EPA 910-R-16-002
April 2016
United States
Environmental Protection
Agency
Food: Too Good To Waste
An Evaluation Report for the Consumption
Workgroup of the West Coast Climate and
Materials Management Forum
t I
TOOG
TO WASTE
I U.S. Environmental Protection Agency
I Region 10
-------
ACKNOWLEDGEMENTS
This report was prepared on behalf of the U.S. EPA Region 10 and the West Coast Climate and
Materials Management Forum by the Forum's Consumption Workgroup with support from
EcoPraxis and Tetra-Tech under subcontract to Toeroek Associates.
Special thanks to all of the EPA staff and Forum members from state and local governments
who reviewed drafts, offered revisions, and contributed to the final product. Specifically, we
acknowledge the contributions of:
Dr. Viki Sonntag of EcoPraxis, the primary researcher for this report, and
EPA Region 10 staff: Ashley Zanolli, the Campaign Planning Team Lead; Theresa
Elaine, the Regional Contract Manager and Peer Learning Group Lead; and EPA
NNEMS Fellows Megan Curtis-Murphy and Sanne Stienstra. Questions about the report
may be directed to Ashley Zanolli, Environmental Engineer, U.S. EPA Region 10. To
learn more about the West Coast Climate and Materials Management Forum, visit
www.westcoastclimateforum.com.
A hearty thanks also goes out to the community partners and EPA regional representatives,
listed below, upon whose work this evaluation is based. The report would be far from complete
without their vital participation.
Community Partners:
Addison County Solid Waste Management District, Addison County, Vermont:
Teresa Kuczynski and Donald Maglienti
City of Aurora, Aurora, Colorado: Nathan Owens
City of Gresham Recycling and Solid Waste, Gresham, Oregon: Karen Guillen-
Chapman and Nick Isbister
City and County of Honolulu, Hawaii: Alexandra Lavers
City of Iowa City Land and Recycling Center, Iowa City, Iowa: Jennifer Jordan
City of Palo Alto Public Works Department, Palo Alto, California: Wendy Hediger
Kanu Hawaii and Kupu, Hawaii: Cyrus Howe, Nicole Chatterson and Kaleo Ten
King County Solid Waste Division, King County, Washington: Karen May and Josh
Marx
Sustainable Jersey City, Jersey City, New Jersey: Debra Italiano
Naropa University, Boulder, Colorado: Santiago Giraldo-Anduaga
Village of Oak Park Environmental Services and Seven Generations Ahead, Oak
Park, Illinois: Karen Rozmus and Jen Nelson
Rhode Island Food Policy Council, Rhode Island: Krystal Noiseux, Dave Rocheleau,
and Leo Pollock
San Benito County, California: Lisa Jensema
Seattle Public Utilities, Seattle, Washington: Karl Woestwin and Veronica Fincher
Thurston County Solid Waste, Thurston County, Washington: Terri Thomas and
Gabrielle Byrne
EPA Regional Representatives:
Christine Beling, Region 1; Lorraine Graves, Region 2; Julie Schilf, Region 5; Gayle
Hubert, Region 7; Virginia Till, Region 8; and Shannon Davis and Jenny Stephenson,
Region 9
April 2016
-------
TABLE OF CONTENTS
EXECUTIVE SUMMARY 3
1.0 INTRODUCTION 6
1.1 FTGTW OBJECTIVES AND DEVELOPMENT 7
1.1.1 Campaign Design Principles 8
1.2 RESEARCH OBJECTIVES 8
2.0 CAMPAIGN DESCRIPTION 10
2.1 KEY WASTE PREVENTION BEHAVIORS 10
2.2 CBSM TOOLS 11
2.3 IMPLEMENTATION 13
3.0 ANALYSIS OF CAMPAIGN RESULTS 14
3.1 CAMPAIGN DESCRIPTIONS 14
3.2 CAMPAIGN EFFECTIVENESS 22
3.2.1 Effectiveness of Behavior Change Strategies and Tools 22
3.2.2 Effectiveness of Messaging 28
3.2.3 Effectiveness of Outreach and Engagement 31
3.2.4 Household Factors Affecting Reach and Effectiveness 34
3.2.5 Effectiveness of Implementation Support and Learning Community 35
3.3 CAMPAIGN IMPACT 36
3.3.1 Baseline Amount 39
3.3.2 Percentage Reduction in Wasted Food 42
3.3.3 Household Factors Affecting Impact 43
3.4 CAMPAIGN COSTS, ENVIRONMENTAL BENEFITS AND FIT WITH
EXISTING PROGRAMS 46
3.5 FINDINGS SUMMARY 48
4.0 RECOMMENDATIONS 51
4.1 EMERGING BEST PRACTICES 51
4.2 LESSONS LEARNED 52
4.3 FURTHER APPLIED RESEARCH 53
REFERENCES 55
APPENDIX A: PRE-AND POST-CHALLENGE SURVEY INSTRUMENT 57
APPENDIX B: CAMPAIGN DESCRIPTIONS 59
APPENDIX C: FTGTW COMMUNITY-SCALE MEASUREMENT METHODOLOGY DRAFT
PROTOCOL. ...64
April 2016
-------
FIGURES
Figure 1: Challenge's Effect on Increasing Awareness of Wasted Food in Household 22
Figure 2: Likelihood of Continuing to Use Tools and Strategies 22
Figure 3: Usefulness of Strategies and Tools 24
Figure 4: Baseline Awareness 27
Figure 5: Ranked Motivations to Reduce Wasted Food 28
Figure 6: Food Loss and Waste Data Map 36
Figure 7: Per Capita Wasted Food Reduction Averages by Weight 39
Figure 8: Per Capita Wasted Food Reduction Averages by Volume 39
Figure 9: Waste Fraction Factors 42
Figure 10: RIFPC Preventable Food Waste Household Averages by Test Group 44
Figure 11: Staff Hours, King County 2012, 2013 and 2014 45
Table 1: Selected Behaviors and Associated Benefits and Barriers 9
Table 2: Campaign Descriptions 14
Table 3: Outreach and Engagement Descriptions and Results 18
Table 4: Volume versus Weight Measurement Tradeoffs 25
Table 5: Household and Per Capita Average Percentage Reductions by Project 39
Table 6: Total Baseline Food Waste per Person per Week (Iowa City, Seattle, Honolulu) 38
Table 7: Baseline Data Comparison with Previous Waste Studies 40
April 2016
-------
Community-Based Food Waste Prevention -
Food: Too Good to Waste Evaluation Report
EXECUTIVE SUM MARY
This report presents the findings from Food: Too Good to Waste (FTGTW), a partnership of
Community-based Social Marketing (CBSM) campaigns aimed at reducing wasted food from
households.
Preventing wasted food represents a significant opportunity to keep the valuable resources
used to produce and distribute food from going to waste. Over 40 percent of the food produced
or imported for domestic consumption in the United States is lost, with over one-fourth of
household food purchases by weight going to waste.1
The report includes an analysis of seventeen FTGTW campaigns conducted in ten states from
October 2012 through December 2014. The campaigns focused on assisting households to
make small shifts in how they shop, prepare and store food to prevent it from being wasted.
The evaluation addresses both the effectiveness and the impact of the FTGTW campaigns. It
confirms that CBSM campaigns can bring about a notable reduction in preventable food waste
at the household level.
CAMPAIGN EFFECTIVENESS FINDINGS
A principle objective for this report is to determine the extent to which FTGTW campaigns result
in the desired behavior changes. This includes assessing how effective campaigns are in
generating and sustaining the desired behaviors leading to reductions in wasted food.
FTGTW uses CBSM messaging and tools to engage households in wasted food reduction
strategies. The messaging and tools are designed to address barriers and emphasize benefits
to changing behaviors, in this case, behaviors associated with wasting food intended for human
consumption, that is, the edible portions of food.
The tools include both tools that support specific behaviors, such as a Fruit and Vegetable
Storage Guide, and those that support a broader shift in awareness of wasted food as both an
environmental and economic issue, for example, a community workshop presentation. We refer
to these as behavior change and outreach tools respectively.
The findings on the effectiveness of the three CBSM components - behavior change strategies
and tools, messaging and outreach tools - are summarized next.
1 Economic Research Service, United States Department of Agriculture. 2012. Loss Adjusted Food Availability Data Series.
Available at http://www.ers.usda.gov/data-products/food-availability-(per-capita)-data-system/.aspx.
April 2016
-------
Behavior Change Strategies and Tools: Households found the FTGTW strategies and tools
both useful and easy to use. Of particular significance is the finding that households who
measured their food waste are highly motivated to reduce wasted food. In effect, household
food waste audits increase awareness by countering habitual behaviors and activating waste
aversion - a dislike of wasting resources in one's possession.
Messaging: Feeling bad about throwing away food (waste aversion) and wasting money
appear to be equally strong motivators to reducing wasted food. The evaluation also found that
increasing awareness of the indirect environmental effects of wasted food through messaging is
challenging with mixed results. While the campaigns generated an expressed and often
enthusiastic interest in reducing wasted food, there is the need to particularize environmental
messaging to the household level for greatest effect. An example of such is: throwing away an
apple is equivalent to flushing the toilet seven times.
Outreach and Engagement: The general rule for successful campaigns was to engage
participants early and often. Additionally, campaigns designed to leverage social networks and
create social norms were among the most effective in terms of outreach and engagement, while
community-scale direct outreach was more effective than recruitment through indirect means
such as social media outreach. It was also found that without a focused effort campaigns can
fail - competing priorities in two campaigns presented significant hurdles to success.
CAMPAIGN IMPACT FINDINGS
A second major evaluation objective is to determine if a shift in household food waste behaviors
has the potential to result in waste tonnage reduction. To achieve quantifiable reductions in
wasted food at the community level, it is necessary to engage a significant percentage of the
general population in adopting the behaviors as well as engaging and sustaining behaviors that
have a significant impact at the household level.
Given the small sample size of the majority of the campaigns, the focus was on measuring the
amount of food going to waste in individual households both before and after adopting
strategies to reduce wasted food, that is, impact at the household level. However, requesting
households weigh their waste appears to be an effective means of determining the potential for
reduction in small to medium-sized sample populations.
The total baseline amount of wasted food per person per week ranged from 2.2 pounds to 3.5
pounds. This is comparable to the EPA estimate of 2.5 pounds of landfilled residential waste per
person per week.
Campaigns that are successfully implemented can result in a significant reduction in
preventable (edible) food waste at the household level. The magnitude of the potential reduction
in preventable waste is 50% or more or approximately a half pound per person per week. This is
roughly equivalent to a 20% reduction in total food waste.
CONCLUSIONS AND NEXT STEPS
The FTGTW campaign results establish that behaviors to reduce wasted food are complex with
many complicating factors influencing these behaviors and food management practices in
general. At the same time, this evaluation shows that even small budget CBSM campaigns can
generate notable reductions in preventable food waste at the household level.
Based on the evaluation findings, it is strongly recommended that future campaigns consider
incorporating a household measurement tool or strategy into their campaigns.
April 2016 ~~4
-------
The next steps in advancing household prevention of wasted food as a priority are to measure
the impact of FTGTW campaigns at the community-level and to identify supportive policies for
scaling-up FTGTW campaigns.
April 2016
-------
1.0 INTRODUCTION
This report presents the findings from Food: Too Good to Waste, a Community-based Social
Marketing campaign aimed at reducing wasted food.
Food: Too Good to Waste (FTGTW) is a project of the West Coast Climate and Materials
Management Forum (the Forum), an EPA-led partnership of western cities and states that are
developing and sharing ways to integrate sustainable materials management policies and
practices into climate actions. In 2014, the project was expanded to include participation of
communities in EPA Regions 1, 2, 5, 7 and 8, in addition to Regions 9 and 10. EcoPraxis and
Toeroek Associates were contracted to perform the analysis contained in this report with
guidance from the EPA and their partners.
For both environmental and economic reasons, wasted food is emerging as an issue of
significant consequence. Over 40 percent of the food produced or imported for domestic
consumption in the United States is lost to the landfill and over a quarter of household food
purchases by weight go to waste.2 Food waste has been identified as a major source of
greenhouse gas emissions and other negative environmental impacts. By one estimate, food
waste accounts for more than one quarter of total freshwater use in the U.S.3
The purpose of FTGTW is to develop and test a Community-based Social Marketing (CBSM)
approach to changing food consumption behaviors with the intent of reducing wasted food from
households and its associated environmental impacts.4
Understanding the patterns of household food consumption and waste behaviors can increase
our chances of developing successful strategies to reduce wasted food and its environmental
impacts. To this end, the Forum researched food waste behaviors and potential behavior
change strategies in developing the campaign's messaging and tools.5 A branded toolkit
containing a variety of tools to support campaign implementation was made available to Forum
participants in the fall of 2012.6
This report presents the results of seventeen FTGTW campaigns conducted from October 2012
through December 2014. Data for the analysis came from household food waste
measurements, household participant questionnaires, and interviews with the organizations
implementing the campaigns. While the data from these early implementations are limited, the
analysis provides useful insights for conducting future FTGTW campaigns.
Sections 1 and 2 of the report set the evaluative context for the results reported in Section 3.
Section 1 gives an overview of the campaign and the research objectives, while Section 2
Economic Research Service, United States Department of Agriculture. 2012. Loss Adjusted Food Availability Data Series.
Available at http://www.ers.usda.gov/data-products/food-availability-(per-capita)-data-system/.aspx
Hall, Kevin D. et al. 2009. The Progressive Increase of Food Waste in America and Its Environmental Impact. PLoS ONE 4, no.
11.
The USDA defines food waste as food and beverages that were once available for human consumption but are discarded without
being eaten. Food waste is a sub-component of food loss. (See Muth, Mary et al. 2007. Exploratory Research on Estimation of
Consumer-Level Food Loss Conversion Factors. USDA ERS Report.) Alternatively, food loss and waste (FLW) is defined in the
Food Loss and Waste Protocol developed by the World Resource Institute (2015) as food and associated inedible parts removed
from the food supply chain, where food is any substance intended for human consumption.
Environmental Protection Agency (EPA). 2012. Food: Too Good to Waste Pilot: A Background Research Report for the West
Coast Climate and Materials Management Forum. EPA 910-R-12-006.
West Coast Climate and Materials Management Forum. July 2013. Toolkit Implementation Guide for the Food: Too Good to
Waste Pilot.
April 2016
-------
provides a description of the CBSM messaging and tools used in the pilot. Section 3 relates the
campaign findings. The report concludes with recommendations on how best to conduct and
scale-up future campaigns.
1.1 FTGTW OBJECTIVES AND DEVELOPMENT
FTGTW aims to engage households in efforts to reduce wasted food and its impacts through a
CBSM campaign. A second purpose is to analyze results that will help in the design of future
CBSM programs to reduce wasted food.
CBSM is an approach to driving behavioral change
through community initiatives that remove barriers to
desired behaviors, while simultaneously enhancing
those behaviors' advantages.7 It relies on a series of
key steps as an approach to designing programs (see 2- 'dentify Carriers and benefits of
sidebar) desired behaviors
3. Design pilot program with
Steps in a CBSM Approach
1. Identify desired behaviors
behavior change strategies
and messaging
4. Implement pilot program
5. Evaluate pilot program
6. Replicate and scale-up
successful pilot strategies
The first step in developing a CBSM campaign is to
select which behaviors to promote, beginning with a
determination of how the issue under study is affected
by a particular sector. In this case, the issue was food
waste and the sector was households. Information was
gathered to identify and compare behaviors of interest
in terms of their impact, penetration and probability.
The assessment provided guidance in identifying which behaviors are potential candidates for
large-scale change.
The second step in campaign development is to identify barriers and benefits associated with
the behaviors selected for change. This step involved both a literature review and focus groups.
A report providing a review of the research detailing food waste estimates was issued in
September 2012.8
In step three, the information gathered through the background research and focus groups was
then used to design behavioral change strategies and associated messages and tools. In the
fall of 2012, five pilots were conducted and subsequently evaluated in step four. This analysis
helped to further refine the FTGTW strategies, messaging and tools.
The final step in instituting a CBSM program is to roll out the pilot's successful strategies across
the sector of interest.
In all, seventeen campaigns were conducted from the fall of 2012 through the end of 2014.
Interested community partners, primarily local government agencies with a responsibility for
solid waste management, took the lead in implementing the campaigns in their communities.
This report documents the evaluation findings and emergent best practices.
7 McKenzie-Mohr, Doug. 2011. Social Marketing to Protect the Environment: What Works. Sage Publications Inc.
8 Environmental Protection Agency, ibid.
April 2016
-------
1.1.1 Campaign Design Principles
In addition to embodying CBSM principles in the FTGTW strategies and tools, the campaign
aims to address the needs of implementing communities. To facilitate implementation and
ensure robustness, the campaign was designed with the following principles in mind:
Remove/minimize barriers to preferred behaviors and emphasize benefits: This is a key
principle of the CBSM approach to behavior change.
Contextualize preferred behaviors: To motivate and sustain behavioral changes, the
campaign aims to draw the linkages between household practices and environmental
and social issues at a broader scale.
Engage at the community level and speak to community values: The ability to adapt the
campaign messages and tools to the needs of the community begins with engaging the
community in its implementation.
Leverage existing resources: Communities are called on to be resourceful in
implementing new programs. This encourages the engagement of community-based
partners in program implementation as well as a networked approach to program
development.
Design for breadth and depth: To achieve quantifiable reductions in wasted food at the
community level, it will be necessary to engage and sustain behaviors that have a
significant impact at the household level (depth) as well as engage a significant
percentage of the general population in adopting the behaviors (breadth).
Collect evidence for policy-making and program design: A solid evidence base that
supports the effectiveness of the strategies and tools is needed to justify scaling up the
campaign to long term, broad scale programs.
1.2 RESEARCH OBJECTIVES
In the design phase, the Forum established several research objectives. In addition to validating
a CBSM approach to wasted food prevention, a major goal for the evaluation is to gather
implementation data to support a full scale prevention campaign.
The objectives include determining:
Campaign Reach and Effectiveness: Did the campaign result in the desired behavior
changes? This includes assessments of participation rates and strategy and tool
effectiveness.
Campaign Impact: Did the campaign result in quantifiable reductions in wasted food?
