provides several key EJ-specific questions to
consider when planning for an exposure or dose-response assessment. It describes the implications of
each question for the data gathering and analytic work that may be necessary to address them. Also
included are examples of analyses from the peer-reviewed literature and/or U.S. government analyses,
which may suggest approaches for analysts to consider during planning, scoping, and problem
formulation.
Planning for an Exposure Assessment
Patterns of exposure to stressors across population groups of concern may vary for several reasons.
Variation may be predominantly a spatial phenomenon, if exposure is highest within close proximity to
pollution sources and that is where the population group of concern is most likely to reside. Exposure
differences may reflect variation in behaviors (e.g., subsistence anglers) or exposures due to specific
dietary or cultural practices of a population group (e.g., exposures to pesticides in reeds used for
basket weaving). Exposure may reflect unique aspects of the use or application of the chemical (e.g.,
exposures to pesticide applicators) or it may be affected by other factors that increase vulnerability for a
specific population group (e.g., greater prevalence of a pre-existing health condition such as asthma).
Questions and Key Considerations
1. Based on the use and release patterns of the environmental stressors of concern, are there
population groups that might be more highly exposed?
Environmental stressors may be used and released in a variety of circumstances. However, even when
the stressor is intended for use in a particular circumstance or location, unintended releases can result.
For instance, the stressor could migrate to an unintended location. One example of this is spray drift
from pesticide applications that result in pesticides falling on "off-target" locations, which may then lead
to increased exposure for certain populations that live in close proximity to the treated fields (e.g.,
farmers, migrant workers, children). Text Box B.1 discusses how the potential risk for exposure due to
pesticide application and residues can be calculated using drift modeling and other methods while
accounting for evaporation of aerosols (i.e., volatilization), and the potential effects to bystanders.
Some factors for consideration when evaluating the use and release patterns of environmental
stressors include evaluating the potential for risks due to intended use and potential migration of the
stressor, prevalence of use, environmental fate, and the toxicological characteristics of the stressor.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
B-1
-------
Text Box B.1 Pesticide Spray Drift Risk Assessment to Bystanders
Farmworkers and their families often live near the fields where they work and can be exposed to
pesticides in a manner different from other population groups because of this proximity. While direct
measures of the degree of drift in the vicinity of fields may be difficult or impossible to obtain,
exposure estimates from these residues may be calculated using drift modeling and methods
employed for typical residential risk assessments.
Spray drift can be characterized as the movement of aerosols and volatile components away from a
treated area as a result of the application process. Bystanders, defined as those who live on, work
in, or frequent areas adjacent to treated fields, can be exposed to spray drift directly or by contact
with resulting deposited residues (e.g., children playing on lawns next to treated fields). The degree
of such effects is governed by many processes (e.g., application method, nozzles used, release
height) and the conditions at the time of application (e.g., wind speed and direction).
To model potential high-end exposure to people living near treated agricultural fields (e.g., via
deposition on residential turf), the EPA used AgDRIFT (V2.1.1) and AgDISP (V8.26) to provide
deposition values for residential lawns, as a fraction of the application rate, at different distances
downwind of a treated field. Analysis of spray drift evaluates risks from pesticides similar to how they
are evaluated for use on turf because this scenario represents the highest potential for exposure
associated with spray drift and considers different lifestages, including children at different
developmental stages. Data from pesticide studies that determined turf residue levels and
dissipation rates after application are often available, and in the absence of these data, default
assumptions can be used. This information is used in conjunction with the standard residential
methods to estimate exposure from treated turf, including exposures from all pertinent routes for
both adults and children.
Conceptual Model for Spray Drift Modeling (U.S. EPA, 2012c)
See draft EPA guidance documents on the consideration of spray drift in pesticide risk assessment (White et
al., 2013; U.S. EPA, 2013c).
Treated Field
Z' Buffer
50' Wide Lawn
X = Distance Downwind From Treated Field
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
B-2
-------
2. Are exposure variabilities predominantly a spatial phenomenon (e.g., due to contaminant hot
spots)? Is proximity to a source a reasonable proxy for estimating exposure to stressors of
concern?
For environmental stressors that are dispersed locally in ambient media, exposure may be effectively
captured using proximity to the source as a surrogate measure. Further detail about these methods can
be found in Chakraborty et al. (2011) and in Chapter 6 of this technical guidance.
3. Can exposure variability be estimated using ambient contaminant concentrations, either
measured or modeled? Are data available or can data be modeled at a reasonable spatial scale
appropriate for available demographic and socioeconomic data?
