BENEFITS FROM IMPROVEMENTS
IN
CHESAPEAKE BAY wATER QUALITY
Volume III
of
BENEFIT ANALYSIS USING
INDIRECT OR IMPUTED MARKET METHODS
(Budget Period II).
DEPARTMENT OF AGRICULTURAL AND RESOURCE ECONOMICS
SYMONS HALL
UNIVERSITY OF MARYLAND
COLLEGE PARK 20742
-------
BENEFITS FROM IMPROVEMENTS
IN
CHESAPEAKE BAY WATER QUALITY
Volume // /
of
BENEFIT ANALYSIS USING
INDIRECT OR IMPUTED MARKET METHODS
(Budget Period II)
Prepared • nd Edited by
Nancy E. Bockstael, Kenneth E. McConnell • nd Ivar E. Strand
University of Maryland
Sections Contributed by
Douglas Larson • nd Bruce Madariaga
University of Maryland
Final Draft - April 1988
Principal Investigators
Nancy E. Bockstael • nd Kenneth E. McConnell
EPA Contract No. CR-811043-O1-O
Project Officer
Dr. Peter Caulkins
-------
"The information in this document haa been funded wholly or
in part by the United States Environmental Protection Agenty
under Contract No. 811043-01-0. It haa been subject to the
Agency*, peer and administrative review, and it has . been
approved for publication as an EPA document. Mention of
trade names or commercial products does not constitute
endorsement or recommendation for use. "
-------
Ac knowledgement s
First and foremost, we wish to thank Dr. Peter Caulkins, our
project officer, who has exhibited patience, wisdom, and
appreciation.
Douglas Orr and Firuzeh Arsenjani contributed to the analysis
in this volume. Special thanks are given to Marion Story for
her commitment and technical assistance.
All opinions and remaining errors are the sole responsibility
of the editors. This effort waa funded by USEPA Cooperative
Agreement number CR-81 1043-01-0.
-------
Forward
This is the second of two reports constituting the final report for budget
period II of Cooperative Agreement #811043-01-0, which was initiated and
supported by the Environmental Protection Agency. From the beginning, this
cooperative agreement haa dealt with improving methods of measuring the
benefits of environmental improvements. Budget period I of the agreement
produced two documents which considered theoretical, conceptual and
methodological issues involved, in using hedonic models (Vol. I) and
recreational demand models (Vol. 11) evaluating environmental improvements.
The second budget period's work has extended the work of the first,
especially in the area of recreational demand models. Volume I of budget
period II's final report looks at the theoretical issues of measuring the
benefits of quality changes, the conceptual issues surrounding perceptions of
water quality and methodological issues related to estimating models with
sample selection problems.
The report which follows is the second part of budget period II's final
report. This report provides information on the recreational activities which
take place on the Bay, as well as the monetary values people place on these
activities. While not commissioned with the intent ofhelping in the process of
revising the Bay plans, we hope that the discussions in this report will do
just that.
-------
Executive Summary
For more than ten years, the Chesapeake Bay has been the focus of an
impressive amount of research and an array of environmental programs. The
Chesapeake Bay Program, a cooperative effort by the federal government,
Maryland, Virginia, Pennsylvania, and the District of Columbia, represents a
coordinated commitment to enhancement of the water quality of the Chesapeake
Bay.
Large commitments of money have been made to clean up the Chesapeake
Bay. Yet there is little understanding of the nature and extent of the
benefits which are derived from these massive commitments. Raising this issue
does not imply that either the programs are misguided or need to be justified
on some benefit-cost criterion, for many believe that the cleanup process is an
expression of a fundamental morality that despoiling our surroundings is
wrong. But understanding more precisely how people benefit from cleaner
water in the Bay can help in allocating resources to clean up the water, for
funds must be allocated temporally, spatially and functionally. Perhaps
knowledge about the benefits from water quality improvements can help with
those decisions.
Even though the returns from the Program derive from human benefits--
human use and human health--the specific objectives and implementation
strategies are designed to affect chemical and biological characteristics of the
Bay. The connection between human benefits on the one hand and reductions
of nutrients and toxic materials" on the other remains implicit. Perhaps the
clearest link is between human use and fisheries and wildlife management.
Here the vehicle for linking the strategy and the goal is at least under-
standable, even if the details of this linkage remain indistinct. In other
cases, however, we are left confused as to how the policies impact on humans
and how we would evermeasure the success of these policies in terms oftheir
achievements.
Thie report attempts to focus attention on the human use of the
Chesapeake Bay. It describes something about the nature and level of that
use. It also considers what we know and what we do not know about the
relationship between chemical and biological characteristics of the Bay and
human use. This relationship must be understood in order to address the
more complex measurement of human benefits.
One objective of the report is to provide estimates of Values of
Chesapeake Bay recreational activities and willingness-to-pay estimates of
improvements in water quality associated with these activities. Available data
has been used together with what is known about estimating environmental
benefits. While Chapters 3, 4, 5 and 6 reflect our beat efforts at this task, it
should be kept in mind that benefit estimates have an elusive nature. A
number of different studies have been assembled, and an array of methods
and specifications has been used to provide as much information as is
currently available on the topic.
iv
-------
The benefit estimates themselves do not represent this report's most
important contribution, however. We seek to describe, model and to some
extent explain recreational uses of the Chesapeake. We may have serious
doubts about the precision of willingness-h-pay estimates, but we still learned
a great deal about the factors which matter to people in using the Bay, the
obstacles to their increased enjoyment of the"^ Bay and the distributional
implications of improving the Bay.
Specific information contained in this report includes:
. Maryland boaters, beach users andsportfishermen alter their behavior
in response to poor water quality, as scientifically measured.
• Demographic factors, such es income and location of residence,
influence observed use of the Bay.
. Cent ingent valuation experiments (hypothetical responses) reveal an "
annual willingness to pay in increased taxes of over $100 million for
improvements in Chesapeake Bay water quality.
. Observed behavior of Maryland western shore beach users reveal an
annual willingness to pay for 20 percent improvements in water quality
of between $2 million and $26 million.
•Many of the gains from water quality improvement are concentrated in
the area of heaviest use around Annapolis, Maryland.
The estimates give magnitudes for the annual benefits to residents of the
Baltimore-Waehington area of improving water quality in the Bay in the range
of from $10 million to over $100 million. There are numerous sources of error
and random elements in these estimates, and several activities and populations
have been omitted. But based on these numbers, it seems plausible to
estimate that the annual returns to cleaning up the Chesapeake are at least of
this order of magnitude.
The long-run annual benefits will be higher than these estimates, however.
Firat, as people learn that the Bay has become cleaner, they will adjust their
preferences toward Bay recreation. This is especially true of people who do
not currently use the Bay and are largely excluded from the analysis.
Second, the population and income ofthe area have grown since 1984, and
both are likely to grow more, increasing the demand for and value of
improvements in water quality. Finally, we have ignored the value (both use
and existence value) which households outaide the Baltimore-Washington area
may have for the Bay. The Chesapeake Bay is a nationally prominent
resource. Its improved health is of value to many who will never use it.
In conclusion, we hope this volume will provide a stimulus to decision
makers to refocus attention on human uses of the Bay. Human uses and the
protection of human health have always been the central theme of clean water
legislation, but because of difficulties in relating these to specific standards,
they have often dropped from sight in the formation ofthe actual programs.
We hope to shed some light on ways in which Bay cleanup policies might be
related to the behavior and preferences of actual and potential users of the Bay.