Campaign Implementation Costs: What is the cost to implement pilots and, by extension,
full scale campaigns?
Environmental Impact: What are the estimated environmental benefits for a campaign to
reduce wasted food?
Program Fit: What is the fit of a wasted food prevention campaign with existing strategic
plans and programs such as climate protection and healthy food programs.
In addition to collecting measurement data from participating households, several of the
community partners chose to do surveys of the Challenge participants before and after the
Challenge. The data from these questionnaires provide information on demographic patterns of
household waste and also on the effectiveness of the Challenge itself. The survey instrument
can be found in Appendix A.
April 2016 8
-------
Data collected during the implementations for the purpose of this evaluation included:
Household measurements of wasted food amounts.
Survey data from household participants collected by community partners.
Quantitative data on recruitment and retention.
Observations of community partners regarding what worked.
The evaluation also includes an assessment of the value of these data for determining the
effectiveness and impact of the campaigns. One question that has been raised repeatedly in
past studies of household food waste is the bias in the results that arises from households
documenting their own waste.9
1.3 ENVIRONMENTAL GOALS
As noted above, wasted food is emerging as an issue of significant consequence for both
economic and environmental reasons. Consumption-based greenhouse gas (GHG) inventories
point to food production as a large generator of GHG.10 And, according to the UN's Food and
Agriculture Organization (FAO), wasted food contributes a total of 3.3 billion tons of GHGs a
year to the planet's atmosphere, making wasted food the third largest emitter if it were a
country.11 The same FAO report estimates that the volume of water used in producing wasted
food is equivalent to the flow of Russia's Volga River.
Ultimately, FTGTW aims to reduce the environmental impacts of wasted food. At the same time,
making an evidence-based connection between household food waste prevention and
environmental benefits is a complex undertaking. Proving that a CBSM approach to reducing
wasted food works is an important first step in establishing this connection.
Beyond this, it will be important to understand the institutional incentives and barriers to
developing programs to prevent wasted food. The need for such programs is apparent,
especially as the benefits of such programs could make a significant difference with the
outcomes of some of our greatest environmental challenges.
9 See, for example, FUSIONS, 2014, Report on Review of (Food) Waste Reporting Methodology and Practice.
10 See, for example, Stockholm Environment Institute, US Center, 2011, Consumption-Based Greenhouse Gas Inventory for Oregon
- 2005.
11 UN Food and Agriculture Organization. 2013. Food Wastage Footprint: Impacts on Natural Resources Summary Report.
http://www.fao.org/docrep/018/i3347e/i3347e.pdf
April 2016
-------
2.0 CAMPAIGN DESCRIPTION
The FTGTW Campaign aims to engage households in efforts to reduce wasted food and its
impacts through a CBSM campaign. This section of the report describes the key waste
prevention behaviors selected as the focus for the campaign, as well as the CBSM messaging,
tools and the toolkit implementation guide.
2.1 KEY WASTE PREVENTION BEHAVIORS
Food waste in households is the result of "a complex interrelationship between multiple activities
and the context in which they are performed."12 In addition, these activities are influenced by other
competing concerns, such as work, family, and other social relationships, often leading to a
disassociation between values and behaviors. As such, there are numerous opportunities to
reduce wasted food though not all are equally consequential.
In FTGTW's development stage, five behaviors were selected on the basis of their potential
impact for reducing wasted food from households. The background research informing behavior
selection was published in a report, Food: Too Good to Waste Pilot, a Background Research
Report for the West Coast Climate and Materials Management Forum The five behaviors were
later modified to incorporate findings from the first five pilots. The five final selected behaviors and
the associated benefits and barriers are listed in Table 1.
Table 1: Selected Strategies and Associated Benefits and Barriers
Behavior
^^^^^^^^^^^^^^^^^^^^^^H
Get Smart: See How Much
Food (and Money) You Are
Throwing Away
Smart Shopping: Buy
What You Need
Smart Storage: Keep
Fruits and Vegetables
Fresh
Smart Saving: Eat What
You Buy
Smart Prep: Prep Now, Eat
Later
Benefit
^^^^^^^^^^^^^^^^^^^^
Waste aversion*
Waste aversion
Saving money
Waste aversion
Health
Saving money
Waste aversion
Convenience
Saving money
Health
Barrier
^^^^^^^^^^^^^^^^^^^^^^^^^H
Dynamic lifestyle**
Time
Habitual behavior
Dynamic lifestyle
Time
Habitual behavior
Knowledge
Time
Not enough room in fridge
Gratification
Convenience
Skills
Knowledge
* Waste aversion is a dislike of wasting resources in one's possession. It is considered a benefit as
preventing waste leads to fewer feelings of guilt. ** Dynamic lifestyle refers to a high degree of
unpredictability in everyday activities, such as which meals are consumed at home.
12 Quested, Tom, et al. 2013. Spaghetti Soup: The Complex World of Food Waste Behaviors, Resources, Conservation and
Recycling 70 43-51.
13 Environmental Protection Agency, ibid.
April 2016
10
-------
These behaviors cover a range of household food management strategies - from food
purchasing, storage and preparation to choices about what to eat when. The associated
behavior change tools for these are described in the next section of this report.
It should be noted that, in addition to the barriers cited in Table 1, by the time a food item is
thrown out, the opportunity to prevent this has already passed, in some cases, many days
beforehand. This lag contributes to a disconnection between the activity that led to wasted food
and the actual occurrence of the waste.
2.2 CBSM TOOLS
FTGTW uses CBSM messaging and tools to engage households in strategies to reduce wasted
food. CBSM messaging and tools are designed to address barriers and emphasize benefits to
changing behaviors, in this case, behaviors associated with wasted food.
The knowledge and experience of Forum participants informed the design of the tools and
messaging, as did the literature on CBSM tool types. In particular, the McKenzie-Mohr study of
CBSM campaigns focused on changing household environmental behaviors was useful in
providing tool design principles and examples.14 CBSM experts also contributed to the design of
the campaign and the recent literature on behavioral economics was also consulted, particularly
with regard to the importance of loss aversion in averting wasteful behaviors.15 In addition, the
Forum interviewed and consulted with several food waste reduction experts and programs, both
here in the United States and abroad. Lastly, the strategies, tools, and messaging were tested
in five pilot campaigns in late 2012 - early 2013 and subsequently modified to improve their
effectiveness.
Two principal target populations were selected in crafting the campaign messaging and
strategies: (1) families with young children and (2) young adults (of ages approximately from 18
to 30). The two principal target populations were chosen on the basis of previous research that
indicated these two demographics generate the largest amounts of wasted food in households.
The specific campaign tools are described next. The complete set of tools can be viewed at
www.westcoastclimateforum.com/food. The tools include both tools that support specific
behaviors, such as the Fruit and Vegetable Storage Guide and the Shopping List Template, and
those that support a broader shift in awareness of wasted food as both an environmental and
economic issue, for example, the infographic/poster and the community workshop presentation.
We refer to these as behavior change and outreach tools respectively.
Behavior Change Tools
Fruit and Vegetable Storage Guide (Smart Storage Tool): The fruit and vegetable storage
guide is designed as a prompt tool for household use. A prompt is a visual aid to remind
households of a desired behavior. The guide provides useful information for keeping produce
fresh and is available in both English and Spanish. Prompts are particularly useful when
designed to engage people in positive behaviors and are presented in close proximity to where
the action takes place. The guide was printed in bright colors on a half sheet suitable for posting
on the refrigerator.
14 McKenzie-Mohr, Doug. 2011. Social Marketing to Protect the Environment: What Works. Sage Publications Inc.
15 See Thaler, Richard H. and Cass R. Sunstein. 2008. Nudge: Improving Decisions about Health, Wealth, and Happiness. Penguin
Books.
April 2016 ~TT
-------
Meals-in-Mind Shopping List Template (Smart Shopping Tool): The shopping list template
provides an easy-to-use tool for making a shopping list with meals in mind. It was designed to
create awareness around how much food will be needed for upcoming meals and, as such, is
intended to be a step towards meal planning. It was hypothesized that the tool would be
effective with young adults for whom meal planning is considered a loss of time since the
decision on whether to eat at home or eat out is driven by their "dynamic lifestyle". The tool also
focuses on the cost saving strategy of using up food that has already been purchased. The
template is available in both English and Spanish.
EAT *
Mm ml TO WASTE
FIRST!
Eat First Prompt (Smart Saving Tool): The Eat First prompt
can be placed on a designated shelf in the refrigerator to
corral items with limited shelf life. The prompt is pictured to
the right.
Food: Too Good to Waste Challenge (Get Smart Tool): In
practice, the Food: Too Good to Waste Challenge
incorporates implementation of the four other strategies but
its focus is on measuring the amount of food going to waste in households.16 Challenge
participants are provided instructions and instruments to measure the amount of food going to
waste in their households.
Challenges support behavioral change as a form of commitment in a group context while also
drawing attention to the need for new behaviors. In addition, several implementing organizations
offered incentives to participating households for completing the challenge.
The FTGTW Challenge serves a dual purpose as both a measurement and a behavior change
tool. The Challenge presents an opportunity to collect data from household participants by
which to evaluate the pilot's effectiveness and impact in addition to raising awareness of food
waste behaviors in participating households through feedback (measurement of amount of food
going to waste).
The period of time households measured waste before adopting waste reduction strategies is
referred to as the "baseline period" and the amount of food wasted during that time as the
"baseline amount".
Three variants of the Challenge were tested as described next:
Volume Measurement: The measurement tool challenges households to become aware of
how much food goes to waste in their homes by measuring the amounts of food thrown out
before and after adopting strategies to reduce waste. A measurement bag or bucket is used to
collect and measure preventable and inedible food waste.17 (In some implementations, only
preventable waste was collected). The tool consists of printed bags or buckets in which to
collect waste, instructions and worksheets. The instructions provide guidance on how to
participate in a month-long challenge. The worksheets are to be used to collect data on
preventable and non-edible food waste from household participants during the challenge.
16 Not all campaigns implemented Challenges. Of those that did, some solely engaged Challenge participants while others
conducted broader scale campaigns where only some participants took the Challenge. However, the Challenge was the sole
means of collecting data on the amount of food going to waste.
17 Preventable food waste is food that is intended for human consumption but is not eaten for any reason (e.g. mold, plate waste).
April 2016
-------
Weight Measurement: In the weight measurement variant of the Challenge, weight
measurements are substituted for the volume measurements.
Photo Diary: The photo diary can be used to document changes in food waste behaviors in
households participating in the challenge. This tool provides guidance on how to participate in a
two-week photo diary version of the challenge and worksheets to capture collected data.
Outreach and Engagement Tools
Message Map: The message map includes messaging for all 5 key waste prevention behaviors
selected during the campaign design (see Table 1). It can be used to tailor outreach materials to
individual campaign objectives.
Infographic/Poster: The purpose of the infographic is to tell a story about why wasted food
matters. The infographic provides a means of contextualizing wasted food as an issue in
relation to its environmental and economic impacts. This infographic can be used online, or as a
poster or handout at community events and venues, such as sustainability fairs, farmers
markets, and local grocery stores.
Workshop Presentation for Community Participants: The workshop presentation tool is a
slide show with accompanying narrative to be used at community workshops. This presentation
is intended to provide a space to engage households in thinking through strategies and provide
potential actions to reduce wasted food. The workshop presentation helps to establish food
waste aversion as a social norm. Workshop participants are also asked to make a commitment
to reduce wasted food. Research shows that public commitments are strong motivators in
making behavioral shifts.18
Incentives: Incentives were used to engage households in taking the Challenge. Some
incentives, such as scales and compost buckets, also reduced barriers to participation. Other
incentives included grocery certificates and culinary tools for participating and/or completing the
challenge.
2.3 IMPLEMENTATION
The tools in the FTGTW Toolkit are designed to be adaptable to the needs of the implementing
communities based on their objectives and resources.
The Implementation Guide provides local governments or other implementing organizations a
description of the Campaign and tool kit, and information on how the organizations might launch
a new wasted food prevention challenge or incorporate this campaign into existing programs.19
The guide is intended to support the implementing organizations, also referred to as community
partners in this report, in making a number of implementation choices by explaining the trade-
offs associated with the various choices.
Peer group learning calls were conducted monthly for community partners. These calls were an
important opportunity for communities to learn from each other by sharing successes, newly
developed resources, and lessons learned in real time.
18 McKenzie-Mohr, ibid.
19 West Coast Climate and Materials Management Forum, ibid.
April 2016 13
-------
3.0 ANALYSIS OF CAMPAIGN RESULTS
As of the end of 2014, seventeen FTGTW campaigns have been conducted. This section of the
report presents the findings from these initial implementations. Section 3.1 describes the
campaigns and the following sections relate the findings in terms of the research objectives.
Section 3.5 provides a summary of the key findings and their implications for running a
successful campaign.
3.1 CAMPAIGN DESCRIPTIONS
The campaign descriptions establish the reference conditions for the subsequent analyses.
Table 2 provides a summary description of each of the seventeen campaigns. More extensive
descriptions of select individual campaigns can be found in Appendix B.
Below is a synopsis of the descriptions in the two tables. The campaign descriptions are arrayed
in the table roughly in order of their implementation dates.
FTGTW Partner: While the majority of FTGTW partners were local government agencies with a
responsibility for solid waste management, non-profits were also represented. Typically, the
non-profits had broader organizational objectives, such as food policy (Rhode Island Food
Policy Council, Campaign 9) or environmental stewardship (Kanu Hawaii, Campaign 16).
Community Location and Urban/Rural Classification: The seventeen campaigns
represented a diversity of locations and urban/rural classifications. The community locations
included states from across the U.S. and spanned a range of classifications from small rural
towns to large cities.
Time of Year: Campaigns were held in various months and all seasons. A few campaigns were
held over the holidays leading up to mid-December.
Campaign Objectives: All of the campaigns had food waste reduction as their primary
objective. In the first five pilots, the focus was assessing tool effectiveness and gauging the
potential impact of waste reduction strategies on the amount of household food waste (King
County, San Benito, and Honolulu). In addition, the Seattle Public Utilities campaign provided
useful baseline information on the amount of household food wasted. In the late 2013 and the
2014 campaigns, partners sought to familiarize themselves with implementing a campaign and
the wasted food issue. Two organizations, King County (Campaigns 8 and 15) and Kanu Hawaii
(Campaign 16), scaled-up earlier pilots.
Campaign Focus: The campaign focus describes in general terms the implementing
organization's approach to achieving their objectives. Each campaign had a unique focus
according to their individual community's needs and resources. At the same time, for most of
the partners, limited resources meant a campaign targeting a small sub-set of the population.
These organizations tended to focus on Challenges. Broader scale media campaigns were
conducted in King and Thurston Counties in Washington, in Palo Alto, CA and in Hawaii. Finally,
King and Thurston Counties and Gresham, Oregon also did direct community-scale outreach,
such as tabling at farmers markets and community events or giving presentations to faith
groups.
April 2016 14
-------
Table 2: Campaign Descriptions (Page 1 of 3)
1
2
3
4
5
6
FTGTW Partner
King County
Solid Waste
Division
San Benito
County
Naropa
University
City and County
of Honolulu
(CCH)
Seattle Utilities
City of Palo Alto
Public Works
Department
Community
Location
Fall City, WA
San Benito
County, CA
Boulder, CO
Honolulu, HI
Seattle, WA
Palo Alto,
CA
Urban-Rural
Classification
Peri-Urban
Town;
population
2,000
Rural County;
population
57,600
Mid-Sized City;
population
103,000
Urban;
population
375,000
Urban;
population
650,000
Suburban;
population
66,000
Time of
Year
Fall 2012
Fall 2012
Oct- Nov
2012
Feb-Mar
2013
Jan-Apr
2013
Jun2013-
Present
Campaign Objectives
Test effectiveness of
FTGTW messaging and
tools.
Test food waste
reduction strategies to
inform the county's
future food waste
collection plans.
Bring awareness of food
waste and composting
to students.
Test CBSM strategies
and tools including a
cookbook with recipes
and food waste
prevention tips.
Gather baseline data on
food waste, specifically,
how much of the food
waste in Seattle's
residential waste
stream is edible.
Quantify and reduce the
amount of edible food
waste.
Campaign Focus
King County developed
and tested an elementary
school curriculum on food
waste that invited active
participation of both the
students and their
families.
Households were asked to
photo document
instances of their
preventable food waste
for four weeks.
The campaign had an
educational focus.
CCH was interested in
food waste management
solutions that would both
lower the costs of
landfilling as well as offset
the cost of importing food
to the island.
This campaign focused on
obtaining data on current
food waste management
practices.
Broad media campaign
that included FTGTW info
in outreach for curbside
compost collection
services.
Community
Partners
Local
elementary
school through
Green Schools
Program.
Local food bank
in Hollister,
California.
Naropa
University
Sustainability
Council
Master degree
student in
Environmental
Sciences from
the Univ. of
Gothenburg,
Sweden.
None
None
Target
Audience
The target
audience was
families with a
child in the 4th
grade at the
town's public
school.
Lower income
Hispanic
families.
Student body
and faculty of
university with
enrollment of
approximately
400.
The principal
audience was
young adults.
The average age
of participants
was 34.
Seattle's
residential
population.
General
population.
April 2016
15
-------
Table 2 Continued (Page 2 of 3)
7
FTGTW Partner Community
City of Gresham
Location
Gresham,
Recycling and OR
Solid Waste
8
King County
Urban-Rural
Classification
Suburban;
population
109,000
Time of
Year
Summer
Campaign Objectives
Campaign Focus
Raise awareness and Built partnerships with
Community
Partners
Local grocers
2013- change attitudes and 1 food related organizations
Summer behaviors. (e.g. grocers) to spread
2014 FTGTW messaging.
Target
Audience
General
population
King County, Rural to Urban; Fall 2013 Raise awareness of the Broad county-wide media Puget Sound Families with
Solid Waste WA population benefits of reducing campaign; Produced short Consumer young children
Division
9
10
Rhode Island
2,000,000
Rhode
Food Policy Island, Rl
Council
City of Iowa City
Landfill and
Recycling Center
11
Iowa City, IA
Mostly Urban;
Feb - Oct
population 2014
1,000,000
Mid-sized City;
food waste and videos on FTGTW (Food) Coop with a strong
encourage residents to strategies in partnership focus on the
try behaviors that help with food co-op grocer.
reduce wasted food.