Ambient concentrations can be used to identify and assess spatial variability in exposure that may
contribute to exposure differences between population groups. Two types of ambient concentration
information exist: data from ambient air quality monitors, and modeled estimates of ambient
concentrations averaged over a period of time. Monitoring data generally offer a more accurate
estimate of the level of exposure to a stressor. However, obtaining monitoring data at a level of
geospatial resolution that allows for the evaluation of differences may not be feasible for a number of
reasons, including: (1) some environmental stressors may not be routinely monitored; and (2) coverage
for routinely monitored stressors is insufficient to provide the level of geospatial resolution required to
discern differences, as most monitoring data are available only down to the county level. This lack of
detail is problematic given that racial, ethnic, and income diversity, as well as differences in ambient
concentrations, could vary widely with the level of geospatial resolution. An example of an alternative
strategy for evaluating multi-pollutant settings is provided in Text Box B.2.
Modeled data can sometimes serve as a surrogate for monitoring data when high quality data inputs
are used. Ambient air quality modeling methods have been developed to estimate ambient
concentrations of a plume beyond its point of release, based on relevant factors such as meteorology
and chemical characteristics (e.g., reactivity and solubility). However, the predictive accuracy of models
is not comparable across stressors. Important considerations for using modeled data should include the
predictive accuracy of the model for the stressor in question and the ability to predict ambient
concentrations for smaller geospatial units such as census tracts. Data provided at a larger geospatial
scale than the census tract may not support assessment of differences in exposure. Analysts may
consider the use of screening models to highlight concerns about exposure differences, which can be
evaluated in greater detail with more sophisticated models at a later stage.
4. Are bio-monitoring data available for population groups of concern, including those with potentially
elevated exposures?
Although analysis using bio-monitoring data can be time consuming, it may be the most accurate way
to estimate exposures for population groups of concern. A literature search for previous assessments of
differential exposure using survey data should be conducted prior to beginning of such an analysis. An
important resource to consider is the National Report on Human Exposure to Environmental Chemicals
(CDC, 2022). Human exposure data in this report are presented by lifestage, race, ethnicity, and
income to the extent that such detailed breakouts are possible.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
B-3
-------
Text Box B.2 Understanding Environmental Inequality Using Different Policy Instruments
Using a case study approach, the EPA evaluated the viability of a multi-pollutant, risk-based pollution
control strategy as an alternative to a traditional pollutant-by-pollutant approach to air quality (Fann et al.,
2011). The study used spatially resolved air quality, population, and baseline health data from the Detroit
metropolitan area to perform within- and across-group comparisons of exposure and risk. The objective
of the study was to demonstrate how states might design air quality attainment strategies that: (1) attain
tighter standards, (2) maximize human health benefits of air quality improvements, and (3) achieve a
more equitable distribution of air pollution-related risk.
The assessment of EJ concerns followed four steps: (1) identify and model exposure to population
groups vulnerable to PIVh.s-related mortality and morbidity effects in the baseline, based on fine scale air
quality modeling and population characteristics such as education attainment, race, and poverty level; (2)
design an emission control strategy that maximizes air quality improvements among these population
groups, primarily by reducing emissions of directly-emitted PM2.5; (3) compare the multi-pollutant, risk-
based strategy with the traditional pollution control strategy for attainment by modeling the air quality
effects of each strategy and comparing the results with the baseline scenario; and (4) calculate the
change in exposure/risk inequality from the baseline to assess whether a multi-pollutant risk-based
strategy results in a more equal distribution of exposure and risk than a traditional pollution control
strategy. The findings from this study revealed that the population risk reduction approach produced
greater net benefits.
Risk-Based, Multi-Pollutant Modeling Framework (Fann et al., 2011)
Select control
strategy A
Multi pollutant
modeling platform
Integrated
emissions
inventory
1
Multipollutant
control measures
1
Air quality
modeling system
I
Estimated change
in air pollutant
concentrations
Health impact
assessment
Fig. 1. Risk-based, multipollutant
modeling framework.
Revise to incorporate
population-oriented emission
controls
Using similar indicators of inequality as Fann et al. (2011), Mansur and Sheriff (2021) compare the
distributions of nitrogen oxide (NOx) emissions from California's Regional Clean Air Incentives Market
cap-and-trade program to estimated counterfactual emissions under a prescriptive regulatory approach.
Using U.S. Census data, they estimate individual exposure from facility emissions using NOx emissions
as a proxy for health effects by race, ethnicity, and poverty status. They find that trading led to a more
equitable distribution of emissions reductions across all groups.
Additional information about the Detroit multi-pollutant project can also be found at
https://www.epa.qov/sites/default/files/2021-01/documents/detroitpres final09.pdf.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition B-4
-------
When using exposure biomarkers to draw inferences about exposure differences for a source-specific
regulatory action, analysts should carefully consider the extent to which measured levels reflect
exposure, and whether biomarkers represent total exposure to an environmental stressor from multiple
sources. Comparisons at this stage are often focused on point estimates or, at most, deterministic
models rather than complex probabilistic models. Analysts may use simple, well-established
comparative methods such as ratios to examine between-population group comparisons or may apply
more complex approaches such as analysis of variance or regression techniques as needed.