-------
Table of Contents
Chapter Page
1. Enviromental Programs for the Chesapeske Bay 1
The State of the Bay's Water Quality 1
The Current Environmental Programs 2
The Role of This Report 4
2. The Role of Human Use Activities in Defining Goals and Strategies
for the Chesapeake Bay 7
Systematic Evidence of the Link Between Perceptions
and Behavior 8
Seme General Attitudinal Patterns 10
Two Propositions about Water Quality and Behavior .... 12
Results of Focus Group Experiences 17
Sumnary of Focus Group Experience 22
Conclusions 23
3. Recreational Uae of the Bay and Willingness topsy Estimates
for Improvement to the Bay as a Whole ... 25
Recreational Use of the Bay 26
Aggregate Willingness to Pay ,, . . 29
Analysis of Willingness to Pay Responses 30
Results of Analysis by Subgroups of Respondents . . • • • 32
Regional Comparisons ... 36
Existence Value 38
Existence Value Experiment ... 39
Interpretations of Results • • - 40
Conclusions . * 41
4. Effect of Chesapeake Bay Water Quality on Beach Use 42
Tbe SurveyandtheData 44
The Data 45
The Varying Parameter Model 47
The Varying Parameter Model Estimates 50
Estimated Benefit Changes 53
Discrete/Continuous Choice Model 59
The Choice Among Sites 59
Estimation of Discrete Choices Among Beaches ........ 64
The Number of Trips Decision 67
Estimated Benefit Changes 68
Discussion 70
5. Recreational Boating and the Benefits of Improved Water Quality . 71
A Profile of Boaters and Boat Owners 71
The BoatOwnerSurvey 71
Boaters and Boat Owner Characteristic 71
The Importance of Water Quality to Boaters 77
The Behavior of Boat Owners Who Trailer Their Boats 77
TheGeneralModel 77
The Data 80
TheEstimatedModel 81
-------
Modelling the Behavior of Boat Owners Who Do Not Trailer
Their Boats 83
Calculating Estimates of the Benefits of Water Quality .... 85
Improvements for the Trailered Boat Sample 85
6. The Benefits for Recreational Fishing: Striped Bass 90
A Description of the Data 92
TheBasicModel 95
Empirical Results 96
7. Summary and Conclusions 100
Demand for Chesapeake Bay Recreational Activities 100
Estimates of Benefits from Water Quality Improvements .... 102
Caveats 103
Estimates 105
Reference 109
Appendices
The Random Digit Dialing Telephone Survey Procedures 121
Telephone Survey Instrument 125
The User Survey and Sampling Procedure 139
User Survey Instrument 145
vii
-------
Tables
Numbers Page
2.1 Parameter Estimates and Standard Errors for Weibull
Distribution 13
2.2 Effect of Perceptions on Use 14
3.1 Participation Rate in Chesapeake Bay Activities by Activity
and Area, 1984 27
3.2 Joint Participation Rates in Chesapeake Bay Activitiae
By Activity and Region, 1984 . . • —- ,......,,,.... 28'
3.3 Percent of People Willing to Pay Additional Taxes for Water
Quality Acceptable for Swimming, by Amount of Tax 30
3.4 Logistic Model Estimates Related to the Probability
a Respondent Will Accept a Tax Increase 33
3.5 Estimates of Utility Parameters by Income Group 34
3.6 Expected Value of Willingness to Pay for Water Acceptable
for Swimming, by Income Group and User Group, 1984 34
3.7 Estimated Willingness to Pay for Acceptable Water Quality
by Participation and Racial Composition of Household, 1984 ... 35
3.8 Estimated Aggregate Willingness to Pay for Water Quality
Acceptable for Swimming, by Classification and Scenarios, 1984 . 36
3.9 Logistic Model Estimates Related to the Probability a
Respondent Will Accept a Tax Increase to Improve Chesapeake
Bay Water Quality, by Geographic Area 37
3.10 Estimated Willingness to Pay for Acceptable Water Quality
by Region, Participation, and Racial Composition of Household,
1984 38
3.11 Summary Results of Contingent Valuation Experiment
on Existence Value 40
4.1 Average Values of Regression Variables for Visitors, by Beach . . 51
4.2 Tobit Estimates for Beach Demand Model, by Beach 52
4.3 Annual Benefits per Beach User froa a 20 Percent Decrease
in Pollutant, by Beach, 1984 56
4.4 Annual Benefits per Beach User from a 10 Percent Decrease
in Pollutant, by Beach, 1984 57
-------
4.5 Annual Losses per Beach User from a 20 Percent Increase
in Pollutant, by Beach, 1984 58
4.6 Aggregate Benefits/Losses to Users from Changes in Chesapeake
Bay Water Quality, by Beach, 1984 60
4.7 Logit Regression for Selection Among Sites 66
4.8 Log it Analysis for Selection Between State Parks and
Local Beaches „»,„.,,,»,,,».... 66
4.9 Ordinary Least Squares Regression for "Choice Occasions"
or Trips 68
4.10 Comparison of Benefits Based on a Varying Parameter Model
and-Discrete/Continuous Choice 70
5.1 Average Characteristics of Boaters and Non-Boaters
in Baltimore/Washington SMSA, 1984 73
5.2 Average Characteristics of Boat Owners and Non-Boat Owners
in Baltimore/Washington SMSA, 1984 73
5.3 Characteristics of Boats and Boat Owners from
Boat Owners' Survey, 1983 74
5.4 Numbers of Trailered Boats and Boats Kept in the Water
By Residence, 1983 76
5.5 Percent of Boaters Who Fish or Swim While Boating,
Boat Owners' Survey, 1983 77
5.6 Factors Cited Most Important in the Selection of a Boating
Area in Maryland - Percent Response from 718 Boat Owners"
WhoTrailerTheirBoats 78
5.7 Factors Cited Most Important in the Selection of a Boating
Area in Maryland - Responses from 788 Boat Owners Who Keep
Their Boats in Marinas 78
5.8 Estimated Tobit Demand Coefficients: Maryland Counties, 1983 . . 82
5.9 Estimation Results from Second Stage 83
5.10 Estimated Demand for Boating Tripe - Boats Kept in Marinaa ... 84
5.11 Estimated Demand for Boating Trips - Boats Kept in Marinas.
Fishing Behavior 85
5.12 Par Boater Annual Benefits from a 10% Decrease in Pollutant
by Geographical Area 87
IX
-------
5.13 Per Boater Annual Benefits from a 20% Decrease in Pollutant
by Geographical Area 88
5.14 Per Boater Annual Losses from a 20% Increase in pollutant
by Geographical Area 89
6.1 Characteristics of Striped Bass Fishermen and Other
Fishermen/Hunters in the Sample 91
6.2 Sample Distribution Number of Fishermen, Days of Striped
Bass Fishing in 1980, and Catch Rate, by Regions 93
6.3 Tobit Estimation of the Demand for Striped Bass Fishing 97
6.4 Aggregate Consumers' Surplus for Striped Bass Fishing:
Effect of Changing Catch Rates 98
7,1 Aggregate Benefits for Three Water-related Activities
from a "20%" Improvement in the Chesapeake Bay's
Water Quality, 1987 dollars 106
7.2 Aggregate Benefits from Water Quality Irrprovements-
Contingent Valuation, 1967 dollars 107
-------
PigUrea
Numbers Page
2.1 The Sampling Region for the Telephone Survey and Location of
the Beaches Used in the Intercept Survey 9
,2.2 Annual Net Change in Swimming Habits 1970-1964 15
2.3 Cumulative Net Change in Swimming Habits 1970-1964 16
XI
-------
Chapter 1
Environmental Programs for the Chesapeake Bay
Over the past ten years, the Chesapeake Bay has been the focus of an
impressive amount of research and the beneficiary of a great many
environmental programs. Concentrated efforts began in 1976 when the
Congress directed the U. S. Environmental Protection Agency to conduct a
five-year study of the Bay's resources and water quality. The study, which
focused on three major problems of the Bay--nutrient overenrichment, toxic
substances? and the decline of submerged aquatic vegetation--prompted action.
In 1983 the three surrounding states, the District of Columbia, and the federal
government signed a pact, the Chesapeake Bay Agreement of1983, committing
them to improve and protect water quality of the Chesapeake Bay through
coordinated environmental enhancement activities. In late 1987 a new
agreement was signed.
The State of the Bay's Water Quality
Concern over the Chesapeake Bay stems from declines in direct and
indirect measures of the quality of the Bay's waters. The most apparent
measures are related to the productivity of the Bay. Reduction in fish
landings, combined with an awareness of the increasing loads ofpollutants and
their consequences, led scientists to assess the Bay's water quality.
The use of the term "quality" in assessing the Bay connotes a set of
standards goals, or ideals with which the current conditions ofthe Bay can
be compared. The quality of the water depends on one's standards, and the
relevant standards depend on intended uses and frame of reference. For
example, if the most desired use of the water were for transportation, then
the Bay's current water quality would be quite satisfactory. At the other
extreme, if one's standards are derived relative to the state ofthe Chesapeake
Bay three centuries ago, its current quality is clearly too low.
Since the thrust of the Chesapeake Bay program cornea from observed
declines in ecosystem productivity, it is useful to summarize the nature of
those declines. Summary measures give the status of the Bay as a whole, but
mask considerable differences in quality between the upper and lower Bay and
among the various river systems and inlets of the Bay. The following
measures suggest the nature of the thinking that led to the conclusion that
the Chesapeake Bay waa declining in quality.
There are two kinds of evidence of the historical decline in the Bay's
water quality. First there are scientific measures which are indicators of
impairment of the Bay as a functioning ecosystem. A common measure of water
quality is the level of dissolved oxygen in the water. This is oxygen available
to various plant and animal life. its absence can eliminte higher forms of life
from ecosystems. Studies have shown that the extent ofwater with little or
no dissolved oxygen has increased by 15-fold in the last 30 years.
-------
Another indicator of water quality is the level of nutrients (nitrogen and
phosphorus, mainly). These nutrients, while not harmful per se, enhance algal
growth, whose decay increases the demand for oxygen. The increase of
nutrients in the Bay's waters is an indirect consequence of population growth,
changing technology and industrial and agricultural expansion in the area.
The decline in submerged aquatic vegetation (SAV) is another indicator of
the decline in the Bay's water quality. The decline in SAV is connected with
turbidity and growth of epiphytes and phytoplankton, by excessive
nutrification. The loss of SAV means less suitable habitat for spawning finfish
and shellfish. There are ofcourse many other indirect measures of the
declining health of the Bay. They all reinforce the notion that human factors
are destroying the traditional ecological linkages of the Bay.