Test campaign
Testing by four cohorts
Providence
adult female in
the household
Recruited 4
effectiveness with provides opportunity to Housing cohorts of ten
different demographics, refine outreach and Authority each: friendlies,
in particular, to messaging to social group. RIFPC list-serve
understand how low
income households
respond to waste
reduction as a food
security strategy.
Mar-Jul
population 2014
71,600
Gain experience in
helping residents
reduce food waste.
Educational effort to teach
subscribers,
Refuse Division,
high-income
apartment
residents; low-
income PHA
residents.
Select
residents benefits of City of Iowa City neighborhoods
reducing waste and chosen to
support for curbside 1
collection of food waste
with yard waste.
Thurston County Thurston Small Feb-Dec Engage Thurston County Test impact of combined None
Solid Waste | County, WA | Town/Rural; 2014 in reducing household broad-scale media
population
262,000
food waste.
awareness campaign with
on-the-ground household
engagement.
represent a
variety of
incomes and
ages.
General
population
including
outreach
through
schools.
April 2016
16
-------
Table 2 Continued (Page 3 of 3)
12
13
14
15
16
17
FTGTW Partner
Village of Oak
Park
Environmental
Services
Addison County
Solid Waste
Management
District
Sustainable
Jersey City
King County
Solid Waste
Division
Kanu Hawaii
City of Aurora
Community
Location
Oak Park, IL
Addison
County, VT
Jersey City,
NJ
King County,
WA
Hawaii, HI
Aurora, CO
Urban-Rural
Classification
Suburban;
population
52,000
Small
Town/Rural;
population
37,000
Urban
Rural to Urban;
population
2,000,000
Urban
Urban;
population
346,000
Time of
Year
April-May
2014
May-Jun
2014
(before
garden
season)
Jun-Aug
2014
Fall 2014
Sept - Oct
2014
Nov-Dec
2014
Campaign Objectives
Promote waste
reduction by linking it to
cost savings.
To facilitate the
district's source
reduction efforts.
Test integration of
waste prevention and
Bokashi composting
strategies.
Raise awareness of the
benefits of reducing
food waste, both
financial and
environmental and
encourage residents to
try strategies.
Test how to leverage
social networks to
accelerate community-
wide awareness and
adoption of FTGTW
strategies.
Create awareness
around greenhouse
emissions from wasted
food.
Campaign Focus
Test cost savings message
that participating in
FTGTW will offset the
monthly cost of
participating in the
compost program.
Vermont has enacted law
that will ban all food from
landfills by 2015; no
compost curbside
collection for most towns
Educational effort with
Bokashi composting
program participants.
Test effectiveness of
different CBSM outreach
methods and messengers.
Empowering people to
build more sustainable,
and resilient communities
rooted in personal
commitments to change.
Determine best practices
in encouraging citizens to
contribute to reducing the
City's environmental
footprint.
Community
Partners
Seven
Generations
Middlebury
Natural Foods
Coop; Addison
County Re-
Localization
Network
Jersey City
Environmental
Commission;
community
gardens
Master
Composters
The RISE
program, Kupu.
Kaiser
Permanente;
Denver Urban
Gardens
Target Audience
Village residents
who did not
participate in
the compost
pilot
Families with
children
Sustainability
focused adults
and community
gardeners
Farmers market
customers
Social network
of 20,000+
members
City employees
April 2016
17
-------
Community Partners: A few of the implementing organizations used CBSM consultants.
Others partnered with local non-profits with whom they had existing relations. Food coops were
another choice of partner for several campaigns (Campaigns 8, 13). In 2014, King County also
partnered with Master Composters for outreach.
Target Audience: "Target audience" or "target population" refers to the demographic the
implementing organization hopes to engage in the pilot. The most common target audience was
a small sample of the general population. In some cases, there was additional effort to reach
young adults (for example, the Naropa University campaign) or families with children (King
County). The Rhode Island Policy Food Council targeted four different demographics including
high- and low-income cohorts.
While the goal of food waste reduction was common to all campaigns, the communities' needs
and resources determined the implementation means. Differing levels of resources included
familiarity with the CBSM approach, community size and demographics, and partner
relationships, as well as varying levels of available staff and funding. Community partners made
implementation choices based on a variety of such factors. These factors are cited below in the
assessment when relevant to the campaigns' outcomes. Table 3 provides an overview of the
implementation details by campaign.
Outreach and Engagement Methods: Campaigns used those outreach and engagement
means adapted to the needs of their communities and available resources. Many of the
campaigns that conducted challenges held workshops while those engaged in broader scale
outreach relied on social media. The recruitment methods were especially varied, ranging from
tabling at farmers markets to invitation letters.
Conducted Challenge: All of the campaigns used the full complement of FTGTW strategies
with the exception of the second year campaigns in King County and Hawaii (Campaigns 8 and
16) which did not conduct Challenges. In many campaigns, outreach was limited to recruiting for
the Challenge. In effect, the Challenge was synonymous with the campaign. Broader-scale
campaigns, however, engaged a wider sweep of participants in using strategies and tools
independent of the Challenge.
Length of Challenge: Most Challenges ran from 4 to 6 weeks. Participants in Seattle's baseline
waste audits weighed their preventable and inedible waste fractions for 13 weeks.
Fraction Measured and Method Used: Campaigns made the choice of which waste fraction to
measure and whether to use weight or volume measurement in relation to their objectives and
resources. In general, there was a balance between these two choices.
Target Population Size: The majority of campaigns targeted a demographic or demographics
within their population but also chose to do blanket outreach to their communities at large. The
Rhode Island and Iowa City campaigns (Campaigns 9 and 10) are examples of more targeted
outreach.
Implementation Outcomes: The last three columns in Table 3 convey implementation
outcomes, including the number of households recruited for Challenges and a given Challenge's
sample size and retention rate. These are referred to in the assessment of Outreach and
Engagement effectiveness.
April 2016 18
-------
Table 3: Outreach and Engagement Descriptions and Results (Page 1 of 3)
1
2
3
4
5
Community
Outreach and Engagement
Location Methods
Fall City,
The invitation to participate was
King emailed to families, followed by a
County visit to the classroom. The
2012 curriculum was ongoing
San Benito
throughout the five weeks. All
families who completed the
challenge were given a grocery
store certificate.
Fliers were placed in food bank
County bags to invite participation,
Boulder
followed by phone calls and email
(when available), while the
workshop presentation was used
to recruit participants from the
seniors and the moms' group.
Conducted Length of Fraction Measurement Target Number of
Final Retention
Challenge Challenge Measured Method Population Households Sample Rate
Yes
Yes
Initial outreach was through an 1 Yes
email announcement. Participants
Honolulu
were also recruited through
tabling at the University's
Sustainability Fair at the end of
which the workshop presentation
was made. All subsequent contact
was through email.
Recruitment was made by email
Yes
using personal contacts in two
Seattle
social networks. Two workshops
were held, the second came after
the baseline weeks to introduce
strategies.
Study participants were recruited
through a short article in the
utility's newsletter that is mailed
to residential customers with
their bi-monthly bill. Participants
weighed daily and submitted
results weekly on-line.
Yes
5 weeks
4 weeks
4 weeks
All solid
waste
All solid
waste
All solid
Volume
Photo diary
Weight
waste
4 weeks
All solid
Weight
waste
13 weeks
Preventable
and non-
edible
fractions -
baseline
only
Weight
Size
110 families
Recruited
47 families
Size
13 families
submitted (another
Weekl
data
560
500
210
All utility
customers
11 families
completed
five
weeks)
20
65
10
1
individuals
picked up
Challenge
materials
17
125
14
119
28%
50%
2%
82%
95%
April 2016
19
-------
Table 3 Continued (Page 2 of 3)
6
7
8
9
10
11
Community
Outreach and Engagement
Location Methods
Palo Alto
Gresham
Campaign outreach means
included traditional media
Conducted Length of Fraction Measurement Target Number of
Final Retention
Challenge Challenge Measured Method Population Households Sample Rate
Yes
(limited)
(newspaper and publication print
ads), a bill insert, and social media
(online, Facebookand Pandora
ads). Transitioning to greater
emphasis on peer-to-peer and
direct-personal-contact outreach
in 2014.
Tabling at farmers markets, art
Yes
walk, car show and at grocery
King
stores.
Strategy was to drive awareness
No
County and engage with the community
2013 through broad-based
communication such as
Rhode
Island
advertising, social media and
media relations.
Engagement workshops focused
on hands-on demonstrations of
waste prevention strategies.
Campaign arrange for low-income
participants to receive community
credits as incentive to participate.
Yes
Iowa City, Intense personalized recruitment: Yes
IA
Thurston
County,
an invitation to participate was
sent to select neighborhoods
followed by door hangers and
neighborhood open houses.
Extensive social media presence
was combined with in-person
WA educational presentations made
to a variety of community groups
and schools.
Yes
Size Recruited Size
6 weeks
Preventable
fraction
only
6 Weeks
Preventable
Volume
City
Not Not
Not
population available available available
Weight
fraction
Not
only
Not
Not
City
population
Not
31
Not
14
Not
45%
Not
applicable applicable applicable applicable applicable applicable applicable
6 weeks
Preventable
fraction
only
Both weight
and volume
6 weeks 1 Preventable Weight
and non-
edible
fractions
4 weeks
Preventable
fraction
only
Volume
Not
available
300
households
County
population
39
52
80
22
29
42
56%
56%
53%
April 2016
20
-------
Table 3 Continued (Page 3 of 3)
12
13
14
15
16
17
Community
Location
Oak Park, IL
Addison
County, VT
Jersey City,
NJ
King
County
2014
Hawaii, HI
Aurora, CO
Outreach and Engagement
Methods
Blanket recruitment through
notice in Village of Oak Park's
newsletter.
Outreach occurred through
partners and educational
workshops. Also, a recruitment ad
was placed in local newspaper
offering grocery coupon incentive
for participation.
Recruitment to Bokashi
composting program with
community gardens with FTGTW
presented as first step in Bokashi
process.
Trained master composters to
disseminate FTGTW materials and
recruit Challenge participants at
farmers markets.
Messages focusing on reducing
residential food waste were
shared through Kanu Hawaii's
20,000+ member network on an
almost daily basis. These
messages included images, text,
and some videos.
Promoted through city's intranet
website that most employees visit
at least once per workday.
Conducted
Challenge
Yes
Yes
Yes
Yes
No
Yes
Length of
Challenge
6 Weeks
6 weeks
6 weeks
4 weeks
Not
applicable
6 weeks
Fraction
Measured
All solid
waste
Preventable
and non-
edible
fractions
All solid
waste
Preventable
fraction
only
Not
applicable
All solid
waste
Measurement
Method
Volume
Volume
Volume
Volume
Not
applicable
Volume
Target
Population
Size
City
population
District
population
1200
network
members
Farmers
Market
customers
Not
applicable
2650 city
employees
Number of
Households
Recruited
25
Not
available
25
71
Not
applicable
72
Final
Sample
Size
12
31
Not
available
53
Not
applicable
24
Retention
Rate
48%
Not
available
Not
available
75%
Not
applicable
33%
April 2016
21
-------
3.2 CAMPAIGN EFFECTIVENESS
A principle evaluation objective is to determine the extent to which FTGTW campaigns result in
the desired behavior changes. This includes assessing how effective campaigns are in both
generating and sustaining the desired behaviors. The result of these changes, that is, the
impact on the amount of food going to waste, is covered in Section 3.3.
A CBSM campaign typically consists of behavior change strategies and tools, messaging, and
outreach and engagement. The effectiveness of these three components are evaluated in turn:
Section 3.2.1 evaluates how effective the behavior change strategies and tools are at engaging
households in the desired behaviors, while Section 3.2.2 and 3.2.3 look at messaging and
outreach effectiveness respectively. Key findings in each section are bolded.
Household factors affecting reach and effectiveness are discussed in Section 3.2.4.
Lastly, in addition to what the community partners did individually to implement campaigns,
another critical factor in their success was the support provided by the EPA and to each other.
This support is addressed in Section 3.2.5.
3.2.1 Effectiveness of Behavior Change Strategies and Tools
This section of the report focuses on assessing how useful households found the FTGTW
strategies and tools in changing their behaviors. It answers such questions as: Are the
strategies and tools useful and easy to implement? How likely are households likely to continue
to use the strategies and tools? And do the strategies and tools increase awareness of wasted
food in the household? The assessment includes an analysis of the campaign participants'
experience and quantitative measures of their engagement. Post-challenge participant
questionnaires provided the main source of quantitative data.20
In general, households gave high marks, both qualitatively and quantitatively, to the
FTGTW strategies and tools. Numerous comments from campaign participants indicated they
greatly valued their experience.21 Three such comments are:
"This was a wonderful learning experience and taught us all how to be more mindful of
the food we consume and the food we bought but didn't consume." (King County)
"I think this was a really good idea. I hope more people decide to do this." (Rhode Island)
"This was a very wonderful learning experience and I have already noticed a change in
my house in regards to the amount of money we are saving because we plan ahead."
(Aurora)
Post-challenge survey responses support these observations. Figure 1 shows that 86% of the
respondents found the strategies and tools useful (n=70).22 Participants who already considered
themselves knowledgeable in preventing wasted food tended to be more neutral in their
evaluation of the strategies and tools' usefulness.
20 In addition to collecting measurement data from participating households, the Rhode Island, Iowa City and Aurora campaigns
chose to do surveys of Challenge participants before and after the Challenge.
21 Comments were solicited along with measurement data by the majority of campaigns conducting Challenges.
22 The combined number of post-survey responses for the Rhode Island, Iowa City and Aurora campaigns was 70 (or n=70).
April 2016 ~22~
-------
Figure 1: Usefulness of Strategies and Tools
Survey Response to the Statement:
"In general, I found the strategies and tools useful."
Strongly agree Agree Neutral Disagree Strongly disagree
Iowa City Challenge participants (n=26) were also asked whether they found the strategies and
tools easy or hard to use. 95% responded that they found them easy or somewhat easy to
use.
96% of the households also indicated that they are likely to continue to use the tools and
strategies as shown in Figure 2, indicating there are no on-going significant barriers to use of
the strategies and tools.
Figure 2: Likelihood of Continuing to Use Tools and Strategies
Survey Response to the Statement: "I am likely to
continue to use the tools and strategies.
Agree Somewhat agree Somewhat disagree Disagree Don't know/unsure
The evidence on the effectiveness of specific strategies and tools indicates that strategies
associated with a tool are considered more useful than those without a tool. This is
consistent with CBSM research. As an example, the Eat First Prompt was developed in 2013
after other tools. In later campaigns after the introduction of the prompt, the associated strategy,
Smart Saving, was seen as equally or more useful to the Smart Storage and Smart Shopping
strategies, where before it was not. Also, the Smart Prep strategy which has no associated tool
April 2016
23
-------
was rated as least useful of the five strategies. It is also likely that there are time barriers to the
Smart Prep strategy.
Although the CBSM literature advises campaigns to concentrate on one behavior at a time, new
research suggests a focus on practices has a greater impact, especially when there is a
cascade of behaviors with a compounded impact as is the case with food waste behaviors.23 It
is difficult to interpret from the available data whether targeting one or two strategies versus a
suite or mix of strategies would be more effective. So while there may be some expressed
preferences for certain strategies, it may be premature to limit the suggested strategies to only
those perceived as the most useful.
To test whether targeting two strategies was more effective than targeting the suite of FTGTW
strategies, Gresham conducted two rounds of the Challenge. In the first round, participants were
given information on the Smart Saving: Eat What You Buy Strategy as well as measuring their
waste (the Get Smart strategy). In the second round, all five strategies were introduced to the
participants. The group that used all five strategies reduced their waste by a greater amount
than did the one focusing on the Eat What You Buy strategy. As the samples in the Gresham
challenges were extremely small, caution should be used in interpreting the results but they do
support the idea of introducing a greater number of strategies.
In addition, there is a great deal of variability on household food management practices related
to a number of demographic and lifestyle factors. Even within a target demographic there is
considerable variability in practices. It is therefore difficult to design a campaign that singles out
the strategy that is most relevant for each demographic or lifestyle. On the other hand,
campaigns featuring a range of strategies allow each household to focus on the strategy
or strategies that work best for them but still targeted toward the ultimate behavior of
reducing wasted food at home.
A key measure of strategy and tool effectiveness is the degree to which they shifted awareness
of the households' tendency to waste food. The most effective strategy in this regard was the
Get Smart Strategy wherein households measured the amount of food going to waste in their
households as part of a challenge. (As noted previously, the FTGTW Challenge incorporates all
five strategies but its primary focus is having households measure the amount of food going to
waste in their households.)
Previous research indicates that most people underestimate the amount of food going to waste
in their households.24 For example, in a recent national survey, Dr. Roni Neff at John Hopkins
found that 73% of Americans thought they wasted less food than average.25
An unexpected finding from the early pilots was that having households measure their waste
strongly motivates their desire to reduce waste. Initially, the Challenge measurement tools were
intended as a means of collecting data to evaluate the impact of waste reduction strategies but
the measurement tools' value in raising awareness quickly became evident.
One possible intervention to a lack of reflectivity around household waste practices, that is,
automatic or habitual behavior, is to create a feedback mechanism to stimulate awareness. The
23 Quested, I.E., ibid.
24 Cox, Jayne, and Phil Downing. 2007. Food Behaviour Consumer Research: Quantitative Phase. WRAP Report; Glanz, Robert.
2008. Causes of Food Waste Generation in Households - An Empirical Analysis. Applied Sciences.
25 Neff RA, et al. 2015. Wasted Food: U.S. Consumers' Reported Awareness, Attitudes, and Behaviors. PLOS ONE.
April 2016 ~24
-------
measurement tools used in the Challenge served this purpose. In addition, it appears that
measuring waste taps into our neuro-psychological tendency to dislike waste which, in turn, is a
strong motivator to try strategies to reduce waste.