Comparisons may focus on specific segments of the distribution or on the proportion of a population
group represented within a percentile category. At times, several years of data may need to be
combined to obtain sufficient sample size to conduct analysis in the tail of the distribution, subject to
resource, analytic, and data constraints.
As discussed in Section 5.3.3, use of biomonitoring data has both benefits and limitations. While a large
population survey (e.g., 2014-2016 Survey of the Health of Wsconsin or NHANES) may suggest the
existence of exposure difference, locale- or site-specific surveys (e.g., New York City Health and
Nutrition Examination Survey) can yield more detailed insights into the dimensions of the differences.
For example, analysis of Survey of the Health of Wisconsin and NHANES data demonstrate that
Wsconsin residents have lower PFAS levels in serum than the U.S. as a whole. More data are needed
to evaluate the effects of PFAS on non-White or low-income populations (Schultz, 2023).
5. Are there population groups that may experience greater exposure to stressors because of their
unique food consumption patterns, behaviors, or use of specific products?
Analysts can consider whether the population group of concern has higher levels of exposure to a
stressor due to food consumption patterns that differ from those of the general population (e.g., unique
diets or greater reliance on hunting and fishing for food), behaviors (e.g., hand-to-mouth behavior of
young children), or through greater use of specific products (e.g., personal care and cleaning products).
Understanding potential exposures from these types of sources will allow for more accurate estimates
of exposures to the stressor(s) of concern. Differences in exposures from ingestion may be due to
several factors, including regional variability in dietary habits, and cultural, ethnic, or religious
practices.123 A population group of concern may consume certain foods at higher rates than members
of other groups or consume parts of animals or plants not commonly consumed by the general
population. For example, children in Tribal communities may consume as much as fifteen times more
fish than children in the general population (U.S. EPA, 2011a). Additionally, some population groups
may eat food predominantly from specific locations. Likewise, subsistence fishers may consume fish far
more frequently and obtain it only from local waterways. If fish from these waterways have higher levels
of a contaminant, subsistence fishers and their families may have higher exposure levels (U.S. EPA,
2011a). Similarly, some cosmetics may contain lead. Analysts can evaluate the exposure pathway
(e.g., dermal or inhalation), frequency of use, and identify the populations most likely to use these
products in unique ways (Burger and Gochfield, 2011). The EPA's Guidelines for Human Exposure
123 "Variability in exposure occurs because of location, occupation, activities within a location, socioeconomic status, consumer
preferences, dietary habits and other lifestyle choices. Behaviors relative to lifestage can be particularly influential
determinants for exposure, especially for infants and toddlers and for the embryo/fetus during pregnancy. Lifestage, health
status, sex and genetic differences also can be important factors that determine dose. The drivers for human activities are
complex and, unlike stressors, cannot be predicted using first-principle models based on physical/chemical properties. Instead,
human activities are treated as stochastic properties (random variables) described by population distributions based on
available (e.g., observational or modeled) data" (U.S. EPA, 2019a). See Chapter 8 in U.S. EPA (2019a) for detailed discussion
of uncertainty and vulnerability in exposure assessments.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
B-5
-------
Assessment (U.S. EPA, 2019a) include information on considerations of lifestages, vulnerable groups,
and population groups of concern in exposure assessment, including Tribal and Indigenous populations
and other racial or ethnic groups.
Text Box B.3 illustrates how the five scoping questions for integrating EJ into an exposure assessment
could be posed to evaluate dietary risks from pesticide residues.
Text Box B.3 Example of Scoping Questions for Integrating EJ Considerations into
Assessments of Dietary Risk from Pesticide Residues
To ensure that EJ considerations are explicitly considered in dietary risk assessments for pesticides,
risk assessors could consider the following scoping questions when evaluating whether risk
concerns may exist.
• Based on the pesticide use patterns, are there population groups that might be more highly
exposed to pesticide residues because of their unique consumption patterns (e.g., subsistence
diets or other cultural practices)?
• Is it likely that the pesticide or its metabolites/degradates will bioaccumulate such that
increased exposure and risk might be expected for certain population groups (e.g., lifestages;
regular consumers offish, shellfish, or game)?
• Is the pesticide used on, or likely to be found in, foods that are consumed in substantially
higher amounts by certain population groups (e.g., lemon grass)?
• Does the pesticide have an atypical or unusual use pattern that could result in unusual
exposures for certain population groups (e.g., use in non-traditional agriculture, or locally-
restricted use)?