There are other signs of declining water quality more cogent to the lay
public. Landings ofwell-known anadromous species such as rockfish and
shad have dropped precipitously in the past several decades. Oyster harvest
and oyster reproduction have also declined in the past decade. There is some
ambiguity in the use of landings as a measure of water quality, of course. A
considerable increase in effort devoted to harvesting fish has happened to
coincide with the increase of effluents over time. Further, natural phenomena
such as hurricane Agnes (1972) induce cyclical variations in finfish and
shellfish reproduction. Nevertheless, there can be little doubt that the quality
of the Chesapeake Bay's waters has declined, both in terms of the ecological
health of the estuary and the benefits to humans of its use.
The Current Environmental Programs
The foundation of the Chesapeake Bay Program is the Clean Water Act, the
ongoing federal environmental legislation dealing with water. Under the Clean
Water Act, appropriations have been made available annually to the Chesapeake
Bay Program, providing both its operating budget and its grant funds. The
relationship between the federal legislation and Chesapeake Bay activities goes
beyond funding, ofcourse, since the Clean Water Act establishes the
guidelines by which states then set specific water quality standards. The
Water Quality Standards Handbook is the most recent document which contains
the guidelines prepared by EPA to assist states in implementing 1983 revisions
of the water quality regulations.
The Handbook defines acceptable approaches by which water quality baaed
effluent limitations may be determined. Whether the pollutant specific or
biomonitoring approach is taken, however, states must adopt criteria which are
sufficient to protect the "designated uses" of a water body. Determination of
designted uses requires an "attainability analysis," i.e. physical, chemical and
biological studies to identify the suitable potential uses of the water and to
determine whether these uses have been impaired. There is, throughout, a
clear sense of the central position which human use activities should play in
the setting of standards and the overriding obligation states should feel
toward the protection of human health where 'people are involved in
recreational uses of aquatic resources.
-------
The first plan of the Chesapeake Bay Commission waa the Chesapeake Bay
Restoration and Protection Plan of September 1985. This is currently the
central document describing the goals of the Chesapeake Bay Program and the
means by which these goals are being achieved. The general goals as stated
in the plan are to
"Improve and protect the water quality and living resources of the
Chesapeake Bay estuarine system (in order) to restore and maintain
the Bay's ecological integrity, productivity, and beneficial uses and
to protect public health." l
The goals of the Restoration and Protection Plan are boad, and include both
ecological and human health, as well as productive use by humans. By and
large, however, there is no clear connection between the goals of the Bay
Program which emphasize human health and human use and the means by
which humans benefit from implementation.
To accomplish the broad goals, specific objectives and implementation
strategies have been developed. Many of these strategies are designed to
reduce orcontrol nutrients. Major strategies to control point sources of
nutrienta include plane to provide grants to design, construct, operate and
maintain sewage treatment facilities, and plans to support phosphorous removal
projects at treatment plants. Plans to support nitrogen removal at treatment
plants have not been proposed, except for an experimental project conducted
by the State ofMaryland in the Patuxent estuary.
The primary stratagy established to control non-point sources of nutrients
to the Bay has been to subsidize the implementation of "Best Management
Practices" (BMPs) to reduce runoff from urban, forested, and in particular,
agricultural lands. Maryland, Virginia, and Pennsylvania have instituted cost-
sharing programs to promote agricultural BMPs. Among the agricultural BMPs
supported through cost-sharing have been: strip cropping, buffer stripping,
terrace and diversion construction, animal waste system installation, and
reduced tillage planting. Some of these practices are employed to reduce
sediment, pesticide, and herbicide runoff, as well as nutrient runoff. A
secondary strategy to control non-point source pollution is to control urban
runoff, in particular combined sewer overflows. Tactics to control combined
sewer overflow include revamping ofsewer systems and building holding
ponds. The state of Maryland has also enacted legislation banning the use of
phosphate detergents and controlling development along the Bay's shoreline.
The Chesapeake Bay Restoration and Protection Plan hae enacted a series
of additional policies to reduce or control the level of toxic materials in the
Bay, Among these policies are programs to support pretreatment plans to
reduce the discharge of metals and organics from sewage treatment plants
resulting from industrial sources, to fund dechlorination processes to reduce
chlorine discharges into critical finfish and shellfish areas, and to implement
oil spill response plans.
•"•Chesapeake Executive Council, Chesapeake Bay Restoration and Protection
Plan U. S. Environmental Protection Agency, Sept. 1985.
-------
Lastly, the Chesapeake Bay Restoration and Protection plan has instituted
a series of policies designed directly to "provide for the restoration and
protection of living resources and their habitats and ecological relationships" 2
in the Chesapeake. Among these policies were programs to develop
comprehensive fisheries management plans, expand oyster repletion activities,
improve waterfowl and wildlife habitat, and re-establish submerged aquatic
vegetation.
There is no "clear connection between the implementation strategies
mentioned above and the goals of the Chesapeake Bay Program. The goals of
the program are couched in terms of human benefits--human health and human
use--but the specific objectives and implementation strategies are designed to
affect chemical and biological characteristics of the Bay. The connection
between human benefits on the one hand and reductions ofnutrients and toxic
materials on the other remains implicit, unsubstantiated and unarticulated.
Perhaps the clearest connection is between human use and fisheries and
wildlife management. Here the vehicle for linking the strategy and the goal is
at least understandable? even if the details of this linkage remain indistinct.
In other cases, however, we are left confused as to how the policies impact on
humans and how we would ever measure the success of these policies in terms
of their achievement of the Program's goals. In implementation, the focus on
human use seems to have been lost.
The Role of This Report
This report attempts to focus attention on the human use of the
Chesapeake Bay. The report describes something about the nature and level
of that use. It also considers what we know and what we do not know about
the relationship between chemical and biological characteristic of the Bay and
human use. This relationship must be understood in order to grapple with
the more complex measurement of human benefits.
Large commitments of money have been made to clean up the Chesapeake
Bay. Yet there is little understanding of the nature and extent of the
benefits which are derived from these massive commitments. How do people
gain from the cleanup? Asking this question does not imply that either the
programs are misguided or need to be justified on some benefit-cost criterion,
for many believe that the cleanup process is an expression of a fundamental
morality that despoiling our surroundings is wrong. Whataver the motivation
for environment, al improvement, we believe that understanding more precisely
how people benefit from cleaner water in the Bay can help in allocating
resources to clean up the water. Moral imperative are of limited usefulness
in the tactics of cleaning up the Bay, Even with commitments for a cleanup of
the Bay, one must allocate those funds temporally, spatially and functionally.
Perhaps knowledge about the benefits from water quality improvements can
help with those decisions.
Chesapeake Bay Restoration and Protection Plan, Chapter 2, page I,
-------
One of our objectives is to provide some initial estimates of values of
Chesapeake Bay recreational activities and willingness-to-pay estimates of
improvements in water quality associated with these activities. We have used
available data together with what we know about estimating environmental
benefits (see Bockstael, Hanemann and Strand, 1985; and McConnell, Bockstael
and Strand, 1987) to determine these "ball park" willingness-b-pay figures.
While Chapters 3, 4, 5 and 6 reflect our best efforts at this task, it should be
kept in mind that benefit estimates have an elusive nature. Much has been
written about the imprecision ofnon-market benefit estimation, and we will
have more to say on the question in this report. The usual difficulties are
compounded by generally sketchy data. No new surveys were conducted for
this study, and only one ofthe surveys used in the subsequent chapters was
designed with benefit estimation in mind. Nonetheless we have put together a
number of different studies and have used an array of methods and
specifications to provide as much information as is currently available on the
topic. We provide estimates ofaggregate benefits for a variety of recreational
activities. We have also tried to provide information on the relative
magnitudes of benefits which are likely to accrue to different groups of users
and to improvements in different geographical areas.
The benefit estimates themselves do not represent this report's most
important contributions however. We seek to describe, model and to some
extent explain recreational uses of the Chesapeake. The report represents an
attempt to begin to understand the preferences and behavior of individuals
toward the Bay. Models ofbehavior are essential to benefit estimation. Even
in the face ofhuge uncertainty over benefit estimates, the underlying
behavioral models can provide useful and reliable information. We may have
serious doubts about the precision of willingness-to-pay estimates, but we can
still learn a great deal about the factors which matter to people in using the
Bay, the obstacles to their increased enjoyment of the Bay and the
distributional implications of improving the Bay.
The Restoration and Protection Plan is an interim plan, "the first iteration
of the planning effort implemented in response to this commitment. " As such
it is a first move in the direction ofChesapeake Bay improvements but it is
sub ject to revision and fine-tuning as goals of environmental improvement
become clearer and information about problems, technology and costs becomes
more sophisticated.
What we hope this volume will provide is a stimulus to decision makers to
refocus attention on human uses of the Bay, as the goals and the strategies
for achieving these goals are fine-tuned in the coming year. Human uses and
the protection of human health have always been the central theme of clean
water legislation, but because of difficulties in relating these to specific
standards, they have often dropped from sight in the formation ofthe actual
programs, We hope to shed some light on ways in which Bay cleanup policies
might be related to the behavior and preferences of actual and potential users
of the Bay.