A number of participant observations were made indicating the effectiveness of the Challenge in
raising awareness and thus motivating people to reduce waste, among them:
"[I was] shocked at how much we've actually been wasting." (Honolulu)
"This was a great way to teach/show our kids how much food gets wasted and how
to change our habits to be more efficient." (King County)
"Participating definitely made me more conscious about my patterns, including
buying too much food and not eating it before it went bad." (Thurston County)
Challenge participants were also asked in the post-Challenge survey whether the Challenge
raised their awareness of how much food was going to waste in their households. Their
combined responses (n=69) are shown in Figure 3. 91% of the respondents agreed that
participating in the Challenge raised their awareness of food going to waste in their
households and over half strongly agreed.
Figure 3: Challenge's Effect on Increasing Awareness of Wasted Food in Household
Survey Response to the Statement:
"lam now more aware of food going to waste in my household"
Strongly agree Agree Neutral Disagree Strongly disagree
In addition, 42 Challenge participants in Thurston County, WA rated the Challenge as the most
effective of eight different tools. While, in Seattle, the amount of wasted food trended down even
though no food reduction messaging or strategies were introduced as part of their home waste
audits.
The take-away in these observations and data is that tracking the amount of wasted food
motivates action to reduce wasted food. In effect, feedback increases awareness by
countering habitual behavior and activating waste aversion. While there are barriers to
engaging people in measuring food waste over an extended period of weeks, tools designed to
provide feedback and raise awareness, such as the FTGTW Challenge, highly motivate
households to reduce wasted food.
April 2016 "~2?
-------
As described in Section 2.2, campaigns used different Challenge procedures according to their
objectives and available resources. While all the methods were effective in raising awareness,
they were not equally effective in measuring the amount of food going to waste. For example,
the San Benito pilot used a photo diary to document food waste. The method was too complex
to obtain an accurate measure of the amount of food going to waste as evidenced by
inconsistencies in the data. This method was not offered in the 2013 and 2014 campaigns.
Three implementation choices appear to influence the impact results: whether to use volumetric
or weight measurement; what portions of the waste stream to measure; and the length of the
Challenge.
The trade-offs between volume and weight measurement methods are summarized in Table 4,
followed by a discussion of the trade-offs.
Table 4: Volume versus Weight Measurement Trade-offs
Trade-off
Volume Measurement
Accuracy of Measurement Less accurate
Weight Measurement
More accurate
Ease of Use in Households
Less time consuming/more
convenient
More time consuming/less
convenient
Clarity of Method to
Households
Costs
More complex
Less complex
Lower costs
Higher costs
While volumetric measurement is very effective in creating a visual sense of how much goes to
waste that households can refer to on an on-going basis, it is less exact than weight
measurements given the considerable variability in what goes to waste in different households
and at different times in the same household. This has to do with some foods being denser than
others. Volume measurements are less accurate than weight measurements for additional
reasons: some participants might tamp down the volume, while others do not; people may have
difficulty accurately reading the volume if measurement indicators are on the outside of non-
transparent containers; and volume measurements are less precise than weight as participants
were instructed to round off to a fraction of the volume. Also, when different size containers are
used, it makes it more difficult to compare measurements. In sum, volumetric measurement is
more geared to providing households a convenient waste yardstick but weight
measurements provide for more accurate accounts of food waste, both for the household
and as a data source for implementing organizations.
Another measurement variable in the different Challenges was the portion of waste recorded. It
seems important to have households focus on that portion of food waste they might prevent.
While data on the total amount of waste is useful for solid waste agencies, there is a great deal
of seasonal variability in the amount of inedible waste which may obscure a reduction in
preventable waste. In the Seattle and Honolulu campaigns, participants recorded both the
preventable and non-edible portions of waste but separately. From a research perspective, this
data is important to determining what the limits might be on waste reduction at the community
level.
It should also be noted that the instructions on separating preventable waste, that is, edibles,
from inedible waste were not always provided and this was largely left to households to
determine. In particular, discarding food waste that is not already separated during food
preparation, for example, a discard chicken carcass with meat still on it, introduces a level of
April 2016
26
-------
uncertainty into the accuracy of the measurements. It is recommended that Campaign
instructions be amended to address how to separate from inedible from preventable waste
when discarding.
Collecting data on both the edible and inedible fractions of household food waste allows
for a comparison to municipal waste stream data. By contrast, collecting data on the
preventable fraction alone focuses household attention on the potential impact from
adopting waste prevention strategies. Seasonal variability of the total waste stream may
mask reductions in the preventable fraction.
Measurement periods varied from four to six weeks for the different pilots. In the Honolulu pilot,
several Challenge participants felt that four weeks was too short a time period either to collect
data or to establish waste reduction habits. In the 2014 campaigns, it was strongly
recommended that households measure for two weeks to establish a baseline amount and then
four weeks to determine the impact of the waste reduction strategies. Campaigns that opted to
do fewer than these recommended time frames reported poor results. Also, the length of the
Challenge did not seem to correlate to the retention rates (see Table 3). Routinized data
collection seemed to make participation less problematic as evidenced by the high retention rate
in Seattle where they measured wasted food for 13 weeks.
In sum, less than six weeks is too short a time to provide an accurate idea of the
reduction potential but longer Challenges may affect participation rates, although the evidence
from different pilots seems to suggest that the means of outreach are a more important factor in
determining participation.
A final question with regard to the efficacy of the Challenge as a measurement tool is the
accuracy of having households perform their own waste audits. The most commonly
recommended method of quantifying household food waste is to conduct waste characterization
studies which involves the use of trained waste handlers to measure the waste.26 FTGTW has
pioneered self-audits as a cost-effective means of collecting data on household waste patterns.
To determine whether waste audits are sufficiently accurate compared to a waste
characterization study is beyond the scope of the FTGTW mandate. However, the trade-offs
between the two methods merit comment. Self-auditing provides for more granular data than do
waste characterization studies and, significantly, are instrumental in raising household
awareness. Attributing changes in the amounts wasted to behavior changes is more difficult with
waste characterization studies than with self-audits. The major drawback of self-audits is the
resources needed to obtain a statistically robust sampling are cost prohibitive.
More research is indicated to verify whether audits introduce bias in the results towards under-
reporting the amounts wasted. In the FTGTW campaigns, it was found that clear and simple
Challenge instructions to households increase the accuracy of data collection and reporting.
Based on the measurement results, requesting households weigh their waste appears to be
an effective means of determining the potential for reduction in small to medium-sized
sample populations, especially when there are limited resources to conduct waste
26 World Resources Institute. 2015. Draft FLW Protocol Accounting and Reporting Standard (FLW Standard), http://www.wri.org/our-
work/project/global-food-loss-and-waste-measurement-protocol/documents-and-updates#project-tabs
April 2016 ~~27
-------
characterization studies. Appendix C describes a community-scale measurement protocol that
combines waste characterization approach with self-audits.
3.2.2 Effectiveness of Messaging
Data on the effectiveness of the messaging content came from three sources: the pre- and post-
Challenge surveys; observations from campaign staff; and social media statistics.27 In addition,
the results from a recent nationally representative survey on U.S. consumers awareness,
attitudes and behaviors related to food waste are presented.28 This survey was conducted by
Dr. Roni Neff of the John Hopkins Center for a Livable Future.
One concern during the design phase of the campaign was that the campaign messaging not be
perceived as "preachy" or condescending. There were no reported incidents of complaints about
the tone of the messaging. Instead, campaign volunteers and staff reported significant
expression of interest at tabling events and lively and spirited conversations at
workshops.29 These results are likely an effect of CBSM principles - rather than instruct
participants not to waste food as might be typical of a more traditional education campaign,
CBSM offers interested participants tools and strategies for achieving the desired behavior
changes.
The positive interest probably also reflects the issue's relevance to households. In the John
Hopkins national survey, 88% of the respondents indicated interest in reducing food waste,
while only 12% of the respondents said that they were not at all interested. Likewise, in
Gresham, 78% of participants in an event tabling survey (n=94) answered that they would like to
reduce wasted food in their households.
One measure of messaging effectiveness is the increase in the percentage of the target
population that expresses awareness of wasted food as an issue which requires action.
Responses to pre-Challenge surveys serve to provide a baseline for establishing how much of
the population is aware of wasted food as an issue. The questionnaires asked respondents
whether they had seen or heard anything about the problem of wasted food in the last year.
Their responses are shown in Figure 4. 42% and 44% responded yes in Iowa and Rhode Island
respectively, which compared to 42% in the national survey.
Figure 4: Baseline Awareness (In the past year, have you seen or
heard anything about the problem of wasted food?)
Yes Of "yes" responses, could state Of "able to state" responses,
what they heard indicated awareness of
household waste
Rhode Island Iowa City Aurora National Survey
27 This section looks at messaging content. Message delivery is considered in the context of outreach in the next section.
28 Neff RA, et al. 2015. Wasted Food: U.S. Consumers' Reported Awareness, Attitudes, and Behaviors. PLOS ONE.
29 Event tabling is the presentation of relevant materials, usually laid out on a table, at public events.
April 2016 ~~2S
-------
Respondents were also asked to state what they had heard. If we consider the "yes" and "able
to state" responses, the data indicate that awareness of wasted food as a problem is not
necessarily linked to taking action to reduce waste at home. While 42 to 44% might be
considered relatively high percentages, only half of the "yes" respondents in Rhode Island could
say what they had seen or heard and only 10% of the these responses, in turn, indicated an
awareness of household waste. Other responses had to do with such things as composting and
landfilling waste or food recovery efforts. In Iowa City, 100% of the "yes" respondents could
state what they had seen or heard and over 40% of the "able to state" responses indicated an
awareness of wasted food as a household issue. These results suggest that the Iowa City
sample showed a higher awareness of wasted food as a consumer problem, prior to the
Challenge, than did the Rhode Island and Aurora samples.30
To understand the impact of campaign messaging on awareness, the Iowa City and Rhode
Island pre- and post-Challenge questionnaires asked to what extent different rationales
motivated the respondents to minimize the amount of food being thrown out by them. The rating
system was as follows: (1) Not at all (2) A little (3) A fair amount (4) A great deal. Their
averaged responses are shown in Figure 5.
It can be seen from the figure that over half of the rationales concerned respondents "a fair
amount" to "a great deal", (between 3 and 4 on the scale). The rankings indicate that feeling
bad about throwing away food and wasting money appear to be equally strong
motivators to reduce wasted food. In a similar question, the national survey included a
slightly different selection of motivations. It was found that saving money was the most
important motivation with setting an example for children ranking second with parents. They
found no geographical differences in the rankings.
Figure 5: Ranked Motivations to Reduce Wasted Food
f. o
3,5
3.0
2.5
2,0
1.5
1 0
0,5
0,0
Wasted
money I
spent
buying the
food
Wasted Thatthere The wasted Feeling bad The The amount
time I spent are people
shopping, without
storing, enough to
and/or eat
preparing
food
energy and
water
resources it
took to get
the food to
my plate
about
throv/ rg
away food
that could
K=ve beer-
eaten
contribution of food that
of wasted ends up in
food to landfills
global
warming
i Rhode Island Pre
i Iowa City Pre
Rhode Island Post Iowa City Post
1 One possible explanation for Iowa's higher awareness is their proximity to food production.
April 2016
29
-------
Comparing the rankings before and after the Challenge reveals some tentative conclusions
about FTGTW messaging. Messaging intending to raise awareness of wasting energy and
water resources by throwing out food was a key component of the FTGTW campaigns. Concern
for wasted resources rose in both Rhode Island and Iowa City, in Iowa City it ranked third
following the Challenge. However, In Rhode Island it was second to last of the seven
motivations after the Challenge. Both in Rhode Island and Iowa City, the lowest ranked
motivation prior to the Challenge was the contribution of wasted food to global warming. In Iowa,
the degree of concern fell from before to after the Challenge, where in Rhode Island it increased
by half of a scale point. Overall, these responses suggest that increasing awareness of the
indirect environmental effects of wasted food through messaging is challenging with
mixed results.
Campaigns also found that there is the need to particularize environmental messaging to
the household level for greatest effect. An example of this is equating throwing away an
apple to flushing the toilet seven times. During the Thurston County outreach presentations to
community groups, the outreach staff observed that talking about the amount of resources
embodied in a single food item made a greater impression on the audience than talking about
wasted resources in the abstract. Also, pointing to the socio-economic consequences of wasting
embodied resources, such as linking rising food prices to wasting the scarce resources used to
produce and distribute food, make the significance of the environmental impacts easier to
understand.
Additional information on messaging effectiveness comes from two of the community partners
who collected data on their social media campaigns - Thurston County Solid Waste and Kanu
Hawaii. Three measures were used to evaluate the effectiveness of individual messages: reach
(the number of unique users who received at least one impression of the post), engagement
(the sum of clicks, likes, engagements and shares) and stickiness (engagement divided by
reach). The Kanu Hawaii Facebook campaign was conducted in August 2014. The content of
their Facebook posts relate to a variety of environmental and socio-economic issues. By
contrast, the Thurston County Facebook campaign began in March 2014 and continues but their
Facebook community was built from scratch and their page is dedicated to raising awareness
around wasted food.31 At the end of 2014, they reached over 2,500 page likes. A summary of
the media statistics on the top individual messages follows.
In the 20,000+ Kanu Hawaii network, a posting on the shelf life of different foods had the
greatest reach at 38 thousand impressions, exceeding the number in their network. The
message with the greatest amount of stickiness was a posting on how to reduce browning of
avocados at 12.4%. In Thurston County, the post with the greatest reach at 2900 was one
asking if people had taken the Challenge, while the message with the greatest stickiness at 25%
was one asking opinions on whether supermarkets should sell fruits and vegetables that do not
meet their "normal" standards.
It should be noted that both campaigns' choices about what to post reflected their perception of
their communities' needs, so the content was not necessarily equivalent. In Hawaii, there was a
conscious decision to "emphasize the importance of local perspective in the campaign's
content". Messages harvested from FTGTW materials were discarded as being too "mainland"
and global. In Thurston County, only a handful of posts included content pointing to the
consequent environmental problems of climate change or wasted resources, so determining
31 Behavior change tools were distributed as part of the Thurston County campaign but not in Kanu Hawaii's campaign.
April 2016 ~~3o"
-------
how effective this messaging was compared to other messaging types (e.g. tips and hints) is
problematic. However, one post asking people to share the post if they agreed with the
statement that "All food production uses resources that we rarely see or think of" did garner a
stickiness rate of 17% and a reach of 1300.
A more systematic examination of messaging content would need to be undertaken to draw
definitive answers on what type of social media messaging is most effective. One tentative
conclusion is that posts asking the audience to respond through a question or to demonstrate
their knowledge have higher levels of engagement and stickiness. Also, more content with
localized environmental messaging is needed.
Both building a new network to focus on wasted food (Thurston County) and utilizing an existing
network to achieve reach (Kanu Hawaii) appear to have advantages. The results suggest that it
might be interesting to combine these two approaches in future campaigns.
3.2.3 Effectiveness of Outreach and Engagement
In addition to customizing outreach and engagement methods to the needs of their
communities, the amount and type of available resources for a given campaign also had bearing
on the methods used. These differences can lead to qualitatively different results. In
consequence, it is possible to make only general statements about the effectiveness of
particular outreach and engagement methods given the limited number of campaigns
conducted. It is also too early to tell whether the pilots will result in sustained behavior changes.
Both qualitative and quantitative data were collected to evaluate the effectiveness of the
campaigns' outreach and engagement. In addition to tracking outreach efforts such as
Challenge recruitment and retention rates, campaign staff also reported their observations on
their interactions with target audiences.
Generally, it was found that campaigns designed to leverage social networks and create
social norms were among the most effective in terms of outreach and engagement. Clear
campaign objectives along with an understanding of how these related to the implementation
means also led to more successful campaign outcomes as defined by the community partners.
Finally, campaigns that had staff who were experienced in community outreach were able to
benefit from this experience.
To track outreach efforts, campaigns collected data on Challenge recruitment. However, caution
should be used in interpreting Challenge recruitment numbers as a measure of a campaign's
reach, that is, the percentage of the population that changed behavior from exposure to the
campaign. The number of people influenced to change their behavior will be greater than the
number of people recruited to take a Challenge.
While the majority of the campaigns focused on implementing small-scale Challenges, several
of the 2013 and 2014 campaigns had broad-scale media components. These included the King
County campaigns in 2013 and 2014, and the Palo Alto, Thurston and Kanu Hawaii campaigns.
In the small-scale pilots, outreach was directed primarily to potential Challenge participants. By
contrast, the broad-scale campaigns focused on building awareness of the targeted population
using a variety of engagement techniques such as social media, traditional media, and videos.
In between small-scale Challenges and broad-scale Campaigns are community-scale outreach
initiatives such as tabling and presentations to community groups and networks. These sustain
one-to-one and peer-to-peer connections while increasing the numbers reached.
April 2016 31
-------
Table 3 shows the recruitment outcomes. The results indicate there was varying success in
recruiting households to take the Challenge. CBSM research has established that direct
personal outreach is more effective than broad scale media campaigns. The recruitment data
support this conclusion. In particular, community-scale direct outreach, such as used in the
Thurston and King County 2014 campaigns, was more effective than recruitment through
indirect means such as social media outreach.
Campaign staff reported that the response to FTGTW
community-scale direct outreach efforts was
enthusiastic. A ten year veteran of community
outreach campaigns in Thurston County observed that
she had never seen as high a level of interest and
gratitude for bringing an issue forward. Other
campaign volunteers reported that, during tabling,
people commented that they thought FTGTW was a
good use of public dollars.
In terms of venues, tabling at farmers markets (King
and Thurston Counties) led to greater engagement
than did tabling at grocery stores (Gresham) as
reported by campaign staff and volunteers.
The U.K.'s Waste and Resources Action Programme
(WRAP) hypothesizes that direct or one-to-one
outreach allows campaigns to tailor wasted food
reduction strategies to the individuals.32 Another
explanation is that one-to-one and peer-to-peer
outreach engages social learning and the creation of
new social norms, as well as the opportunity to leverage networks.
Leveraging existing networks through cascade training, e.g. "train the trainer", approaches is
one possibility for cost effective community-scale campaigns. King County, in its 2014
campaign, partnered with Master Composters to do tabling at farmers markets and report that
their engagement in outreach elevated the network's knowledge about the benefits of waste
prevention. A variant of this approach is to organize a campaign around "food waste champions"
who, equipped with campaign materials, perform outreach to their personal networks.