• Do the physical and/or chemical properties of the pesticide indicate a potential for long range
transport (e.g., volatility, persistence), especially pesticides that may also bioaccumulate?
• Are there other groups within the population groups of concern (e.g., based on lifestage) who
might be more highly exposed to the pesticide through their diet?
Planning For an Effects Assessment
An effects assessment includes hazard identification and dose-response assessment. Planning,
scoping, and problem formulation for the effects assessment of HHRA present other opportunities to
incorporate EJ concerns into a risk assessment. Planning, scoping, and problem formulation play key
roles in identifying population groups of concern that may exhibit a particular sensitivity to a stressor.
This is also the point at which analysts can consider how demographic and socioeconomic
characteristics might modify effects seen in the general population. Analysts can consider whether
factors particular to a population may alter dose-response relationships for the contaminants in
question.
Below are a few key questions and sample responses that highlight the types and scale of analytic
work that may be required to adequately integrate EJ concerns into an effects assessment.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
B-6
-------
Questions and Key Considerations
1. What population groups are most relevant from a risk perspective for the stressor(s) in question?
The purpose of asking this question is twofold: (1) defining the vulnerable population groups, and (2)
considering what dose-response or concentration-response information is available for those population
groups. The goal should be to achieve as close a match as possible between the information available
in the literature and the characteristics of the population (i.e., care should be taken not to fit a dose-
response function to a population group to which it does not apply). To answer this question, analysts
may need to consider stratification by race, ethnicity, and income, or factors such as educational level,
access to health care, and baseline disease prevalence.
2. Are there population-specific effect assessments for the population groups of concern?
In answering this question, analysts can investigate these factors: (1) Are there known or identified
effect modifiers?; (2) For identified factors that modify hazards of interest, how are they distributed
among population groups of concern?; and (3) Are effect modifiers distributed differently among various
lifestages within population groups? To answer these questions, a review of relevant literature is
necessary to identify potential sources of population group-specific dose-response information or data
on effect modifiers (see Text Box B.4).
3. Are the spatial and temporal scales of the studies supplying the dose-response function
consistent with the spatial scale needed to incorporate EJ concerns, from both an exposure and
outcome perspective?
Ideally, the dose-response functions chosen should match as closely as possible the geographic scale
of the proposed analysis incorporating EJ concerns. Analysts may introduce measurement error if
dose-response functions from studies conducted over smaller geographic areas are applied at a more
aggregate scale. For example, if the study assigned each subject in the cohort a county-level average
response, the study could underestimate the true relationship between exposure and outcome at a finer
spatial scale. Likewise, if the exposure in the study is acute, it cannot be applied directly to incorporate
EJ concerns where the exposure of interest is chronic; rather, the exposure duration being modeled in
the regulatory analysis should be considered.
Analysts may consider adjusting the geographic scale to incorporate EJ concerns for this reason, and
also may need to change the scope if detailed data on factors such as baseline health are available
only at a certain scale (e.g., at the local urban level or at the acute exposure level).
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
B-7
-------
Text Box B.4 Concentration-Response Functions Stratified by Demographic Factors
The literature on particulate matter (PM) provides examples of concentration-response functions
stratified by demographic factors including age and race. Di et al. (2017) analyzed the relationship
between air pollution exposure and mortality in the Medicare population, specifically working to
understand concentration-response relationships by race and Medicaid eligibility. For PM2.5, the risk
of death among men, Black individuals, and people with Medicaid eligibility was notably higher than
for the rest of the population.
1.25-1
¦£ 1.10-
n
£
&
— Overall
population
Female Male Native White Hispanic Asian Black Not Eligible Eligible for
American for Medicaid Medicaid
•2 0.96-
+ 4.- -F-
— Overall
population
Male Native White Hispanic Asian
American
Black Not Eligible Eligible for
for Medicaid Medicaid
Figure. Risk of death associated with an increase of 10 |jg/m3 in PM2.5 concentrations (top) and an
increase of 10 ppb in ozone exposure (bottom)
The EJ analysis for the final PM NAAQS reconsideration (U.S. EPA, 2024m) includes a distributional
analysis of the estimated relative risk of PM2.s-related mortality using the dose-response functions
stratified by race and ethnicity from Di et al. (2017). This analysis improved upon assessments
performed in previous NAAQS regulatory impact assessments by combining these differentiated
concentration-response functions with county-level baseline mortality rates. Using this approach, the
EPA found that the same PM2.5 exposure reduction will reduce the risk of mortality approximately
three times more in Black populations than in White populations.
Technical Guidance for Assessing Environmental Justice in Regulatory Analysis - Second Edition
B-8
------- |