-------
This report on the Chesapeake Bay is part of a larger EPA Cooperative
Agreement. The initial agreement dealt with improving methods of measuring
the benefits ofenvironmental improvement, and did not deal with the
Chesapeake Bay. This report provides information on the recreational
activities which take place on the Bay, as well as the monetary values people
place on these activities. While not commissioned with the intent of helping in
the process of revising the Bay plans, we hope that the discussions in this
report will do just that.
-------
Chapter 2
The Role of Human Use Activities in Defining Goals and Strategies
for the Chesapeake Bay
According to the EPA's Water Quality Standards Handbook. "States must
adopt water quality criteria sufficient to protect designated uses. " In the
process of developing standards, if water body assessments are called for,
they must "characterize present uses, uses impaired or precluded, and the
reasons why uses are impaired or precluded. "
The definition of designated uses which must be protected remains a
murky issue. Underlying much of the document is the implicit assumption that
chemical, physical and biological parameters can be used to define uses. On
the other hand there is some acknowledgement of the reality that human
activity does not necessarily align itself *with physical and chemical water
properties:
"The basis of this policy is that the States and EPA have
an obligation to do as much as possible to protect the health
of the public even though it may not make sense to encourage
use of a stream for swimming or wading because of physical
conditions. In certain instances, particularly urban areas,
people will use whatever water bodies are available for
recreation. "
At the heart of the dilemma is the disparity between the goals which are
couched in terms of human uses and the targets of policy actions which are
callibrated in ambient pollution levels. There is no one-to-one mapping
between human use and scientific measurement. Failure to come to terms with
their relationship has lead to something of a schizophrenia about human
activity and scientific measurement in the Water Quality Standards. This
schizophrenia is not unlike that found in the recreational demand literature
which typically seeks to value environmental amenities by relating behavior to
changes in ambient pollution levels without explaining how people perceive
pollution.
The connection between human activity and scientific measures of ambient
water quality is further obscured by the considerable ambiguity one finds in
both these discussions about the ways in which individuals gain from water
quality improvements. In the EPA Water Quality Standards Handbook, we find
repeatad reference to protecting "uses," i.e. recreational activities, and at the
same time a sense of obligation to protect human health. One might argue
that these two concepts are coincidental, that we are interested in the health
of humans as they participate in recreational activities in the Bay. In terms
of pollutants which the individual can see (or smell or learn about in some
less direct way), an individual's criteria forusing the Bay are likely to exceed
those minimum standards required to avoid health risks. On the other hand,
the individual is totally unaware of health risks stemming from pollutants
which cannot be easily detected. Thus many recreation decisions are not
directly guided by quality characteristics associated with health standards.
7
-------
Whether modelling recreational decisions or developing standards, we face
a dilemma when we try to link water quality and human behavior. The
obvious way to measure water quality is through chemical and physical
readings perhaps supplemented by assessments of the biological resources.
But water quality improvement is undertaken to enhance society's welfare
which is recognized as deriving, in large part, from human use. Does human
use respond to the changes in the chemical and physical composition of the
water which physical and biological sciences measure? Are factors which
affect their health the sole factors which matter to people? Can people
perceive changes in these measures?
These questions not only plague benefit measurement, they are central to
environmental policy making. If gods are fundamentally human oriented and
standards are scientifically based, then the disparity between the two must be
resolved for environmental regulation to achieve its potential. In what follows
we present evidence about the human side of the problem.
First we present descriptive information on the variation in household
perceptions of the Bay based on two surveys. These survey results do not
reveal anything about the formation of perceptions however. To gain some
insight into this process, we use the focus group approach, The material
discussed in an earlier volume of this report is summarized in this chapter,
and insights from our focus group experience are offered which are specific
to water quality in the Chesapeake. From these various sources, we draw
some implications for environmental policy.
Systematic Evidence of the Link Between Percept.ions and Behavior
Evidence on what people think of the water quality ofthe Chesapeake and
how they behave toward the Bay comes from two surveys: an on-site survey
of beach users and a telephone survey (Figure 2.1). Our telephone survey
was conducted May 1, 1984 to September 1, 1984, Research Triangle Institute
(RTI) collected data for the University of Maryland on recreational use and
perceptions of the Chesapeake Bay using a random telephone survey. The
telephone survey was planned and executed jointly with an on-site survey of
beach users at western shore Chesapeake beaches. The 1,044 households with
completed interviews were residents ofthe Baltimore and Washington SMSA's.
Demographic, attitudinal and use data were obtained. Chapter 3 reports on
the analysis of use patterns and activities derived from the telephone survey.
It also provides estimates of willingness to pay for Bay improvements.
In this chapter the attitudinal information obtained from the telephone
survey is examined. This survey allows inferences to be made about the
impact of perceptions on decisions to use the Bay. It also facilitates
expansion of sample patterns of behavior and perceptions to the population.
The phone survey provides information about broad perceptions of the
Bay, but without details about regional variation in quality. Specific regional
information comes from the user survey, which gathers data about patterns of
use and perceptions for 408 users of twelve beaches on the western shore of
the Chesapeake. The user survey is described in detail in Chapter 4 of this
volume.
8
-------
Figure 2.1
The Sampling Region for the Telephone Survey ( jf7 ) and the Beaches (
Used in the Intercept Survey
-------
Some General Attitudinal Patterns
Several important patterns emerged from the telephone sample. Forty-
three percent of the households responded that they had used (31%) or
intended to use (12%) the Chesapeake Bay for recreation in 1984. For the
users, boating (73% of users), sightseeing (69%), beach use (66%) 1, swimming
(64%), and fishing (56%) were quite popular activities. Hunting (5%) and other
uses (20%) were not as prevalent responses. The percent of the Bay users
who visited eastern shore beaches (44%) was nearly the same as the
percentage who visited western shore beaches (47%). Most of the users (81%)
also visited ocean beaches in 1984.
Among those who did not use the Bay, reasons for non-use included:
Not interested in water-related recreation (40% ofnon-users)
Too busy to use (25%)
Takes too long to get there (20%)
Unable to use for health reasons (6%)
Water quality unacceptable (5%)
Costs too much (5%)
Too crowded (3%)
Too many jellyfish (3%)
Other (31%)
Personal preferences and the scarcity of households' leisure time were
important considerations. Trip costs and poor water quality were not cited as
often (5% of the time) but were still recognized as reasons.
The fact that only 5 percent of our telephone sample stated that the
Chesapeake Bay water quality was responsible for their nonparticipation may
diminish one's assessment of the role of water quality to Chesapeake citizens.
For one thing, water quality in the Bay is not homogeneous--it varies
substantially and respondents in our sample recognized the differences.
Suppose respondents living in Annapolis believe Annapolis' water to be
unsuitable for swimming but water at Pt. Lookout to be suitable. These
individuals may respond that time was the prohibiting factor. It takes nearly
three hours to travel from Annapolis to Pt. Lookout. From another
perspective, people who do not visit the Bay because of time constraints may
know little about the Bay's water quality and will not cite water quality as a
problem.
A number of other questions were included in both surveys to learn more
about perceptions of water quality. For example, we asked telephone
respondents,
"Do you consider the water quality in the Chesapeake to be
acceptable or unacceptable for swimming and/or other water
activities?"
Only 43 percent of the telephone respondents answered "acceptable,"
Alternatively stated, 57 percent of a random sample from the Baltimore and
Washington SMSA's found the Bay water quality unacceptable for swimming
and/or other water activities.
10
-------
A similar question was asked in the user survey concerning specific
western shore beaches with which the respondent was familiar. Even here
there was negative reaction to the water quality, especially at certain beaches.
We discovered, in fact, that some households found the Bay's water quality
unacceptable? but nonetheless used it. There are several explanations for this
apparent inconsistency, It is possible that households may find the water
unacceptable for certain kinds of activities (swimming) but not for others
(beach use or boating) . Households may find the water quality acceptable for
activities with short duration or during certain seasons of the year when the
Bay appears cleaner. Finally, as mentioned earlier, the question abstracts
entirely from the natural heterogeneity of the Bay. Some areas may be
unacceptable to just about everyone while others may appear clean to the
majority.
Some insight was provided by the user survey into the specific factors
considered important in visiting a Chesapeake public beach. We asked
individuals to rank each of five factors on a scale of one to five, with five
being the most important to them. The weighted averages and medians were:
Mean Median
Presence of floating debris or oil
Presence of odors
Presence of jellyfish
Presence of cloudy water
Presence of seaweed and other
aquatic plants
4.32
3.44
3.41
1.97
1.85
5
4
3
2
1
These numbers indicate that floating debris and oil is the major quality
criterion, with odors and jellyfish being the next most important.
The question was re-analyzed by considering the differential responses of
users who came into contact with the water (swimmers and waders) and those
who did not (sunbathers, etc.). The contact users cited odor as the most
important or the second most important criterion 56 percent of the time,
whereas non-users cited it as highly important only 16 percent of the time.