In addition to direct personal contact, more personalized appeals to participate also appear to
be more effective. For example, in Iowa City, an outreach letter to potential Challenge
participants began with their having been "chosen" to participate. Over 50 people out of the 300
receiving the letter responded. (See Section 4.1 for further discussion of Iowa City's outreach.)
In this regard, campaigns that focused their outreach were more successful in recruiting
participants as measured by the percentage recruited of the target population size than those
who chose blanket approaches. The recruitment rate for those using non-targeted outreach did
not exceed 2%. Iowa City achieved the highest recruitment rate at 17%. In addition, by
identifying target neighborhoods, their sample was more representative of the target population
as a whole.
32 Quested et al. 2013. Spaghetti Soup: The Complex World of Food Waste Behaviors. WRAP is the U.K.'s premiere food waste
reduction organization. They have conducted extensive research on the causes of wasted food and behavioral interventions.
April 2016
32
-------
To date, creating a new social norm around wasting less
food has not been a primary focus of campaigns. One
outreach tool used by Thurston County and King County in
its 2014 campaign that serves this purpose was the public
commitment, an example of which can be seen in the
accompanying photograph. People visiting the FTGTW
table at famers markets were invited to have their
photograph taken with a chalkboard on which they had
written their "commitment" to trying a reduction strategy. In
turn, these photographs can be shared through social
media sites like Facebook and Instagram. This is a fun way
of making the new behaviors visible.
As noted above, the Challenge recruitment rates should not
be read as a measure of a campaign's overall reach,
especially in broad scale campaigns. There are many
barriers to households measuring their wasted food over
several weeks, from the practical, such as being away from
home during part of the Challenge period, to more
aesthetic reasons, such as a concern about odors. In Thurston County, there was also feedback
that many households already knew they wasted more food than they wanted to and wished to
start trying the strategies immediately instead of measuring their baseline amounts. To
determine the full impact of a broad-scale campaign, and thereby its effectiveness, it will be
necessary to conduct a community-wide measurement program that combines surveying and
some method of waste characterization.
Recruitment rates reflect outreach effectiveness, while retention rates are more an indication of
successful engagement techniques. Almost all of the campaigns achieved a fifty percent or
greater retention rate.
With respect to engagement, the general rule for successful campaigns was to engage
participants early and often. For example, the 2014 King County campaign emailed
participants weekly with encouragement, tips, reminders and incentives in the form of an
opportunity to "win" waste reduction-related products through participation. Frequent
engagement also likely reinforces the sense of belonging to a group or network, thus enhancing
pro-social motivations.
Another successful engagement technique was to engage challenge participants through a
variety of learning techniques. For example, Rhode Island conducted hands-on workshops in
which the workshop presenter, a chef, demonstrated measurement techniques and strategies.
This type of social learning and learning by doing reduces the knowledge barrier to wasting less
food. The efficacy of this approach is likely reflected in the higher percentage waste reductions
achieved in Rhode Island.
Finally, it is clear that without a focused effort a campaign can fail. Two campaigns that
encountered significant hurdles to implementation were the Naropa University and Sustainable
Jersey City campaigns. In both campaigns, there were competing priorities for the organizations
carrying out the campaigns and hence for the Challenge participants.
By contrast, campaigns that expressed learning objectives achieved success by familiarizing
themselves with CBSM principles and techniques which, in turn, resulted in a higher quality
April 2016
33
-------
engagement. The majority of these organizations plan to build on this early success and
continue to scale up their programs.
As noted above, it is too early to assess whether FTGTW campaigns result in sustained
behavior changes. However, there are a few indicators of the potential to effect lasting behavior
change. One is the high percentage of post-Challenge participants reporting that they are likely
to continue to use the tools and strategies (see Figure 2 above). The other is the campaign's
confidence in their success and their plans to scale-up their programs. Ultimately, the question
of how to sustain waste reduction behaviors comes back to creating new social norms which
suggests sustained campaigns and an increase in scale.
3.2.4 Household Factors Affecting Reach and Effectiveness
Given the small sample size of the campaigns and self-selection by Challenge participants, it is
difficult to draw conclusions from the quantitative data with respect to factors influencing
household participation and the reported effectiveness of strategies and tools.
The two major target audiences for the early pilots were households with young children and
young adults. However, in practice, a majority of the 2013 and 2014 campaigns did not target
their outreach or messaging. Of those who did, more targeted families with children. One
challenge cited by a number of campaigns in targeting young adults was identifying networks or
venues through which to reach them. Farmers markets appear to offer access to more diverse
demographics than do presentations at civic or neighborhood groups.
Some hints as to differences in the two target populations' responses to the campaign materials
emerge from comparing the Honolulu and King County 2012 results. Meal planning was judged
to be more of a task by young adults than families with children, although (as noted above)
these results should take into consideration that the Honolulu campaign which targeted young
adults used a more complex meal planning tool requiring more effort.
In Seattle, the baseline audit participants, who were self-selected through responding to a
request for participants placed in the newsletter that accompanies utility bills, skewed heavily
towards the well-educated and those with higher level incomes. While this bias does not predict
who might respond to a FTGTW campaign, it does indicate that outreach needs to be
targeted with the desired demographics in mind.
In the San Benito pilot, most of the participants were Spanish-speaking. Language appeared to
be a barrier to participation as reported by the implementing partner, even with the FTGTW
tools being available in Spanish.
Rhode Island reported a very positive experience in engaging low-income households but also
noted the need for sensitivity around whether some of these households would have food at all.
Many of the participants wanted to know what more they could do around the issue of wasted
food and also offered to be ambassadors to their communities. The incentives they offered for
participation included community hours credits which housing authority residents are required to
fill in exchange for subsidized housing.
In Thurston County, a late 2014 survey of people who chose not to participate in the Challenge
identified two main reasons for their choice. One was the previously mentioned time barrier but
also many of the survey respondents stated that they didn't need to see how much they wasted
since they were already aware they wasted a lot and wanted to move forward with testing the
strategies without taking time for baseline measurements.
April 2016 34
-------
As campaigns scale-up, they are more likely to target the general population. Whether one
should focus on identifying innovators and early adopters over representative samples of the
whole population depends on a campaign's objectives. This also holds true for whether one
should try targeting different demographic segments over outreach to the general community.
3.2.5
Effectiveness of Implementation Support and Learning Community
FTGTW was developed as a partnership that includes the EPA, city, county and state agencies,
and non-profits, each contributing unique perspectives, skills and experiences. This
collaboration provides a valuable opportunity to both leverage resources and accelerate
learning.
Food Recovery Hierarchy
Source Reduction
Reduce the volume of surplus food generated
Feed Hungry People
Donate extra food to food banks, soup Kitchens and shelters
One of the most successful elements of the
partnership are the monthly peer learning
group calls convened by the EPA. Community
partners share their experiences and offer
peer advice in tackling various implementation
issues, while the EPA uses the calls to seek
feedback on FTGTW campaign development
and to efficiently support the network.
In effect, the calls are facilitating the
emergence of a network of waste prevention
practitioners focused on source reduction (see
hierarchy at right). Through these calls,
participants gained valuable experience in
CBSM as a means of community engagement.
This collaboration allows communities to
leverage their limited resources by sharing
the costs of campaign development and
implementation. This network model of information sharing, replicating and implementing has
the potential to generate novel solutions and whole system innovations that no one could
achieve on their own.
Additionally, the importance of participating in the peer group learning calls to learning CBSM as
the foundation of successful campaigns cannot be overstressed. CBSM provides a way to
creatively engage communities in wide-scale behavior changes but it takes both understanding
and experience to successfully execute. The peer learning group calls provide a way for the
implementing organizations to share lessons learned and accelerate their own learning.
April 2016
35
-------
3.3 CAMPAIGN IMPACT
A second major FTGTW program objective is to determine if a shift in household food waste
behaviors has the potential to result in waste tonnage reduction. To achieve quantifiable
reductions in wasted food at the community level, it is necessary to engage and sustain
behaviors that have a significant impact at the household level (depth) as well as engage a
significant percentage of the general population in adopting the behaviors (breadth).
Given the small sample size of the majority of the campaigns, the focus was on measuring the
amount of food going to waste in individual households both before and after adopting the
strategies to reduce wasted food, that is, on measuring impact at the household level or the
"depth" potential. The period of time households measured waste before adopting waste
reduction strategies is referred to as the "baseline period" and the amount of food wasted during
that time as the "baseline amount".
Figure 6, the Food Loss and Waste Data Map, locates the impact results from the FTGTW
campaigns relative to other food loss and waste (FLW) data. The map shows key FLW data in
relation to the material flows the data represent and the data types and sources. It is intended
for use as a tool to assist understanding of the more commonly used data points referencing the
extent of food waste.
Interpreting impact results requires understanding which fraction of waste flows is being
measured which, in turn, depends on clear definitions of what is being measured. Food loss and
waste (FLW) is defined as food and associated inedible parts removed from the food supply
chain, where food is any substance intended for human consumption.33 The USDA defines food
loss as the edible amount of food, postharvest, that is available for human consumption but is
not consumed for any reason, while food waste is defined as a component of food loss and
occurs when an edible item goes unconsumed.34 Food waste as referred to in waste
characterization studies typically includes both the edible and non-edible portions of food.
It is worth noting with regard to calculating impact that there is no data source that provides the
amount of household food purchases by weight. The USDA Loss-Adjusted Food Availability
(LAFA) Data Series estimates for consumer level loss aggregate consumer purchases (by
weight) of food for preparation or consumption at home with purchases made at restaurants and
food service (as seen in Figure 6).35
Figure 6 indicates that the fraction of waste measured in the FTGTW challenges is a portion of
the waste that goes to solid waste destinations and does not include wasted food that was fed
to animals, was backyard composted or went down the drain.
Depending on what data was collected by a campaign, the percentage reductions reflect either
the households' total solid food waste (hereafter referred to as "total waste") or preventable
waste fractions.
33 World Resources Institute. 2015. Draft FLW Protocol Accounting and Reporting Standard.
34 Buzby, Jean C., Hodan Farah Wells, and Jeffry Hyman. 2014. The Estimated Amount, Value, and Calories of Postharvest Food
Losses at the Retail and Consumer Levels in the United States.
35 Available at http://www.ers.usda.goV/data-products/food-availability-per-capita-data-system/.aspx.
April 2016 ~36~
-------
Figure 6: Food Loss and Waste Data Map
Waste Stream
Data Methods and Sources
ERS: 10% of the amount of food
available for human consumption not
including non-edible portions (2010).
ERS: 21% of the
amount of food
available for human
consumption not
including non-edible
portions (2010).
ERS:5.6lbs/person
/week of the amount of
food available for
human consumption
not including non-edible
portions (2010).
Mass Balance
USDA ERS Loss-Adjusted
Food Availability Data Series
Estimates food loss not Including non-edlbie portions of food waste.
Food Loss and Waste Data Map:
Data References for Prevention Potentials
and Waste Disposition Flows
This data map was developed for the food loo Good to Waste (FTGTW) Project.
See reverse for definitions and additional information on estimates.
Waste Destinations
in order of EPA's Food
Recovery Hierarchy
Alternative Waste
Destinations: Animal
feed; industrial products;
fermentation; on-site
composting; land
application; down the
drain
Solid Waste Destinations:
Anaerobic digestion;
commercial composting;
combustion; and landfill.
Animal Feed
>| Backyard Composting |
>j Anaerobic Digestion [
W Commercial Composting
©
FTGTW: 50% reduction in
preventable waste or a 20%
reduction in residential total
solid food waste
Combustion
Landfill
Down the Drain
Waste Characterization
EPA MSW Characterization Reports
Estimates edible and non-edible portions of food waste.
Direct Measurement
FTGTW Home Audits
Estimates edible and non-edible portions of food waste.
April 2016
37
-------
Food Waste Data Map: Definitions and Notes
The Food Waste Data Map (the schematic on the reverse side of this sheet) shows key data in relation to the material waste flows
the data represent and the data types and sources. The map was produced as part of the EPA-supported Food: Too Good to Waste
(FTGTW) Project. It is intended for use as a tool to assist understanding of the more commonly used data points referencing the
extent of food waste,
Primary Definitions
In this schematic, food loss and waste (FLW) is defined as food and associated inedible parts removed from the food supply chain,
where food is any substance intended for human consumption (see http://www.wri.org/our-work/project/food-loss-waste-proto-
col).
The USDA defines food loss as the edible amount of food, postharvest, that is intended for human consumption but is not
consumed for any reason, such as spoilage, moisture loss, and food waste, (http:/Awww.ers.usda.gov/publications/eib-econom-
ic-information-bulletin/eib121.aspx). Food waste as referred to in waste characterization studies typically includes both the edible
and non-edible portions of food.
Recovery is the collection of wholesome food from wholesale and retail sources for distribution to people in need. Food dona-
tions are the contribution of wholesome food by households to programs distributing food to people in need.
Data Type Definitions
Mass Balance: Estimates based on data for raw material input and produced amounts.
Waste Characterization: Data from waste stream samples where the different fractions of waste are weighed.
Direct Measurement (e.g. home audits): Volume or weight measurements conducted at unit of analysis (e.g., household,
business).
Data Source Acronyms
USDA ERS: US Department of Agriculture, Economic Research Service
US EPA: US Environmental Protection Agency
FTGTW: Food Too Good to Waste
Mass Balance Fraction Definitions
All ERS estimates are for food loss. ERS Loss-Adjusted Food Availability (LAFA) data is considered to be preliminary because data
initiatives are underway to improve the underlying loss assumptions. Additional information on the ERS'mass balance loss
estimates can be found at: http^/www.ers.usda.gov/data-products/food-availabi/ity-fper-capita)-data-system//
toss-adjusted-food-availobitity-doaimentation.aspx
Primary Level Loss: Losses occurring between point of production (e.g. farm or processing or manufacturing facility) and
retail such as further processing, trimming, shrinkage, or loss in the marketing and distribution system. Primary level waste
is limited to post-harvest waste, that is, it does not account for on-farm losses. ERS cannot estimate primary level loss
because of data limitations.
Retail Level Loss: Losses at the retail level, such as in supermarkets, supercenters, convenience stores, mom-and-pop
grocery stores, and other retail outlets (but not including restaurants and other foodservice outlets). For fresh fruits, vege-
tables, meat, poultry, and seafood, loss estimates are based on supplier shipment data with point-of-sale data from stores
in large, national supermarket retail chains.
Consumer Level Loss: Losses at the consumer level including losses for food consumed at home and away from home (for
example, restaurants and fast food outlets) by consumers and food services. For most commodities, loss estimates are
based on Nielsen Homescan data (food purchase data) and the dietary intake component of the National Health and
Nutrition Examination Survey (NHANES) (food consumption data).
Available Food Supply: The amount of food available for consumption not including non-edible portions. The estimates
of food loss as a percentage of the available food supply cited on the data map have had the non-edible portions removed
(Buzby et al. February 2014.The Estimated Amount, Value and Calories of Post-Harvest Food Losses at the Retail and Con-
sumer Levels in the United States. USDA ERS Economic Information Bulletin No.121.)
Direct Measurement Definitions
Inedible Parts: Components of a food that the user does not intend to consume. Examples could include bones, rinds, or
pits/stones.
Edibles/Preventable Waste: Food that is intended for human consumption but is not consumed for any reason (e.g. loss
from mold, plate waste).
Baseline Amount: Amount of food wasted before FTGTW intervention.
End Amount: Amount of food wasted after FTGTW intervention.
-------
Table 5 (on the following page) summarizes the results on how much food was wasted during
the baseline period (prior to the households adopting strategies) and during the final week and
computes the percentage reduction. The per capita averages for both total and preventable
waste fractions by weight and volume are shown in Figures 7 and 8 respectively.
The methodology for calculating the percentage reduction was to average the amounts of waste
for the baseline weeks and to compare that to the average of the final week. The reason for
selecting the final week average as the basis for comparison is because the change in behavior
has a cumulative effect as new habits become established. Thus, the predominant pattern is a
downward trend in the amount of food wasted over the period of the Challenge, (as noted above
in Section 3.2.1 in the discussion of the Seattle baseline waste audits). As the samples are
small, the most that can be determined by this method is the potential for wasted food reduction
at the household level. The final week average represents a more accurate estimate of this
potential than does an average of the strategy implementation weeks. In fact, in more than one
campaign a jump was seen in the amount of waste collected during the first week of
implementation.
The following sections of the report discuss first the findings on the baseline amounts of wasted
food, than the percentage reductions, followed by a discussion of the household factors
affecting impact.
3.3.1 Baseline Amount
Table 6 summarizes the findings on how much food was wasted in households in the Iowa City,
Honolulu and Seattle pilots during the baseline period (prior to intervention) and compares it to
figures for tipping weights in King County (where Seattle is located) in 2007, the USDA estimate
for consumer food loss, and the EPA estimate of residential food waste sent to landfills. (The
USDA and EPA figures cover the entire U.S.)
Table 6: Total Baseline Food Waste per Person per Week (Iowa City, Seattle, Honolulu)
Community
Iowa City
Honolulu
Seattle
King County 2007 - Tipping weights
d)
EPA Landfilled Food Waste 201 1 <2>
USDA Consumer Food Loss 2010
(Includes restaurant waste) (3)
Pounds/
household/
week
4.7 Ibs
4.3 Ibs
6.3 Ibs
10. 5 Ibs
~
Number of
people/
household
2.5
1.7
2.3
2.5
~
Pounds/person/
week
2.2 Ibs
3.5 Ibs
2.8 Ibs
4.2 Ibs
2.5 Ibs
5.6 Ibs
Sources: (1) Cascadia Consulting Group. 2007 Waste Characterization Study (2) Municipal Solid Waste
in the United States: 2011 Facts and Figures. May 2013, US EPA. U.S. EPA. (3) Buzby et al, 2014. The
Estimated Amount, Value, and Calories of Postharvest Food Losses at the Retail and Consumer Levels
in the United States.
From Table 6 we see that the total amount of food wasted per person per week in the three
campaigns ranged from 2.2 pounds in Iowa City to 3.5 pounds in Honolulu. This is comparable
to the EPA estimate of 2.5 pounds of landfilled residential waste per person per week.
The comparability suggests that the household measurements reasonably represent the actual
amount of wasted food.
April 2016
39
-------
Table 5: Household and Per Capita Average Percentage Reductions by Project
Project
Sample Avg
Fraction
Method
Number
of People
per
Household
Number Number Avg Avg % Change Avg per Avg per
of of Household Household in Capita Capita
Weeks Baseline Baseline Final Household Baseline Final
Weeks Amount Week Averages Amount Week
Amount Amount
Per Household Per Capita
% Change
in Per
Capita
Averages
Gresham
Honolulu
Rhode Island
Seattle
Honolulu
Iowa City
Seattle
Rhode Island
Addison County
Aurora
King County 2014
Thurston County
King County 2012
Oak Park
17
12
15
107
12
29
107
15
31
23
53
42
13
12
2.3
1.7
2.2
2.3
1.7
2.5
2.3
2
n/a
2.1
2.9
2.9
4.5
3.5
Preventable
Preventable
Preventable
Preventable
Total
Total
Total
Preventable
Preventable
Preventable
Preventable
Preventable
Total
Total
Weight
Weight
Weight
Weight
Weight
Weight
Weight
Volume
Volume
Volume
Volume
Volume
Volume
Volume
6
4
6
13
4
6
13
6
6
6
4
4
5
6
2
2
2
13
2
1
13
2
2
2
1
1
1
2
2.4
1.6
2.4
2.1
4.3
4.7
6.3
69
112
146
141
n/a
190
152
1.9
1.2
0.8
n/a
3.8
5.4
n/a
49
101
81
88
n/a
137
167
-19%
-22%
-66%
n/a
-12%
15%
n/a
-30%
-9%
-45%
-38%
-31%
-28%
10%
1
1.0
1.1
0.9
3.5
2.2
2.8
52
n/a
65
49
n/a
41
45
0.9
0.9
0.6
n/a
2.7
2.4
n/a
36
n/a
38.6
31
n/a
30
48
-15%
-11%
-48%
n/a
-31%
7%
n/a
-30%
n/a
-41%
-37%
-30%
-27%
7%
-------
Seattle's numbers are also similar to those found in previous studies conducted by Seattle
Public Utilities as seen in Table 7 below. This indicates that the Seattle averages are a fair
estimate for how much food goes to waste in Seattle. However, the Seattle Campaign was
conducted from Jan 6 through April 6 of 2013, so it may be that the seasonal average for winter
months is lower than the average for growing season months when more fresh produce might
be purchased.
Table 7: Baseline Data Comparison with Previous Waste Studies
Study
Food: Too Good to Waste
2013
Average Weight per Household
329 Ibs./household/year
Green Cone backyard compost
study
1993
370 Ibs./household/year
Curbside weighing project
1994
328 Ibs./household/year
Adjusted waste composition
study
1994
293 Ibs./household/year
Source: Seattle Public Utilities, Draft Report, Seattle's Food Waste Weighing Pilot
Time of year was probably not a factor in comparing the campaign results for Seattle and
Honolulu with due consideration for the two different campaign locations. The Honolulu
Challenge ran from February 20 to March 20 of 2013, so the two pilots were conducted at the
same time of year. However, the Iowa City campaign was conducted during the early summer
months. We would expect that there would be more total food waste during the summer months
because of an increase of the availability of fresh produce. However, the Iowa City results
indicate a lower baseline amount.
It could be that household size is the explanatory factor in the range of baseline amounts,
especially the higher baseline amount in Honolulu. Previous research says that the larger the
household size, the smaller amount of waste per capita, the explanation being that food tends to
get eaten up at a faster rate.36 The second factor that might explain the differences are existing
regional differences in food waste management behaviors. Prior to the campaign Iowa City
residents were already taking steps to reduce wasted food at home which is reflected in their
lower baseline amounts. There is some evidence to support this supposition which is discussed
in the following section.
In both the Seattle and Honolulu pilots, baseline preventable food waste averaged about
one-third of all food waste. (Preventable waste was not measured in the Iowa City campaign.)
Over the course of the Seattle pilot, weekly averages varied from 26.9% to 41.8%. While the
average percentage of preventable food waste in the Honolulu campaign fell from the baseline
period to the period participants tried strategies, the average was over 40% in the baseline
period.
The results on the amount of preventable food waste as a percentage of total food waste merit
further research. In the U.K. WRAP studies, preventable food waste accounted for 60% of
household food waste compared to the one-third of overall waste in the Seattle and Honolulu
campaigns.37
36 Quested et al. 2013. Household Food and Drink Waste in the United Kingdom 2012. Also, the Seattle data bears out this
hypothesis as discussed in Section 3.3.3 below.
37 Quested et al. 2013. Household Food and Drink Waste in the United Kingdom 2012.
April 2016
41
-------
3.3.2 Percentage Reduction in Wasted Food
Table 5 and Figures 7 and 8 summarize the FTGTW impact findings. The reduction in
preventable food waste ranged from -11% to -48% by weight and from -27% to -39% by volume,
while the total waste change ranged from +7% to -30% by weight and from +7% to -27% by
volume.
Several conclusions can be drawn from these results. First, there is a high degree of variability
in the wasted food reduction potential. It is apparent that the percentage of waste reduction is
affected by campaign variables such as the time of year when the campaign was conducted and
which fraction of waste was measured.
Both Iowa City and Oak Park saw increases in the total amount of wasted food rather than
reductions. That both campaigns measured total waste rather than preventable waste in tandem
with their timing in late spring/early summer probably account for the increases. Participants
noted the individual household amounts in later weeks were due to watermelon rinds and fresh
corn husks. It seems likely that there is enough of a seasonal variation in the amount of non-
edible waste to obscure probable reductions in the amount of preventable waste. This is
bolstered by the fact that preventable waste reductions are more significant than the total waste
reductions as would be consistent with this premise.
Another campaign variable that appears to significantly affect the recorded amount of reduction
was the length of the Challenge. Challenges that lasted longer saw more significant reductions.
The six week Rhode Island campaign resulted in a 48% reduction while the four week Honolulu
campaign resulted in an 11% reduction.
A third campaign variable that affected results is the choice of measurement method. As
discussed earlier, volume measurements are less accurate than weight measurements. While
volume measurement can be an effective and relatively inexpensive means of raising
awareness in individual households, it is less useful than weight measurements in reliably
establishing the potential for waste reduction. For example, the difference in the Rhode Island
results by weight and volume were substantial. The Rhode Island campaign saw an average
48% reduction by weight and only a 30% reduction by volume. The diminished accuracy of
volume measurement may also help explain why the volume of edible waste in King County in
2014 was as great as total waste in 2012 (see Figure 7).
Region may also be a factor in determining the waste reduction potential. One household factor
that greatly affects the potential for reduction is the amount of baseline waste. It is clear that
households starting with a low level of waste have less waste reduction potential (see the
next section for a discussion of the Rhode Island Results). The Iowa City results are an
interesting case in this respect. While Iowa City measured the total waste fraction and thus we
do not know what percentage reduction was achieved in preventable waste, it is also seen from
Figure 7 that Iowa City started with a lower baseline amount of waste than did the other
campaigns. One contributing factor in this lower starting point may be that there is a higher level
of awareness in Iowa City as discussed above in Section 3.2.2.
Possibly the most significant conclusion regarding the impact results is that campaigns that are
successfully implemented can result in a significant reduction in preventable food waste
at the household level. The magnitude of the potential reduction in preventable waste is
50% or more or approximately a half pound per person per week. This is roughly equivalent
to a 20% reduction in total waste.
April 2016 42
-------
The 50% or more potential reduction estimate is based on the Rhode Island campaign results
where the average reduction in preventable waste was 48% by weight and the 41% and 37%
reduction by volume in the Aurora and King County 2014 campaigns respectively. One of the
four Rhode Island cohorts achieved a 70% average reduction (see the next section).
Finally, it is worth repeating that given the small sample sizes and the lack of quantitative data
on reach, these impact results are not scalable to the community level. A next step in
determining impact will be to conduct a community-scale measurement study (see Section 4.3
for further discussion).
3.3.3 Household Factors Affecting Impact
There is a significant amount of variability in the amount of wasted food by household and
therefore, the amount of waste reduction possible. These differences largely reflect household
food management practices but, in turn, these practices are influenced by demographic and
lifestyle factors as seen in Figure 9.
Figure 9: Waste Fraction Factors
How much food
How much consumed
Mow much disposed
households to other source* faft
( Inedible!
purchase animal feed)
How much wasted
(Tolal Solid Pood Wsate Pratlloi
Preventable waste
IPiwnuWc WBMC Irxtxr.)
Factors influencing Factors influencing wa ste:
purchases: * Time of Year
Region Vegstarlanor Not
* Income * How Often Household
Household Sl» Eats Out
* Age * How Often Household
- Number of Earners Shops
The behavior-influencing factors that affect both how much food is purchased and how much of
that is wasted are also shown in Figure 9. Data was collected on the variables in red font in at
least one of the pilots. Factors affecting how much food is purchased may also influence how
much is wasted. These factors' influence on impact are discussed below in Section 3.3.3.
The high degree of variability introduces a large amount of noise into the data with small
samples, particularly since the Challenge participants are self-selected. For that reason, caution
should be used in extrapolating the data. At the same time, the data provide useful indicators of
which demographics and management practices to target for the greatest impact.
A regression analysis of the Seattle baseline data where 123 households measured their waste
for 13 weeks found that:
Income and education significantly affect amount of food waste. Both low and high
income wasted more, while middle income households (inflection point of $67k)
wasted the least.
April 2016 43
-------
Young adults generate the most food waste.
Immigrants waste less than average.
Vegetarians waste more than average.
The amount wasted per person decreases with household size.
The finding that young adults have greater baseline amounts of wasted food is consistent with
earlier studies. As explained in the FTGTW background research report, a dynamic lifestyle in
which plans on where and when to eat frequently change makes food management
challenging.38 With vegetarians, the higher waste amounts are attributable to greater amounts of
inedible waste.
While the Seattle survey also contained questions regarding food management practices, the
findings around these were much weaker. It is surmised that self-reporting food management
practices will be biased and any correlations with amounts of wasted food weak.
Finally, the Seattle analysis confirmed the significance of the feedback effect from measuring
waste: food waste decreased over the length of 13 weeks although no strategies or tools were
introduced to the study participants.
Also contributing to high variance, a number of out-of-the-ordinary but not infrequent household
events can alter food management practices, such as household visitors, vacations, and
sickness. For example, in Honolulu, "irregular" events reported by participants included a family
member placing a large quantity of food in refrigerator unbeknownst those responsible for food
preparation, where after it went to waste; one household moving to a new residence; and
another spending a week at a residence other than their own.
Factors such as diet and households schedules can cause variability as well. These "lifestyle"
factors might also influence waste averages for different demographics. For example, in the San
Benito pilot, which targeted low income-families whose primary language was Spanish, higher
amounts of legumes and grains were wasted and lettuce was the main vegetable wasted.
Photo Examples of Wasted Food - San Benito Pilot
The average amount wasted per household per week for all weeks of the San Benito pilot, both
baseline and post-intervention, was 2.5 pounds. A high percentage of the wasted food was food
that was prepared but not eaten (see photographs above). These patterns may reflect food
management practices that are typical of this demographic.
Further evidence of the effect of household factors on the amount of wasted food comes from
the analysis of the Honolulu campaign results performed by A. Lavers.39 Of the eight
38 EPA 2012 ibid.
39 Alexandra Lavers. 2013. Eat Me First! Development and Evaluation of the Food: Too Good to Waste Household Food Waste
Prevention Program in Honolulu, Hawaii, USA.
April 2016
44
-------
households that saw waste reductions, two distinct clusters were identified: Cluster 1 consisted
of younger participants and predominantly single-member households. Cluster 2 included all
campaign study participants who were over the age of 45. Cluster 1 averaged a 41% reduction,
while Cluster 2 averaged a 24% reduction. The report states, "[l]t is apparent that the difference
in preventable food waste reduction can be attributed to the fact that Cluster 2 households
started with less food waste and thus had less potential for reduction; Cluster 1 simply started
with a larger amount of food waste." An additional five households increased the amount of food
wasted over the Challenge. Three of these households cited "irregular" events that affected the
amount wasted.
The effect of eating out on the amount of wasted food was also analyzed in the Honolulu report.
No discernible effect was seen and the report concluded, "[S]ome households are aware that
they eat out frequently and adjust grocery habits accordingly while others do not."
Figure 10 depicts the waste reduction by week for four cohorts in Rhode Island. The baseline
amount for the high-income households is more than twice that of the low-income households.
Correspondingly, the reduction for the housing authority cohort was 48%, less than the 55%
reduction in the upscale apartment cohort.
3.5
Figure 10: Rhode Island Food Policy Council (RIFPC) Preventable Food
Waste Household Averages by Test Group
Friendlies - 65% reduction
Listserve - 70% Reduction
Upscale Apartments - 55%
Reduction
Housing Authority - 48% Reduction
Week
The Iowa City results also indicate that the amount of food wasted rose with income level
although there was a huge variance in the amounts. Other influences included diet, vegetarians
had more waste than non-vegetarians; household gardening, gardeners had more waste than
non-gardeners; and age, seniors had less waste. Interestingly, peak waste was generated by
the 41-50 year old participants in the Iowa City Challenge.
In sum, almost all of the factors cited in Figure 9 has some effect in the amount of food going to
waste though different factors had greater influence in some campaigns than others.
April 2016
45
-------
3.4 CAMPAIGN COSTS, ENVIRONMENTAL BENEFITS AND FIT WITH EXISTING
PROGRAMS
At the request of the Forum partners, a key FTGTW evaluation goal was to determine the costs
to implement a FTGTW project and to estimate its environmental benefits. This information is
needed to help make the case for the return on larger scale campaigns. One more final
objective was to review the alignment of FTGTW with existing policy and program objectives.
Implementation costs for campaigns ranged from a few thousand dollars for pilots to
above $100,000, not including staff time, for broad scale campaigns.
Given the small scale of the pilots and the shared responsibility for design and development
between the Forum and the implementing partners, campaign costs were modest. Basic costs
included costs to print the CBSM tools, cost of measurement bags or other containers used in
the Challenge, and staff time. In addition, some pilots used incentives to engage individual
households, the costs of which include both time to procure the items and/or the cost of the
incentives themselves. In several campaigns, partnerships were made with businesses who
offered the incentives in-kind.
For the 2013-2014 campaigns, the cost to use FTGTW materials was negligible to participating
communities, while the opportunity to collaborate on tool and campaign design led to reduced
costs for these activities.
Staff time included time spent in the following: gaining organizational approval; campaign
design; materials preparation and modifying the tools for local use; outreach and collecting data;
and time spent evaluating and disseminating campaign results. The allocation of staff time
varied according to campaign objectives and scale; whether outside resources were used; and
the implementing organization's level of CBSM experience. In Iowa City, staff time accounted
for 82% of an $11,000 budget.
The cost data for the three campaigns King County conducted each fall in 2012, 2013 and 2014
(see Campaigns 1, 8 and 15 in Table 2) indicate a progression in the cost structure as shown in
Figure 11. (Volunteer hours are not included).
Figure 11: Staff Hours, King County 2012, 2013 and 2014
250
200
150
100
50
00
(/I
m
u
'o1
Q_
0
ro
ro
01
Q_
ro
0)
-M
ro
c
o
ro
4-»
01
E
01
Q.
E
2012 B201
_^
^
w
"ro
c
<
ro
Q
3 2014
00
c
c
S
01
BO
W
01
z
o
ro
Q.
u
t
ro
Q.
_^;
i_
0
1
01
This progression suggests that, as experience grows, less time as a percentage of total staff
hours will be spent in design and material preparation, and proportionally more in
implementation but also that staff time will increase as campaigns increase in size. Specifically:
April 2016 ~~46
-------
Total staff hours grew from 190 hours in 2012 to 315 hours in 2013 and 745 hours in
2014. The 2014 community-scale direct outreach campaign involved more than double
the staff hours of the 2013 broad-scale media campaign.
In 2012, the pilot year, staff hours spent for design and materials preparation accounted
for 44% and 17% of 188 hours total.
In 2013 and 2014, the percentage of staff time shrunk to 38% and 32% respectively as
the total number of hours increased to 315 hours in 2013 and again to 745 in 2014.
Material costs will also rise as activities are scaled-up. However, the type of campaign
determines how much. Thus, King County spent $42 thousand in 2013, which included the
costs of video production in its broad-scale media campaign but only $20 thousand for its 2014
community-scale direct outreach campaign which focused on tabling at farmers markets,
including $3,700 for incentives. Consultant costs for both years were $59 thousand.
Within the scope of our present work, it is not possible to estimate the environmental
benefits of a FTGTW campaign with any degree of accuracy given the current data sources
and life cycle assessment assumptions. However, the macro data point to considerable
benefits.
To see a return at the community level in terms of environmental benefits and reduced waste
handling costs, larger scale initiatives that see significant diffusion of the program in the
community are indicated. The to-date findings only provide evidence of the potential for
reductions at the individual household level.
For comparison, however, WRAP documented the benefits of a targeted campaign in West
London. The campaign resulted in a 15% reduction in household food waste, on average, and a
35% reduction in households who took action as a result of the campaign (14% of the
sample).40 Annually, this reduction could save West London households an estimated £14
million and prevent 20,000 tons of CO2e. In addition, for every £1 invested in the campaign,
West London boroughs saved up to £8 in avoided disposal costs. The research was funded by
the London Waste and Recycling Board and performed by WRAP in 2012-2013.
With respect to fit of the campaign with existing programs, most campaigns had waste
management goals as their primary objective. On the policy front, household food waste
reduction promises to be significant both in terms of climate action and solid waste. State level
mandates and legislation on climate change and solid waste are driving interest in the need to
reduce waste, not just divert it. For example, both Vermont and Massachusetts have enacted
bans on food waste going to landfills. Because most of Vermont's population is located in small
towns and rural areas, curbside collection of composting materials is not an option. Therefore,
publicly supported food waste prevention initiatives are a strategic answer to meet the state's
ban cost effectively.