On the other hand, the presence of jellyfish was considered to rank as the
first or second factor for the non-contact users 84 percent of the time but
only 37 percent for individuals who were in contact with the water. These
results are somewhat difficult to interpret because we cannot determine cause
and effect. Logically, those moat bothered by jellyfish are likely to refrain
from entering the water. However, those who go into the water are more
likely than those who don't to detect unpleasant odors.
In any event, of the five factors deemed important and perceivable to
beach users, three are characteristics which could be linked with water
quality. It is interesting, for the purpose ofkeeping our perspective, that a
natural factor (jellyfish ) ranks among the unpleasant features ofthe Bay.
11
-------
Two Propositions about Water Quality and Behavior
To investigate the relation between current water quality and the use of
the Bay, we examined two simple propositions.
Proposition 1: The percentage of respondents at a particular beach who
find the water quality unacceptable is related to water
quality as measured by scientific water quality readings
at the site.
Proposition 2: An individual's use of the Bay is related to his/her
assessment of whether the water quality is acceptable.
Affirmation of Proposition. 1 implies a positive relationship between
individual behavior and the objective measures of water quality upon which
environmental policy is based. Proposition 2, if true, indicates that people are
consistent in matching their behavior to their perceptions of water quality.
Both propositions are important in making the connection between
environmental improvements and behavior-based benefit measures.
The Bay is a well studied ecosystem and has been the focus of much
attention by the U. S. Environmental Protection Agency and the states of
Maryland and Virginia (e.g. U. S. EPA, 1983). Also, Maryland counties on the
western shore sample water at the beaches on a monthly basis in compliance
with Maryland health requirements. These various sources provide objective
measures of water quality at the Chesapeake beaches, and allow examination of
the relationship between users' perceptions and objective measures at the
beaches visited.
As mentioned earlier, the user survey instrument contained a question
asking respondents to judge whether specific beaches on the western shore
were acceptable or unacceptable for swimming or other water related activities.
To answer, the respondent was not required to have used the beach but only
to be familiar with it. The water quality at Sandy Point State Park was
familiar to the largest percentage of people (63 percent), whereas only one
person knew about the water quality at Camp Merrick. The percentage of
those familiar with a beach who found the water quality at that beach
acceptable varied from 94 percent at Rocky Point State Park to 12 percent at
a Baltimore Park, a beach used primarily by picnickers.
As a guide to the sample's responsiveness to water quality, the percentage
of people not finding the water quality acceptable (PCNA) at a beach was
regressed on the moot probable fecal coliform count (FCC) for that beach.
The fecal coliform counts were collected at the beaches during the swimming
season by county officials. Unfortunately, the FCC measurement was available
for only nine of our twelve beaches.
One might argue that individuals would have no way of perceiving fecal
coliform. However, a high FCC might manifest itself in odors or may be
correlated with other factors which cause visible changes in the water. Of
course, periodically high counts could cause a beach to be occasionally closed
by the health officials, a practice that could "brand" the water quality at
certain beaches. Since there were five examples of beach closures, the
12
-------
estimation of a relationship between PCNA and FCC should serve as a small
test of the ability of individual to perceive factors correlated with FCC.
To assure that obvious restrictions on the PCNA and FCC variables were
not violated by our functional forms a Weibull distribution was assumed:
(2.1) PCNA = 1 - exp[-(FCC/«)*),
where »is the shape parameter and & is the scale parameter. Using a
non-linear least squares routine, we obtained parameter estimates of 2,537 and
.49 for 6 and • respectively with ratios of parameter values to standard
errors greater than two in both cases (see Table 2.1). These results support
Proposition I. There is an apparent connection between objective measures of
water quality at a beach and households' perceptions that water quality at the
beach is acceptable.
Table 2.1
Parameter Estimates and Standard Errors
for Weibull Distribution
Coefficient
Estimate .495 2,537.
Standard Error .095 923.
The value of the shape parameter suggests that the percent of beach
users who find water quality unacceptable is concave in the water quality
variable. (A sufficient condition for concavity is •< 1, i > 0.) To find the
fecal coliform level for a given level of acceptance, equation (1) is inverted
and estimated coefficients inserted
FCC = 2,537 •(-ln(PCNA))a-oa.
For an acceptance rate of 90 percent, the estimated maximum median fecal
coliform count is in the order of 25 fecal coliform per 100 ml. At fecal
coliform counts of200, 75 percent of the users are estimated to accept the
quality. At counts of 1,200, this estimated ratio drops to 50 percent.
The second proposition was tested using the telephone survey response.
Households were asked whether anyone in their household had changed
(stopped or started) swimming patterns in the Chesapeake because ofwater
quality. Two hundred seven of the 1,044 telephone respondents stated they
had stopped, and 26 stated they had started. Of those who stopped, 75
13
-------
percent believed the water quality was unacceptable. In comparison, 53
percent of those who did not change thought the water quality was
unacceptable. Finally, the water quality was believed unacceptable by 42
percent of those who started to swim.
We regressed the individual binary response as to whether or not they
stopped swimming against their acceptance of the water quality. The following
logistic probability model was estimated:
P = {1 + exp[a0 +
where
0 if the household stopped using the Bay
P =
[ 1 if the household did not stop
WQP " water quality percept ion
(1 if acceptable, 0 otherwise)
A maximum likelihood estimation approach produced the effects reported in
Table 2.2. Water quality perception appeared to have a positive statistically
significant impact on whether the household continued swimming, indicating
some relationship between users' perceptions of water quality and their use of
the water. This result provides support for Proposition 2 that behavior is
related to perceptions. Nonetheless, some people who consider water quality
unacceptable are still observed to swim.
Table 2.2
Effect ofPerceptions on Use
Effect Estimate (al) Standard Error
Intercept 1.54 .1?
Water Quality Perception .57 .13
Sample Size = 503
Additional insight into the first proposition can be gained from an analysis
of temporal changes in household habits of using the Chesapeake. It is the
consensus among scientists that the water quality of the Chesapeake fell
substantially over the period 1950-1980 (EPA, 1983). The living resources of
the Bay have been used as a primary indicator of this decline. Submerged
aquatic vegetation and anadromous fish stocks are among those living
resources whose dramatic decrease over this period has been cited. If we can
14
-------
show that contemporaneous with the decline in objective measures of water
quality, individuals were more likely to quit using the Bay, we have additional
indirect evidence of the link between behavior and water quality,
Responses from the telephone survey were used to develop a time series
on the percentage of households who changed their swimming participation.
The procedure was fairly complicated since the size of the population eligible
to change its behavior varied from year to year. That is, consideration had
to be made for how long the household had lived in the Chesapeake region
and whether they had previously stopped swimming. For example, households
responding that they stopped swimming in 1979 clearly were not eligible to
stop again in 1982.
The time series is shown in Figure 2.2. Approximately one percent of the
eligible households stopped swimming each year in the early 1970's. This
increased to around five percent per year in the early 1980's. The trend is
definitely one of increasing non-participation in swimming over the time in
which it is believed that declines in water quality were occurring. Although
the overall pattern is a diminishing one, there appears to be a possible
modification of the trend near the end of the period.
Figure 2.2
Annual Net Change in Swimming Habits
1970 -1984
Net
Change
in
SwiMiin
Habits
1970
1975
19RO
15
-------
If one is bold, enough to assume no reversals in habit for a particular
household, the individual percentages can be combined to show the cumulative
effect ofwater quality change on a given 1970 population Of households
(Figure 2.3). With this assumption, 30 percent of households that had been in
the area in 1970 would have had a Q ember who ceased swimming by 1984.
Figure 2.3
Cumulative Net Change in Swimming Habits
for a Popu ation in Residence in 1970 and Remaining in Resdence until
1984
TOTAL
CHANGE
IN
SWIWING
HABITS
.1
II7S
mi
16
-------
The decline is interesting for a number of reasons, but mostly because it
shows the potential stock of individuals who could be enticed to return to the
Bay if water quality were improved. Benefits from water quality improvements
will derive from increased desirability of current recreational trips, more (of
these higher quality) trips taken by current users, and finally trips taken by
those who are currently non-users but are enticed into the activity by higher
quality. The above analysis suggests that water quality improvements could,
over a sufficiently long time, attract a large number of new (or returning)
users. To the extent that demand analysis is based on current use and fails
to predict accurately this potentially large number of new entrants, benefit
estimates will be understated.
Results of Some Focus Group Experiences
The above evidence about the relationship among (a) use of the Bay, (b)
subjective perceptions of the Bay's water quality, and (c) objective measures
of specific attributes of water quality in the Bay suggests in general that
changes in water quality have an effect on behavior. In subsequent chapters
we show how economists can use information about changes - in behavior
induced by quality changes to assess the economic gains from water quality
improvement.
Good theory, convincing benefit measurements, and effective environmental
policy do not require that individuals act knowingly and mechanically in
response to changes in ambient quality. In fact, casual observations suggest
that many people have only vague notions about environmental quality, and
act unconsciously in response to changes in quality. However, much can be
gained by understanding better the link between perceptions and behavioral
changes. How are perceptions formed? Which aspects of water quality,
objectively measured, matter most to people? These and other questions about
the formation of perceptions require some insight into individual motives.