The Rhode Island campaign was implemented by the Rhode Island Food Policy Council. In
addition to having a zero waste mission, they were also interested in how reducing wasted food
contributed to food security. It was possible for them to obtain relevant data for this purpose by
segmenting their participants into cohort groups. In other words, they aligned their
implementation methods with their objectives.
40 Quested, Tom and Robert Ingle. 2013. West London Food Waste Prevention Campaign Evaluation Report.
April 2016 ~~47
-------
King County has conducted a consumption-based emissions inventory where food was shown
to be a large contributor to CO2 emissions and thus an important target for emission reductions.
On the solid waste side, King County has done a lot of things like composting, so they wanted to
take their program to the next level and food waste reduction was a next logical step.
The question of how to integrate prevention of wasted food with existing composting initiatives is
still open with respect to the evidence. However, the FTGTW campaigns that were able to
successfully integrate promoting composting with waste prevention framed their objectives
using the EPA's Food Recovery Hierarchy in which wasted food prevention and composting are
seen as complementary strategies but source reduction is the preferred option. WRAP also had
done research supporting this conclusion.41
The challenge is to develop clear messaging that distinguishes between handling preventable
and inedible waste and the best solutions with respect to their environmental impact. WRAP'S
research has shown that composting options may provide a rationale for some households to
waste, that is, legitimize waste generation.
3.5 FINDINGS SUMMARY
A summary evaluation of the FTGTW campaign results relative to the research goals discussed
in Section 1.2 of this report is presented next.
Behavior Change Strategy and Tool Effectiveness:
Households gave high marks to the FTGTW strategies and tools, finding them useful
and easy to use.
Consistent with CBSM research, strategies associated with a tool are considered more
useful than those without.
Campaigns featuring a range of strategies allow each household to focus on the strategy
or strategies that work best for them. Even within a target demographic there is
considerable variability in household food management practices.
The FTGTW Challenge served a dual purpose as both a measurement and a behavior
change tool. 93% of participants said that they were now more aware of food going to
waste in their households with 55% strongly agreeing that this was the case. Tracking
the amount of wasted food motivates action to reduce wasted food. In effect, feedback
increases awareness by countering habitual behavior and activating waste aversion.
Volumetric measurement is more geared to providing households a convenient waste
yardstick but weight measurements provide for more accurate accounts of food waste.
Measuring the preventable portion of wasted food focuses household attention on the
potential impact from adopting the waste prevention strategies.
Less than six weeks (including a two-week baseline measurement period) is too short a
time to provide an accurate estimate of the waste reduction potential.
Self-audits appears to be an effective means of determining how much food is going to
waste for small to medium-sized sample populations.
41 Marsh, Emma. 2013. Food Waste Messages for Maximum Impact: How to Engage Your Residents in Preventions and
Collections.
April 2016 48
-------
Messaging Effectiveness:
Campaign volunteers and staff reported lively discussions and expression of interest at
tabling events and lively and spirited conversations at workshops, reflecting the issue's
relevance to households and the effectiveness of offering tools and strategies to reduce
wasted food along with the messaging.
Feeling bad about throwing away food and wasting money appear to be equally strong
motivators to reduce wasted food.
Increasing awareness of the indirect environmental effects of wasted food through
messaging is challenging with mixed results. There is the need to particularize
environmental messaging to the household level for greatest effect.
Outreach and Engagement Effectiveness:
Outreach and engagement tools that are designed to leverage social networks and
create social norms are among the most effective.
Community-scale direct outreach using one-on-one and peer-to-peer recruitment
techniques was more effective than recruitment through indirect means such as social
media outreach, a finding consistent with previous research on CBSM methods.
Response to direct outreach efforts was enthusiastic.
The general rule for successful campaigns was to engage participants early and often.
Outreach needs to be targeted with the desired demographics in mind.
Campaigns can fail without a focused effort.
Effectiveness of Implementation Support and Learning Community:
The FTGTW peer network allows communities to leverage their limited resources by
sharing the costs of campaign development and implementation.
Peer learning group calls provide a way for implementing organizations to share lessons
learned and accelerate their own learning.
Impact:
The total average amount of food wasted per person per week of 2.2 to 3.5 pounds is
comparable to the EPA estimate of 2.5 pounds of landfilled residential waste per person
per week.
The baseline preventable fraction averaged about one-third of all food waste.
Households starting with a low level of waste have less waste reduction potential.
Campaigns that are successfully implemented can result in a significant reduction in
preventable food waste at the household level. Reductions in preventable waste of 50%
and greater are possible. This equates to an average reduction in preventable food
waste of around a half pound per week per capita or 20% of total food waste.
Implementation Costs:
Campaign implementation costs ranged from a few thousand dollars for pilots to above
$100,000, not including staff time, for broad scale campaigns.
Community-scale direct outreach campaigns involve more staff time than do broad-scale
media campaigns but material costs will be less.
Environmental Benefits:
April 2016 49
-------
There is insufficient data to determine the environmental benefits of FTGTW projects at
the community level. However, the macro data point to considerable benefits.
Fit with Existing Programs:
Most campaigns had waste management goals as their primary objective. On the policy
front, household food waste reduction promises to be significant both in terms of climate
action and solid waste.
Campaigns with a focus on food security, whether at the household (Rhode Island) or
community level (Hawaii), were also successful.
April 2016 50
-------
4.0 RECOMMENDATIONS
The FTGTW strategies and tools have proven effective in helping households to reduce wasted
food. In addition, the greater majority of the campaigns were successful in achieving their stated
objectives. This section of the report discusses recommendations to newly implementing
organizations based on the previous campaigns' experiences. It concludes with an applied
research agenda for strengthening the evidence base.
4.1 EMERGING BEST PRACTICES
Community-based Social Marketing is a grounded approach to changing household behaviors.
Implementing organizations have the ability to adapt the campaign messages and tools to the
needs of their communities. The following campaign examples illustrate emerging best practices
that advance the basic FTGTW model described in Section 2.
Targeting Campaign Outreach, Iowa City, Iowa
Iowa City used an intensive targeted outreach approach to achieve a very high recruitment rate
of 17% compared to the average of 1 to 2% in the majority of the campaigns. In addition, they
selected, in consultation with the city's solid waste manager, defined neighborhoods to
represent the city's demographics (similar to cluster sampling), emphasizing a variety of ages
and income levels. This provided for a high degree of confidence that their results were typical
for the city's population.
Their outreach methods included:
Initial mailing to 300 households in six neighborhoods, including a pre-pilot survey,
background letter, and pre-stamped envelope to return survey.
Door hangers to encourage people to participate in pilot.
Neighborhood open houses to offer participants a convenient location to pick up pilot
supplies and to have a face-to-face introduction with participants.
Yard signs to remind participants of important dates in pilot such as when to start strategies.
Hands-on Community Engagement, Rhode Island
The Rhode Island Food Policy Council had as their objective to test the
FTGTW tools and strategies with different backgrounds and socio-
economic statuses. Their outreach workshops featured live
demonstrations of the strategies and measuring techniques (see the
picture of scales with container at right) which greatly assisted participant
learning. (Most campaigns relied on written materials and slide shows.)
The hands-on workshops likely explain why Challenge participants in
Rhode Island had the greatest percentage reduction in wasted food.
They also shared food from cooking demonstrations at the workshops
creating a sense of inclusion. The workshop presentation was translated
into Spanish for workshops where Spanish was the dominant language.
Creating New Social Norms, King County, Washington
April 2016
51
-------
King County has evolved its outreach and engagement approach
over three campaigns, one each fall from 2012 to 2014. In 2014,
they adopted a cascade training (or train-the-trainer) approach,
working with Master Composters to do tabling and outreach.
One engagement technique they developed in 2014 was taking
pictures of public commitments to reduce wasted food at farmers
market tabling events (see picture at left). They then shared these
through social media.
Making public commitments like these reinforces the intent to follow
through. As evidence, King County had the highest retention rate of
all the campaigns. It can also serve to make emerging social norms
visible. This is especially important with behaviors taking place in
the privacy of one's own home.
Sharing these photos with the person whose picture is being taken so they in turn can share
with their social networks is a powerful way of creating a new social norm around reducing
wasted food.
4.2 LESSONS LEARNED
This section offers lessons learned on how to conduct a successful FTGTW campaign based on
the above evaluation of seventeen campaigns. These lessons build on the implementation
recommendations presented in the FTGTW Implementation Guide which captured the
experience of the first five pilots in 2012.42 The list below both reiterates critical Implementation
Guide recommendations and expands on them.
Develop clear campaign objectives and a clear sense of how implementation choices
(e.g. what activities to conduct; what data to collect; and how to collect and analyze the
data) support these objectives.
Engage target audiences early and often for higher recruitment and retention rates.
Before modifying any campaign materials, particularly Campaign instructions,
implementing organizations should review the need for these changes and vet any
changes carefully. Seemingly minor changes can undermine the data collection and
analysis process.
Having households focus on the preventable waste portion of food waste may help to
simplify the measurement of food waste as well as challenge the perception that it is
only non-edibles that are being thrown out. By measuring the preventable food waste
portion only, households will also see a more dramatic reduction in the amount being
wasted, incentivizing the continued use of the strategies.
The length of time needed for households to establish new food management practices
is at least 3 to 4 weeks. Implementing partners are strongly encouraged to engage
households in trying strategies for four weeks minimum in addition to the two weeks
required for baseline data collection. Shortening these time periods does not appear to
increase participation. Also, campaigns that opted for shorter challenges reported that
this negatively impacted their results and they would not do so again.
2 West Coast Climate and Materials Management Forum, ibid.
April 2016
52
-------
Households can either measure the waste reduction throughout the period of trying
strategies or at the end of this period. However, the analysis should compare the
baseline average to the final week average.
There are trade-offs between volumetric and weight measurements. Volumetric is less
exact than weight measurements but is simpler for participants to execute and does not
require scales. Pilots should consider these trade-offs in relation to their objectives. For
campaigns wishing to collect data, weight measurements are strongly recommended.
A few comments indicated that there was some confusion about the relationship
between composting and strategies to reduce wasted food. It is recommended that a
future campaign along with an academic partner investigate messaging to clarify this
relationship.
The social nature of behavior change suggests that a focus on engaging households in
FTGTW through existing social networks that meet regularly would be very effective.
It is strongly recommended that pilots not schedule household engagement during the
holiday seasons as participants are particularly busy which makes participation difficult.
The importance of peer group learning cannot be overestimated in the success of the
pilots. Participation in the FTGTW peer learning group is strongly recommended.
4.3 MEASUREMENT RECOMMENDATIONS
Beyond the research questions that guided this evaluation is the broader issue of how to
generate large-scale social change. What policy approaches and program initiatives would
support the rapid normalization of preventing wasted food in households?
FTGTW pioneered self-audits as a cost-effective means of collecting data on household waste
patterns. The results suggest that measurement plays a critical role - not only in determining
the impact of waste prevention campaigns but also in raising households' awareness of how
much food they are wasting. As such, providing households the means to measure the amounts
of food going to waste at home becomes a powerful tool to support change in household
practices. For this reason, it is strongly recommended that future campaigns consider
incorporating a household measurement tool or strategy into their campaigns.
At the same time, to advance our understanding of how to accelerate the development of
preventing wasted food as a social norm, we need sufficient data on the community level impact
of campaigns to prevent wasted food. This includes understanding the baseline awareness of
wasted food, both as a cost to households and as an environmental issue. To measure the
impact of a campaign at a community scale, it is also necessary to know how many people
adopt waste prevention practices as well as their average waste reduction. To determine both,
FTGTW has developed a draft community-scale measurement protocol (see Appendix C).
4.4 FURTHER APPLIED RESEARCH
The Food: Too Good to Waste campaign results substantiate the hypothesis that food waste
reduction behaviors are complex with many complicating factors influencing these behaviors
and food management practices in general. At the same time, this evaluation confirms that
CBSM-style campaigns can bring about a notable reduction in preventable food waste at the
household level.
April 2016 53
-------
Ultimately, wasted food is an issue of significant concern for both environmental and socio-
economic reasons. Additional applied research is needed to quantify the impacts of wasted food
and the benefits of food waste reduction programs. A research agenda is outlined below.
Baseline Questions
Baseline research questions concern defining the scope of the problem of wasted food at the
household level and the contributing factors. These are presented next.
What are the average per capita weights of preventable (edible) and inedible fractions?
Different best practice solutions are needed to address the preventable and inedible
fractions.
What is the amount of household in-home food purchases by weight (that is, excluding
restaurant and food service purchases)? Establishing how much wasted food there is as
a percentage of in-home purchases will help determine the return to prevention
practices.
What are the baseline amounts of food waste by food type (e.g. vegetables, meat and
poultry, dairy)?
Campaign Effectiveness Questions
How do we cost effectively include direct contact in community-wide campaigns?
What is the necessary level of engagement to sustain behavioral change?
What tools are best for spurring the development of new social norms around wasted
food?
What policies are needed to support the development of successful campaigns?
Campaign Impact Questions
The intent of the intervention questions is to determine the environmental and socio-economic
impacts of waste reduction campaigns.
What is the impact of a FTGTW campaign on reducing the amount of wasted food in a
community's residential solid waste (i.e. residual waste) and/or mixed compost streams?
What are the reach and engagement rates of campaigns by input? This data is needed
to establish impact at the community level and will help determine the most effective
means.
What is the typical composition of food waste before and after intervention? This data is
needed to determine the environmental and socio-economic impacts.
What are the life cycle impacts by ton of reducing wasted food?
How does the shift to healthier eating, particularly more fruit and vegetables, affect
waste prevention behaviors and the amount of non-edibles?
April 2016 54
-------
REFERENCES
Bureau of Labor Statistics Consumer Expenditure Survey data can be accessed at
http://www.bls.gov/cex/Stables.
Buzby, Jean C., Hodan Farah Wells, and Jeffry Hyman. 2014. The Estimated Amount, Value,
and Calories of Postharvest Food Losses at the Retail and Consumer Levels in the
United States. Economic Information Bulletin. EIB-121.
Cascadia Consulting Group. 2008. 2007 Waste Characterization Study. Report.
Cox, Jayne, and Phil Downing. 2007. Food Behaviour Consumer Research: Quantitative Phase.
WRAP Report.
Economic Research Service (ERS), United States Department of Agriculture. 2012. Loss
Adjusted Food Availability (LAFA) Data Series. Data can be accessed at
http://www.ers.usda.gov/data-products/food-availability-(per-capita)-data-system/.aspx
Environmental Protection Agency (EPA). 2012. Food: Too Good to Waste Pilot: A Background
Research Report for the West Coast Climate and Materials Management Forum. EPA
910-R-12-006.
FUSIONS. 2014. Report on Review of (Food) Waste Reporting Methodology and Practice.
Available at http://www.eu-fusions.org/index.php/publications.
Glanz, Robert, 2008. Causes of Food Waste Generation in Households -An Empirical Analysis.
Applied Sciences.
Hall, Kevin D., Juen Quo, Michael Dore, and Carson C. Chow. 2009. The Progressive Increase
of Food Waste in America and Its Environmental Impact. PLoS ONE 4, no. 11.
Lavers, Alexandra. 2013. Eat Me First! Development and Evaluation of the Food: Too Good to
Waste Household Food Waste Prevention Program in Honolulu, Hawaii, USA.
Marsh, Emma. 2013. Food Waste Messages for Maximum Impact: How to Engage Your
Residents in Preventions and Collections. WRAP Report.
http://www.wrap.org.uk/sites/files/wrap/Food waste messages for maximum impact
WRAP UK.pdf
McKenzie-Mohr, Doug. 2011. Social Marketing to Protect the Environment: What Works. Sage
Publications Inc.
Muth, Mary K, Katherine M Kosa, and Shawn A Karns. 2007. Exploratory Research on
Estimation of Consumer-Level Food Loss Conversion Factors. USDA ERS Report.
Neff RA, Spiker M, Truant P. 2015. Wasted Food: U.S. Consumers' Reported Awareness,
Attitudes, and Behaviors. PLOS ONE.
http://journals.plos.org/plosone/article?id=10.1371/journal. pone. 0127881
Quested, Tom, E. Marsh, D. Stunell, and A.D. Parry. 2013. Spaghetti Soup: The Complex World
of Food Waste Behaviors, Resources, Conservation and Recycling 70: 43-51.
April 2016 ~~55
-------
Quested, Tom, Robert Ingle and Andrew Parry. 2013. Household Food and Drink Waste in the
United Kingdom 2012. WRAP Report, http://www.wrap.org.uk/sites/files/wrap/hhfdw-
2012-main.pdf.pdf
Quested, Tom and Robert Ingle. 2013. West London Food Waste Prevention Campaign
Evaluation Report. WRAP Report.
http://www.wrap.org.uk/sites/files/wrap/West%20London%20Food%20Waste%20Campa
ign%20Evaluation%20Report_1.pdf
Stockholm Environment Institute, US Center. 2011. Consumption-Based Greenhouse Gas
Inventory for Oregon - 2005. Technical Report for Oregon Department of Environmental
Quality. http://www.deq.state.or.us/lq/pubs/docs/ConsumptionBasedGHGEmissions
lnventoryORTechnicalReport.pdf
Thaler, Richard H. and Cass R. Sunstein. 2008. Nudge: Improving Decisions about Health,
Wealth, and Happiness. Penguin Books.
UN Food and Agriculture Organization. 2013. Food Wastage Footprint: Impacts on Natural
Resources Summary Report, http://www.fao.org/docrep/018/i3347e/i3347e.pdf
USDA Economic Research Service. 2010. Loss-Adjusted Food Availability: Spreadsheets.
http://ers.usda.gov/Data/FoodConsumption/FoodGuideSpreadsheets.htm.
West Coast Climate and Materials Management Forum. July 2013. Toolkit Implementation
Guide for the Food: Too Good to Waste Pilot, http://westcoastclimateforum.com/sites/
westcoastclimateforum/files/related_documents/02_ToolKit_lmplementation_Guide_for_
the_Good_Too_Good_to_Waste_Pilot.pdf
World Resources Institute. 2015. Draft FLWProtocol Accounting and Reporting Standard (FLW
Standard), http://www.wri.org/our-work/project/global-food-loss-and-waste-
measurement-protocol/documents-and-updates#project-tabs
April 2016 56
-------
APPENDIX A: PRE- AND POST-CHALLENGE SURVEY INSTRUMENT
In addition to demographic questions, the following questions were asked in pre- and post-
Challenge surveys conducted by the FTGTW partners.