Traditional research methods have not been very helpful in obtaining
these insights. In contrast, focus groups (Reynolds and Johnson; Caldor;
Desvousges and Smith) have been found to be a useful means of investigating
the existence and formation of subjective perceptions on environmental issues
and marketing questions. Focus groups are group interviews conducted in the
form of informal discussion sessions under the guidance ofa neutral
moderator. Participants are encouraged to talk at will and describe personal
experiences, anecdotes, and acquired knowledge. The moderator merely
encourage participation by all, mediates arguments and spurs conversation
and thought through questions carefully designed to give direction to the
discussion. By encouraging participants to reveal thought processes and
levels ofawareness, their motives begin to emerge.
For this study we conducted two focus groups of a quite different nature.
The groups were made up of 10 to 15 individuals who had some common
association with one another. Each session lasted about one and a half hours.
In each group there were Chesapeake Bay users and nonusers; however, the
groups were chosen so as to be heavily weighted towards people familiar with
the Bay.
17
-------
One focus group consisted of students at the University of Maryland in
College Park, who were members of a wildlife conservation group. Many of
these students had taken environmentally related courses. A number of these
students had grown up in close proximity to the Chesapeake Bay, as active
users of the Bay.
The second group consisted of residents of neighborhoods along the
western shore of the Chesapeake in the vicinity of Plum Point, Maryland.
Their participation was solicited through an announcement in a local
newspaper. Many, although not all, of the individuals were retired, and all
lived near the shore year round. Their backgrounds and education were
quite varied.
These groups were polar in several respects. The college group was
young, formally educated in many scientific and environmental matters, and
tended to be active users of the Bay. The Plum Point group was older, often
retired, not necessarily fluent in scientific and technical matters, and often
somewhat passive in their use of the Bay. The groups shared the
characteristics ofhaving no small children and not being actively engaged in
building careers.
In each case, the moderator presented a formal introduction indicating the
general purpose of the gathering and the underlying research. The
introduction was notably vague so as not to bias subsequent responses. For
the remainder of the session the moderator rained questions but did not
attempt to confine individuals' responses. All individual were asked to
respond to moat of the questions so as to avoid dominance by one or two
people.
Examples of the types of questions raised were the following:
What does water quality mean to you?
How do you know when the water quality is poor?
What activities do you pursue on the Chesapeake?
Has the water quality gotten worse over time?
What do you think is the moat serious cause ofpollution in the Bay?
Is water quality different in different parts of the Bay?
Initially, we had several questions in mind which we hoped the focus
groups could help us answer. These included the following:
1. What sources of information do people use in forming their perceptions
of water quality in the Chesapeake Bay?
2. What factors affect their interpretation of this information (e.g. past
experience, attitudes), i.e. what is their standard based upon?
3. In what way does the water quality of the Chesapeake affect people;
i.e., in what sense do they lose when water quality deteriorates?
4. Do their perceptions affect their behavior and how quickly can
behavior be expected to change in response to environment changes?
18
-------
Our focus group results suggest the following answers to these questions.
Question 1.
Appearance was by far the most frequently mentioned signal of water
quality deterioration. Whether or not the individual had been exposed to
scientific information on the subject, he was most likely to report that how
the water looked, felt and smelled were his most important indicators. Even
individuals who no longer used the Bay or certain sections of the Bay based
their decisions on the water's appearance during their last visit.
Without exception, individuals used clarity of the water as an indicator.
The degree of transparent y of" the water was taken for granted to be a
measure of quality. A few individuals, particularly those living on the Bay all
year, noted seasonal differences in clarity, but still used this as their first
quality indicator. Other factors which signalled poor water quality were the
nature of the bottom, discoloration of the shoreline, froth on the water?
floating debris (including man-made) and dead fish. Smell was a clear signal
of poor water quality, but odors were not common, and visual appearance was
used as a more discriminating indicator.
The second most common signal of pollution was "guilt by association."
Individuals frequently stated that they deduced that water quality would be
poor in sections close to activities which they reasoned would generate
pollution. Such activities included sewage treatment plants, housing and
industrial developments, marinas and other heavy concentrations of boats, and
farms (particularly with livestock) . These deductions took place in both
groups but were of a slightly different nature. The college-age group was
relatively more concerned with agricultural operation and with contamination
from boat sewage. The older group seemed to consider development -- with
or without sewage treatment --of greatest concern. Some of this difference
can be accounted for by the spatial location and familiarity of the two groups.
The Plum Point group knew local conditions well but were relatively immobile
and had limited experience with the rest of the Bay. In contrast, the college
students were heavy boat users and therefore extremely mobile. They tended
to have personal experience along large portions of the Bay and its
tributaries. Graphic examples ofmanure pond overflows, run-off from pig
farms, etc., were offered. Residents of the Plum Point area would have little
exposure to agricultural runoffs more common to the upper Bay.
Television, radio and newspapers were the next most common external
source of information. Rarely waa specific information about local Bay
conditions gleaned from the media. Instead, these sources create a general
awareness of environmental problems. In large part the inferences about
activities which create pollution were baaed on information gathered from
these secondary sources.
It was clear that at least some individuals were privy to more objective
and scientific information than that available in the public media, although the
distinction between types of information sources was not always made clear.
Many of the college students had taken courses and subscribed to scientific
journals. They were able to draw more sophisticated deductions about links
19
-------
between water quality and surrounding land uses. Also, these individuals
tended to be distrustful of information obtained from the media.
The final signal of pollution also depended on deduction. Individuals
noticed changes in the amount and diversity of wildlife in and around the Bay
and concluded that these changes were the result of water quality changes.
Specifically mentioned were crabs, turtles, and ducks. One individual
associated a decline in the diversity of finfish species with water quality
deterioration, Another individual argued that increases in fish prices were
due to pollution.
Question 2.
Different individuals seemed to interpret similar information in different
ways. Factors which affected their interpretation included their past
experience and general attitudes. When individuals react to the appearance of
something, they must, by definition, be comparing it to a standard. More
often than not, the standard used by these groups was past experience,
although occasionally individuals seemed to be operating with a less personal
standard such as pictures of clean water in mountain lakes. Frequently
individuals compared the appearance of the water to what they (accurately or
not) remembered experiencing as a child. With the exception ofone individual
(of the older group), everyone remembered the Bay water being cleaner when
they were children. This was true of the 18- and 19-year-olds as well as the
50- and 60-year-olds. When questioned, individuals admitted that both
maturity and publicity had raised their level ofconsciousness about water
quality but still insisted that water quality was poorer now than when they
were children. Also, the college studenta noticed some improvement over the
last fewyears --in terms of fewer dead fish and birds, less heavy oil
present, and the cleanup of dumps along the tributaries - although they
thought the water was dirtier now than ten years ago.
Individuals' interpretationa of information were also clear] y affected by
their general attitudes -- level of trust in political and entrepreneurial forces
and confidence in technology. Among the college students, some indicated
distrust for political processes and commercial enterprises to the extent that
they believed everything was polluted, whether or not they could see it,
These individuals stated that they would need hard scientific evidence to be
convinced that improvements had been made. At the other extreme, notably in
the older group, a few individuals indicated a trust in the scientific
community, regulator y processes and technology, fueling that the populace
would be protected from unsafe conditions through cleanup activities. Some
indicated resignation to the trade-off between the environment and
development. In all cases attitudes affected how individuals interpreted the
same sensory and media information.
Question 3.
Individuals perceived themselves to be affected by water quality in a
number of ways. It was clear from the discussions that both groups were
apprehensive about going into water they perceived to be dirty. A distinction
20
-------
was made between wading and swimming, as many indicated that they did not
dare submerge their heads in water they thought to be polluted. Also many
wore shoes when wading to avoid contact with the bottom. While only a few
actually mentioned bacteria and potential illness, most seemed to have health
and safety factors in mind.
It is difficult to separate this apprehension from the general
unpleasantness associated wit h unattractive water. No doubt we are
conditioned to link clarity with cleanliness and cleanliness with health.
Nonetheless, it seems that even if the individuals were convinced there were
no health risks, they would consider themselves hurt by water quality
deterioration. Many mentioned the unpleasant fee/of the bottom and the
sticky film that swimmers feel on their skin after bathing. Others mentioned
that clear water allowed them activities such as seeing living organisms in the
water, activities which were precluded by murky water. Still others who
never entered the water but only walked along it reported the experience
more pleasurable when the water looked cleaner. In fact, some gave up the
activity when the water looked dirty.
What is interesting however is that no one mentioned toxics or heavy
metals. Some were aware of the term nutrients, and most connected this with
turbidity, algal blooms and slimy bottoms. Others emphasized oil spills.
Particularly in the older group, individuals expected that pollution could be
seen. Those individuals seemed most conscious of health risks, yet indicated
they felt safe going swimming on days when the water looked clear. Few
indicated apprehension about health effects from unseen pollutants.
Lasses also accrued to individuals through perceived reductions in angler
and duck hunting success. Many complained of a decline in the quality of
fishing, crabbing and duck hunting. Others complained of the reduced
variety of finfish available in the Bay. Among the college students were some
who professed a concern for the wildlife in situ. That is, some individuals
indicated reduced enjoyment of non-consumptive wildlife uses.
Individuals also indicated a fear of eating fish and shellfish caught in
polluted waters. For many individuals low catch rates were irrelevant because
they did not dare eat fish caught in local waters.
Interestingly, one individual who did not use the Bay for any recreational
activity indicated that he really did not care what happened to the water
quality in the Chesapeake. His only concern was the quality ofhis drinking
water. Here is a real world example of the concept of "weak complementarity."
Weak complementarily is said to characterize an individual's preferences if he
does not care about the quality of a resource that he does not use.
Question 4,
Earlier in this chapter, survey results were shown to support the
empirical relation between behavior and perceptions. However, frequently
inconsistent behavior was observed--some individuals perceived the water
quality to be poor but continued to use it. The focus groups shed some light
on these anomalies.
21
-------
With only a few exceptions, the individuals in both focus groups were
familiar with . the Bay and were recreational users of one sort or another.
Also, with only a few exceptions, the individuals were deeply concerned about
the water quality of the Bay. In most circumstances individuals had not given
up all forms of recreational use, but they had reduced that use and curtailed
some activities altogether,
We can model an individual's behavior as changing in three ways in
response to perceived changes in water quality: he can alter the choice of
whether to participate in an activity, he can alter the sites at which he
recreates, and he can alter the frequency ofrecreation.
The two groups revealed different types ofreactions to water quality
deterioration. The younger, more mobile group stopped going to certain
places which they perceived to be worse. The older, less mobile, group
stopped participating in certain activities which they perceived to be sensitive
to water quality, particularly swimming. They also curtailed their fish and
crab catch to avoid eating contaminated fish. In both groups, there appeared
to be a frequency dimension to individuals' reactions as well. Many who
found the water quality too poor for swimming generally indicated they would
go in on especially hot days or on days when the water looked especially
clear. The latter suggests that the degree of intra-seasonal variation in
pollution and other causes of turbidity will affect the frequency of
participation in a recreational activity. Of course only those who live near
the shore can assess the water clarity before incurring the coats of the
recreational trip.
While there is no firm evidence for this, many individuals seemed to
participate more in recreational activities than they believed wise. In many
cases, it was because they had been participating in these activities for years
and resisted giving them up and because they perceived no suitable
alternative. In contrast, some individuals had curtailed certain activities
because of bad experiences and indicated that it would take very convincing
evidence to bring them back. All of this suggests that the response of
behavior to perceptions may be significant but may also be a delayed
response.
Summary of Focus Group Experience
In summary, we can construct a set of hypotheses about perception
formation. The list would include
10 Individual associate the quality of the water with its appearance --
specifically its clarity and color.
2. Individuals associate the quality with the amount of floating
(man-made) debris and dead organisms.
3. Individuals associate quality with angler success rates.
4. Individuala deduce quality from surrounding land and water uses.
22
-------
5. Individuals infer things about the quality of water from general
publicity about the environment and/or technology change.
6. Individuals learn specifics about quality from scientific publications,
educational experiences.
The more exposure individuals had to information of the sort included in (5)
and (6), the more likely were they to deduce things about water quality from
surrounding activities. Nonetheless, items (1) and (2) in the above list
dominated, irrespective of age, education, etc.
Individuals perceived themselves to be harmed by poor water quality
through a number of routes:
1. The individual's recreational experience is degraded by unpleasant
appearances floating debris, etc.
2. The individual fears health and safety risks.
3. The individual believes poor water quality reduces catch rates and
variety of species.
4. The individual fears eating fish from areas with poor water quality.
It is worthy of note that both unpleasant appearance and poor fishing
conditions harm individual but also serve M signals of poor water quality.
These signals carry with them suspicions of further losses in the form of
health risks from contact with the water or from eating contaminated fish.
Conclusions
In this chapter we set out to explore the relationship between human
activities and the water quality of the Bay. This relationship is important for
the Chesapeake Bay Program for several reasons. First, human use of the Bay
is the ultimate goal of devoting resources to improving the quality ofthe Bay.
Gaining some sense that people change their use of the Bay with changes in
water quality suggests that Bay clean-up strategies can have significant
value. Second, economists' benefit measures of improvements in water quality
are based primarily on changes in behavior. Knowing that households have
some sense ofwater quality and are affected by this sense ofwater quality
when deciding how to allocate their scarce time and resources gives support
to the methodology of benefit measurement.
We have explored the relationships between perceptions and human
activities in two ways. From two surveys, a phone survey of households and
an on-site survey of beach users, the relationship between objective measures
of quality and perceptions of quality and behavior has been examined. The
telephone survey supports the relationships in several ways. Households that
perceive water quality as unacceptable are more likely to quit using the Bay.
The telephone survey also shows an implicit but positive correlation between
the likelihood of quitting and the Bay's "water quality. The user survey also
provided support for the perceptions link. This survey shows positive
23
-------
correlation between measures of fecal coliform at each of nine beaches and the
proportion of households that found each beach unacceptable.
The focus groups provide insight into how people judge the quality of
water and why they change their behavior in response to quality changes. A
variety of sensible motives contribute to behavior changes. People smell, feel
and see the water and its surroundings. They react when they learn about
changes in water quality from newspapers television and other media.
Of particular importance to policy makers is the clear signal that
individuals suffer from water quality deterioration in more than one way.
Many regulation are implicitly based on health standards, yet health effects
are only part of the story. Irrespective of health risks, individuals were
uniformly adament in arguing that recreation in water perceived to be dirty is
less enjoyable. This dimension is not totally independent ofhealth concerns,
however, since dirty water was additionally considered to be a signal for
health risks. The word "risk" is a key one. Whether or not a given state of
water quality does in fact present a health risk, the individual suffers from
the uncertainty associated with not being able to assess the risk himself.
While we do not go into this problem in great depth in this study, it is
important to keep a few things in mind. Uncertainty is ceteris paribus
undesirable, and there are two sources of uncertainty involved here. One is
the uncertainty associated with not knowing what is in the water and whether
it is potentially harmful. The second is the uncertainty associated with the
actual onset of adverse health consequences if indeed the water was
potentially harmful.
The losses described above pertain to water use that involves contact.
There are still more ways in which individuals perceive themselves to be
harmed by poor water quality. The enjoyment associated with any activity
within sight of the Bay is claimed to be diminished if the water appears dirty.
Finally, to the extent that poor water quality reduces fish abundance and
species variability y, sportfishermen see themselves harmed. Finally, even if
fish catches aren't reduced, perceived poor water quality suggests health
risks associated with eating fish catch.
Together the -two sources of information provide support for the
inferences which we draw in the following chapters. Individuals are aware of
water quality, change their behavior in response to water quality changes,
and derive benefits when the quality of the Bay is improved.
24
-------
Chapter 3
Recreational Use of the Bay and Willingness to Pay Estimates
for Improvement to the Bay se a Whole
A variety of methods have been used to analyze the welfare effects of
water quality improvements. In the introduction to this chapter, a brief
description of the two basic approaches is offered to prepare the reader for
the methods used in this and following chapters. A more thorough examination
of one of the methods? contingent market valuation? is offered in Cummings)
Brookshire and Schulze. Bockstael, Hanemann and Strand supply a thorough
examination of the other method, indirect market valuation,
The indirect market approach uses individual behavior in related markets
to infer values ofnon-marksted goods. For the case in question, water
quality, the researcher observes the demand for goods that are related to
water quality, such as recreational tripe. The usefulness of the approach
depends on the responsiveness of behavior toward water quality. If
individuals value good water qualit y, the y will be drawn to goods or activities
associated with high quality water end will be willing to travel farther and
incur greater costs for this improved experience. Their behavior can be
observed, and from this, values deduced. One drawback to this approach is
that assumptions regarding behavior must be made in order to assess values.
This results in untestable restrictions on behavior implicitly or explicitly
imposed in the modelling process.
Contingent market analysis involves the establishment, in the interviewee's
mind, of a fictitious or hypothetical market circumstance. The interviewee is
asked to respond to the circumstance in a hypothetical manner. By
establishing a scenario to explain the respondent's answers, the researcher is
able to deduce characteristics of the respondents preferences.
The "average" willingness to pay or sell is the predominant value obtained
in most contingent valuation exercises. A question or series of questions is
designed to elicit the respondent3 (hypothetical) bid for or against the policy
in question. The approach can be directed very specifically to the good or
quality change to be valued, and thus, in theory, elicit the amount of money
needed to keep the individual at the same level of satisfaction before and
after an event. The questions can be quite specific and may therefore define
precisely the event or policy to be assessed. The disadvantage of the
approach is its hypothetical nature. Rarely is it possible to test the validity
of the response, through observations on behavior. in addition, the specific
valuation problem may be so remote from the respondent's market valuation
experiences as to leave him unable to respond reliably.
Contingent valuation has been deemed a useful technique (see Cummings,
Brookshire and Schulze) provided it is applied to problems which are closely
associated with common market valuation experiences. Car-son and Mitchell
present evidence of stable contingent valuation estimates for national benefits
of clean freshwater in the U. S.