Pre-Campaign Questions
How strongly do you agree or disagree with the statement: I take steps at home to not waste
food that could have been eaten.
Agree
Somewhat agree
Somewhat disagree
Disagree
Don't know/unsure
In the past year, have you seen or heard anything about the problem of wasted food?
Yes
No
I don't know
If yes, what? [Textbox]
Did this make you more concerned about wasted food?
Yes
No
I feel the same amount of concern as before
To what extent, do any of the following motivate you to minimize the amount of food your
household throws out? Please rate as follows: (1) Not at all (2) A little (3) A fair amount (4) A
great deal
Wasted money I spent buying the food
Wasted time I spent shopping, storing, and/or preparing food
That there are people without enough to eat
The wasted energy and water resources it took to get the food to my plate
Feeling bad about throwing away food that could have been eaten
The contribution of wasted food to global warming
The amount of food that ends up in landfills
Post-Campaign Questions
How strongly do you agree with this statement: I am now more aware of food going to waste in
my household?
Strongly agree
Agree
Neutral
Disagree
Strongly Disagree
April 2016 57
-------
How strongly do you agree with this state: In general, I found the strategies and tool useful.
Strongly agree
Agree
Neutral
Disagree
Strongly Disagree
How strongly do you agree with this statement: I am likely to continue to use the tools and
strategies.
Agree
Somewhat agree
Somewhat disagree
Disagree
Don't know/unsure
To what extent, do any of the following motivate you to minimize the amount of food your
household throws out? Please rate as follows: (1) Not at all (2) A little (3) A fair amount (4) A
great deal
Wasted money I spent buying the food
Wasted time I spent shopping, storing, and/or preparing food
That there are people without enough to eat
The wasted energy and water resources it took to get the food to my plate
Feeling bad about throwing away food that could have been eaten
The contribution of wasted food to global warming
The amount of food that ends up in landfills
April 2016 58
-------
APPENDIX B: CAMPAIGN DESCRIPTIONS
King County. Washington. 2012
FTGTW Partner: King County Solid Waste Division
Community Location: Fall City, WA
Urban-Rural Classification: Peri-urban town with a population of approximately 2000.
Campaign Objectives: Reducing wasted food is a priority for King County in its effort to
achieve Zero Waste. The specific aim of the King County campaign was to test the
effectiveness of the campaign messaging and tools in reducing food waste and to gauge the
impact of a CBSM campaign based on these results.
Community Partners: King County partnered with a local elementary school through their
Green Schools Program. They were assisted by the marketing firm of Colehour + Cohen
who have special expertise in CBSM campaigns.
Target Audience: The target audience for the King County implementation was families
with small children. The campaign was introduced to 110 families with a child enrolled in the
4th grade at the Fall City public elementary school.
Targeted Behaviors: During the pilot, all five waste prevention behaviors were introduced
to the families.
Behavior Change Tools Used in Pilot: This campaign used both the Meals-in-Mind
Shopping List Template and the Fruit and Storage Guide tools. They also structured the
campaign around a modified Challenge as described below under the subheading
"Implementation Choices". In addition, King County developed several other tools including:
a Top Five Ways to Waste Less Food information sheet; Packing a Waste Free Lunch tip
sheet; a blog to keep families informed and motivated; a Food: Too Good to Waste daily tip
PowerPoint presentation; and a Learn More resource list.
Length of Pilot: As one of the first pilots, the King County campaign included time to
develop materials in parallel with the Forum's efforts of nearly a year. The participant
engagement period lasted approximately two months, including time to recruit and assess
and acknowledge the families' participation. The length of the Challenge was five weeks.
Implementation Choices: The invitation to participate in the Food: Too Good to Waste
Challenge was sent via email to the families of the 4th grade children. A King County
representative then visited the classroom to explain to the students why wasted food is bad
for the environment and household economics and distributed the measurement tools (bag
and weekly worksheets). The first week waste collection served to establish a baseline for
the volume of food going to waste. Both preventable and non-edible food waste items were
collected in the same measurement bag to simplify the process. At the start of the second
week of the pilot, all five campaign strategies were introduced. Thereafter, tools were
introduced one at a time at one week intervals. Students were also presented a daily food
waste reduction tip. All families who completed the challenge were given a grocery store
certificate.
April 2016 59
-------
San Benito County. California
Objectives: San Benito County's objective was to test food waste reduction strategies to
inform the county's future food waste collection plans.
Target Population and Sample Size: In San Benito County, lower income Hispanic
families were recruited through a food bank to participate in the pilot. Additional audiences
included a senior citizen center's service clientele and a mom's group.
Targeted Behaviors: San Benito County targeted keeping fruits and vegetables fresh and
using a shopping list with meals-in-mind.
Tools Used in Pilot: This campaign used both the Shopping List Template and the Fruit
and Storage Guide tools as well as the Photo Diary version of the Challenge. In addition,
they used the Workshop Presentation to introduce the challenge to the seniors and mom's
group.
Community Partners: San Benito County partnered with the local food bank in Hollister,
California.
Length of Pilot: Households were asked to complete worksheets recording instances of
their preventable food waste by weight for four weeks.
Implementation Choices: Fliers were placed in food bank bags to invite participation in the
Challenge, followed by phone calls and email (when available), while the workshop
presentation was used to recruit participants from the seniors and the moms' group.
Boulder, Colorado
Objectives: The campaign presented an opportunity for Naropa University to bring
awareness around food waste and the importance of composting to its students.
Target Population and Sample Size: In Boulder, the campaign was introduced to the
student body and faculty of a local university.
Targeted Behaviors: All five behaviors were introduced to participants.
Tools Used in Pilot: The Workshop Presentation was used to introduce the study
participants to the issue and recruit pilot participants.
Community Partners: Naropa University, Naropa Sustainability Council
Length of Pilot: The campaign took approximately three months to complete. Time was
needed to secure permission from the university to conduct the pilot. The four week
Challenge took place in the period leading up to Thanksgiving, beginning in late October.
Implementation Choices: Initial outreach to the community was though an email
announcement that contained basic information about the campaign along with an invitation
to attend a workshop presentation. Participants were also recruited through tabling at the
university's Sustainability Fair at the end of which the workshop presentation was made.
The workshop was followed by a local food panel and dinner in which local foods were
served. Participants were offered both the photo diary and measurement bag options of the
April 2016 ~6u"
-------
Challenge with instructions to measure food waste in weeks 1 and 4. Three raffle prizes
were offered as incentives to participate. Weekly email reminders were sent to participants.
Honolulu, Hawaii
Objectives: The City and County of Honolulu (CCH) are interested in food waste
management solutions that would both lower the costs of landfilling as well as offset the cost
of importing food to the island.43 The Honolulu campaign sought to test CBSM food waste
reduction strategies and tools including a cookbook with local chef-contributed recipes and
food waste prevention tips. It also aimed to see if there was a connection between
preventable food waste and the number of meals outside the home.
Target Population and Sample Size: Out of approximately 210 emails sent, 17 households
were recruited to participate in a four week challenge. The principal audience was young
adults although two households were in their fifties and two had children. The average age
of participants was 34.
Targeted Behaviors: All five behaviors were tested but the "Buy What You Need" strategy
was combined with the "Make a Shopping List with Meals in Mind" strategy and relabeled
"Smart Shopping". In addition, households were encouraged to test recipes for using leftover
ingredients.
Tools Used in Pilot: The Workshop Presentation was used to introduce the Food: Too
Good to Waste Challenge to the household participants. Behavior change tools included: a
food storage guide developed by Eureka Recycling; a menu planner used in the Australian
campaign; an "Eat Me First" prompt; and a cookbook developed by local chefs containing
recipes for using leftover ingredients.
Community Partners: Alexander Lavers researched, directed and managed the campaign
in fulfillment of a Master degree in Environmental Sciences from the University of
Gothenburg, Sweden.
Length of Pilot: The total elapsed time of the campaign was four months. Recruitment took
approximately three weeks, while the length of the Challenge was four weeks. Adapting,
preparing and purchasing materials for the challenge took three weeks as did the data
analysis. In addition, the project organizer spent several months coordinating the
cookbook's development with the contributing restaurants, the graphic designer, and the
county.
Implementation Choices: Recruitment was made by email using personal contacts in two
social networks, the Recycling Branch of the Refuse Division of CCH and a Honolulu
running club. Challenge participants were asked to measure preventable and non-edible
food waste for two weeks after which they measured both types of waste for an additional
two weeks while trying food waste reduction strategies. Non-campaign study cookbook
recipients will receive an option to fill out a survey on their experience with the
43 Alexandra Lavers. 2013. Eat me First! Development and Evaluation of the Food: Too Good to Waste Household Food Waste
Prevention Program in Honolulu, Hawaii, USA.
April 2016
-------
cookbook/toolkit; in return for their responses they will receive a coupon to a restaurant
featured in the cookbook.
Seattle, Washington
Objectives: Seattle Public Utilities sought to gather baseline data on food waste,
specifically, how much of the food waste in Seattle's residential waste stream is edible food.
Target Population and Sample Size: The target population was Seattle's residential
population. Study participants were recruited through a short article in the utility's newsletter
that is mailed to residential customers with their bi-monthly bill. This initial article resulted in
over 500 responses. Approximately 170 customers still expressed interest in participating in
the campaign after receiving a follow-up letter that described what was expected in terms
their participation. 119 households sent data routinely.
Targeted Behaviors: The objective of this study was to collect baseline data on the
composition of household food waste so no behavior change strategies were suggested to
participants.
Tools Used in Pilot: This campaign focused on obtaining data on current food waste
management practices for a representative sample of Seattle Public Utilities customers.
None of the strategies were introduced in this phase of the work.
Community Partners: None
Length of Pilot: The length of campaign implementation was approximately 8 months from
initial outreach to families, through data collection and preliminary analysis of results.
Implementation Choices: A subsequent letter to respondents detailed what was expected
of participants in the pilot: separation of edible and non-edible food waste for 13 weeks,
daily weighing and recording of the food waste weights, and weekly transcribing of this
information to a Survey Monkey questionnaire.
Iowa City. Iowa
FTGTW Partner: City of Iowa City Landfill and Recycling Center
Community Location: Iowa City, IA
Urban-Rural Classification: Mid-sized city with population of 71,600
Campaign Objectives: Gain experience in helping residents reduce wasted food.
Campaign Focus: Educate residents on benefits of reducing waste and support for
curbside collection of food waste with yard waste.
Community Partners: Iowa City Solid Waste Division.
Target Audience: Representative sample of city's population reflecting its age and income
diversity.
Outreach and Engagement Methods: Five neighborhoods with a combined population of
about three hundred households were selected in consultation with the city's solid waste
April 2016 62
-------
manager to represent a variety of ages and income levels. A letter invitation highlighting
their selection was sent to the households along with a survey and a pre-stamped envelope.
The households were asked to return the survey if they chose to participate. Participants
were also offered food scrap curbside collection services for the duration of the study. A
door hanger encouraging participation was later placed at each of the targeted homes.
The pilot organizers held "neighborhood open houses" where they could answer participant
questions and address concerns as well as deliver the materials needed to participate.
These events also presented an opportunity to connect with others in the neighborhood
who had not initially chosen to participate. This was the most personal method of
outreach.
Initially about 50 households indicated they would participate. While participants were self-
selected they were also pre-chosen for their representativeness of the general population.
Length of Pilot: Iowa City conducted a five week Challenge with a baseline period of one
week.
April 2016 63
-------
APPENDIX C: FTGTW COMMUNITY-SCALE MEASUREMENT METHODOLOGY DRAFT
PROTOCOL
Date: August 4, 2014
Updated: Octobers, 2014
Intended Audience: This protocol has been developed to assist organizations responsible for
food waste disposal, such as states, counties and municipalities, in evaluating the impact of a
FTGTW campaign on the amount of wasted food entering a community's solid waste and
or/compost streams.
Research Objective: Determine the impact of a Food: Too Good to Waste (FTGTW) campaign
on reducing the amount of wasted food in a community's residential solid waste (i.e. residual
waste) and/or mixed compost streams using a robust sampling strategy.
Scope: The protocol addresses food waste from food consumed at home, that is, food entering
the home. This would include restaurant food brought home but not restaurant food left on the
plate at the restaurant or discarded at the restaurant. Additionally, the protocol does not account
for food being disposed through other means than in the solid waste or compost streams, such
as backyard composting or flushing liquid wasted foods down the drain.
Research Questions:
What percentage of the target population took steps to reduce food waste during the
campaign?
What was the extent of food reduction in households that acted to reduce wasted food
using campaign strategies?
What is the impact of a FTGTW campaign on reducing the amount of wasted food in the
residential solid waste stream and/or mixed compost stream?
How did campaign (awareness, self-reported new knowledge, self-reported impact)
affect observed change in level of waste?
Community Selection Criteria:
Resources sufficient to mount and evaluate community/neighborhood FTGTW campaign
over six month period
Support from decision-makers
Expertise in statistical analysis or access to such expertise
o Note: This protocol's intent is to obtain statistically robust results that can be
scaled to a community's population. This requires randomized sampling (i.e.
participants are not self-selected) and a fairly large sample. Information on how
to perform a randomized sample study can be found at [SOURCE TO BE
DETERMINED].
Commitment to long-term tracking of wasted food in residential stream
Data Collection Method: Waste composition analysis linked to household survey
April 2016 64
-------
Overview of Waste Composition Analysis Method: A statistically robust sample of households is
recruited to participate in the study. Authorized collectors retrieve the contents of residual waste
and/or compost bins by individual household prior to the introduction of the FTGTW campaign.
The food waste is then sorted and separated from other waste components by trained sorters
and weighed. At the end of the campaign, waste is again collected, sorted and weighed to
determine impact of campaign. A follow-up survey is conducted to determine which households
were reached by the campaign and then acted to reduce food waste.
Pros: Minimum intrusion into household routine.
Cons: Costly, especially if there are two streams to sample. Sorting after mixing may
introduce uncertainties compared to food waste being collected separately at source.
Note: This protocol was designed to be consistent with the Global Food Loss and Waste (FLW)
Measurement Protocol. Development of the FLW Protocol is being coordinated by the World
Resources Institute (WRI) in conjunction with the Consumer Goods Forum (CGF), FAO,
FUSIONS, UNEP, World Business Council for Sustainable Development (WBCSD), and WRAP.
Interested communities will be invited to test the FLW protocol in 2015.
Impact Measurement Protocol
Pre-Planning
Scope out potential partnerships for both campaign and data collection/measurement
tasks.
Investigate potential funding sources.
Secure support of decision-makers.
Planning
Assess fit with community selection criteria (see above).
Determine community objectives and additional research questions, if possible, using a
community engagement process.
Define geographical scope and campaign's target audience (e.g. general population,
families with children, young adults).
Describe and quantify planned FTGTW campaign activities.
Schedule campaign and data collection activities.
Survey and Sampling Design
Define waste producing unit (e.g. household characteristics).
Determine if cluster analysis is desirable and feasible based on available resources.
Information on how to perform a cluster analysis can be found at [SOURCE TO BE
DETERMINED].
Draw up population of potential analysis/survey participants. Potential sources include
waste collection service customer rolls.
April 2016 65
-------
Determine sample size(s) based on level of confidence required and variability in wasted
food levels among units in the population (if there is available data) and draw random
sample.
Determine whether residual waste, compost or both streams are to be sampled
o Note: If the compost stream is the designated stream for wasted food, it may be
possible to sample just the compost stream provided there is available historical
data to support that the majority of food waste is being disposed in this stream.
Determine temporal scope, adjusting for seasonality in wasted food
Determine waste categories (if other than all food waste)
o Note: This protocol is designed to measure all food waste. All food waste
consists of both edible food and food's associated inedible parts (e.g. egg shells,
bones, hard melon rinds). If the research objectives include the quantification of
edible versus non-edible food fractions or the amount of food waste by food
categories, it will be necessary to construct definitions for these sub-categories
and design appropriate sampling and sorting procedures.
Prepare pre- and post-campaign survey instruments. (See FTGTW Standardized
Questions for candidate questions.)
Recruit and train surveyors (possibly community volunteers)
Secure permission from households to sample waste
Survey households to assess their level of awareness around food waste.
Acquire permission to sort their household's waste (opt in). By requesting permission to
access waste bins and sort waste at the end of the survey, the likelihood of household's
agreement will be greater.
Waste and Data Collection
Determine who will perform waste collection and schedule, pre- and post-campaign. The
pre-campaign waste collection should be scheduled to follow the initial survey after
some time (three weeks to a month) to allow for a reversion to normal behaviors that
may have been affected by the survey interviewing.
Determine how waste will be handled (e.g., no compaction to allow for sorting) and
necessary equipment.
Determine sorting procedure, training and personal protective equipment. This should be
consistent with additional research objectives such as determining extent of reduction in
preventable and non-edible waste fractions.
Develop data transmission requirements and form for recording data.
Coordinate with waste handlers.
Track campaign inputs, e.g. number of farmer's market tabling events, number of media
articles, etc.
Post-campaign survey
Perform follow-up survey to identify households who took steps as a result of the
campaign. The survey should assess their awareness of the campaign, what they say
they learned from it, how they describe its influence on their behaviors, and what
aspects of the campaign they saw as being particularly influential.
April 2016 ~66~
-------
Ask about anything specific to their household that may not be learned based on waste
collection, anything that changed since initial collection.
Data Analysis and Reporting
Link weight measurements and survey data.
Provide information on the representativeness of sample for whole population.
Scale-up results to general population using appropriate weighting factors. Information
on how to scale-up results can be found at [SOURCE TO BE DETERMINED].
Resources
Quested, Tom, and Robert Ingle. 2013. "West London Food Waste Prevention Campaign
Evaluation Report."
Describes a similar study conducted in West London and study results.
April 2016 67
-------
United States
Environmental Protection
Agency
Region 10
1200 Sixth Avenue, Suite 900
Seattle, WA 98101-1128
------- |