25
-------
It seems reasonable to attempt some contingent valuation for the
Chesapeake Bay problem as long as caution is exercised in the interpretation
of the results. For us, derivation of the contingent values thus obtained is
not intended as an end unto itself, but rather information to support the
results of additional analyses.
Recreational Use of the Bay
During the summer of 1984, a telephone survey ofover 1,000 households
in the Washington, D. C. and Baltimore Statistical Metropolitan Sample Areas
(SMSA's) was conducted. A description of procedures can be found in
Appendix A. Appendix B is a copy of the survey instrument. One objective
of the survey was to provide a complete inventory of beach use by residents
in the Baltimore/Washington SMSA's ( see Figure 2.1 ), which include the District
of Columbia, several counties and incorporated cities in Northern Virginia and
much of central and southern Maryland. Restricting the geographical area in
this way biases the sample of individuals toward urban residents. However,
this area includes a large percentage of the population surrounding the Bay.
In the subsequent discussion, the percentage figures reflect the sample
response rates corrected by sampling weights to define unbiased estimates.
The projected total number of households purported below to participate in
various activities are estimated as the product of these weighted response
rates and the approximately two million households residing in the
Baltimore/Washington SMSA (the 1980 Census reported 1,876,144 households).
On the basis of the telephone survey, 43 percent of the region's
households are estimated to have used or intended to use the Bay for some
recreational activity in 1984. Participation rates varied across the region (see
Table 3.1) with Anne Arundel County having the highest percentage use (69%)
and the District of Columbia the lowest (21%). Of the remaining areas,
Northern Virginia had the next lowest participation rate (37%) and Montgomery
County the next highest rate of participation (48%).
The households used the Bay for a variety of recreational activities.
Swimming/beach use was the most popular, with a projected 740,000 households
participating. The next most popular activity was boating which attracted a
projected 620,000 households. Sightseeing (estimated 586,000 households) and
fishing (estimated 477)000 households) were also very popular. The projected
number of households who used the Bay in conjunction with hunting totalled
only about 45,000. There were an estimated 170,000 households that reported
other uses of the Bay.
As one might expect, households often participate in more than one
activity. For the major use activities of swimming, fishing and boating, Table
3.2 shows the percentage of" respondents who participated in one activity or
more. Roughly speaking, about one-third of the households participated in all
three activities, one-third participated in two ofthe three activities, and the
remaining one-third participated in a single activity. This distinction has
importance for benefit estimation; if any one household's participation were
limited to only one activity, independent behavioral studies of each activity
26
-------
Table 3.1
Participation Rate" in Chesapeake Bay Activities
By Activity and Area, 1984
Northern District of Montgomery
Virginia" Colurbia county
\ Participation in
CB Activity (1984) 22 9 37
\ Participate or'
Intend to Participate 37 21 48
\ Participate CB
Fishing 12 8 16
\ Participate CB
Swiming 10 R 24
\ Participate CB
Boating 17 8 26
\ Participate CB
Hunting 1 0 1
I
Prince George's
end Charles Anne Arundel
Counties County Baltimore6 Others^
34 60 36 36
46 69 45 42
18 33 19 21
21 33 23 23
24 4B 24 33
0 523
aWeighted percentage, representing a random sanple of Baltimore-Washington, B.C. SMSA's
^Includes Fairfax, Arlington, Prince William and Loudon counties and the cities of Alexandria, Fairfax and Falls Church.
"Includes Baltimore City and portions of Howard and Baltimore counties.
dlncludea Carroll and Harford counties and portions of Howard and Baltimore counties.
-------
GO
Table 3.2
Joint Participation Rates3 in Chesapeake Bay Activities
By Activity and Region, 1984
Overall
Northern Virginia
District of Cblunbia
Montgomery County
Prince George's and
Charles counties
Anne Arundel County
Baltimore City
and County
Other Maryland
Fishing
3
3
1
8
3
3
3
One Activity
I Swimingk I
14
15
10
22
17
3
13
Two Activities
Boating
10
11
12
7
10
9
15
Fishing
Swimming
9
6
15
8
7
7
12
4
Fishing
Boating
9
9
5
9
11
10
6
20
Swimming
Boating Fishing
20
15
15
22
19
34
20
18
Three Activities
Swiraning Boating
34
38
54
26
31
33
37
40
tWeighted percentages, representing a random ample of the Baltimore-Washington, B.C. SMSA's.
'-Swiraning includes beach use.
-------
could be aggregated to provide the basis of a total benefit estimation for
improved water quality. Multiple participation and the interdependence among
activities prevents straightforward addition of benefits calculated in demand
studies of individual activities.
While it may be necessary eventually to undertake a comprehensive benefit
analysis of all Bay activities? there is enough current information to shed some
light on the value of the recreational use of the Bay. Independent studies
are useful, if for no other reason than to establish "conditional" relationships
between activities and key factors. This may facilitate future studies by
isolating key factors for which information is critical. Moreover, by analyzing
a series of partial systems, bounds may be established on the total potential
benefits.
Aggregate Willingness to Pay
This portion of the chapter employs the contingent valuation technique to
value improvements in water quality in the Chesapeake Bay. The hypothetical
circumstance posed to survey resondents involves the alteration of the Bay's
water quality from its current condition to an improved condition which, in
the respondent's view, is acceptable for swimming. . Because individuals'
perceptions of water quality are not easily linked to objective measures (see
Chapter 2) and because individuals do not easily understand these scientific
measures, the hypothetical circumstance was framed in terms of the
respondent's acceptability. This limits the specific application of the results,
since there is no simple way to determine at what point clean-up efforts raise
the water quality to an acceptable level for everyone. However, the evidence
presented in Chapter 2 offers some guidance as well as some historical
perspective.
The households responding to the contingent valuation experiment are a
subset of the telephone survey of the Baltimore-Washington SMSA's. Each of
the randomly selected households was asked:
"Do you consider the water quality in the Chesapeake to be
acceptable or unacceptable for swimming and/or other water
activities?"
Of the 959 respondents, over one-half (57 percent) found the water quality
unacceptable. '1'hose who responded that it was unacceptable were asked:
"Would you be willing to pay ($A) in extra state or federal
taxes per year if the water quality were improved so that you
found it acceptable to swim in the Chesapeake?"
The amount of money ($A) was varied randomly from $5 to $50 over the
sample. The percentage of respondents who answered "yes" is shown in Table
3.3.
29
-------
Table 3.3
Percent of People Willing to Pay Additional Taxes
for Acceptable Water Quality for Swimming, by Amount of Tax
Amount of Tax Increase $5 $10 $15 $20 $25 $30 $35 $40 $45 $50
Percent of
Respondents Willing
To Pay Tax Increase 64 66 63 70 58 46 57 47 47 53
If sample sizes were big enough, monotonically decreasing percentages
over the entire range would likely be revealed. Nonetheless, the percentages
are in general declining as the amount of the tax increases. Of those who
were presented ataxof $5, $10, $15, or $20, an average of 66 percent agreed
(hypothetically) to accept the tax burden in exchange for acceptable water
quality. Of those presented a tax of $25, $30, $35, or $40, the average
percentage dropped to 52 percent.
Analysis of Willingness to Pay Responses
Hanemann (1984) describes a method for analyzing a central tendency in
willingness to pay from questions with "yes" or "no" answers. Let the
respondent derive utility from the nonmarket good, water quality, and from
money income (y) which can be used to purchase marketed goods. Also let a
vector (x) of individual characteristics affect his utility. Utility is given by
u i(l ,y;x) when the water quality is acceptable and U. (0,y;x) when it is not.
The functions u, , and u. are not known, and thus are considered stochastic
to the researcher; That is
(3.1) Uj(j, y;x) = v(j,y; x) + i/j j = 0,1
where vj independently and identically distributed random variables with
mean zero.
When offered swimmable water at a tax of $A, the individual will accept
the tax providing that
(3.2) v(l, y-A; x) + vl»v(0, y;x) + i/0
and decline otherwise. In this framework, the individual's response becomes a
30
-------
random variable with probability density
Po = Pr{accept tax in exchange for swimmable water}
= Pr{v(l, y-A; x) - v(0, y;x) •"<> - v^
pt = Pr{not accept tax in exchange for swimmable water} = 1 - po.
Define *) = v0-vl and let ?•»)(•) be the cumulative distribution function of 17.
Then the probability ofaccepting the tax in exchange for swimmable water
equals FTJ(AV) where 4v is the difference between the deterministic portions of
the utility function in the two states (see equation (3.2)}.
At this point any of a number of utility functions, individual
characteristics, and density functions to complete the analysis could be
chosen. Like Hanemann, we chose a logistic cumulation distribution function.
Also, we chose a linear function (see Sellar, Chavas and Stoll for the
limitations of this form) for v(0), such that
(3.3) v(j, y) = aj + 9 y, f > 0
where the arguments of x have been temporarily suppressed. The difference,
Av, is (BJ - aO - PA, which gives a probability model of the form
(3.4) F(q) = n F(-aO + al - /YAi) ~ [1 - F(-ao + a, - #A~)]
its, itso
where So is the set of individuals refusing to pay the tax, and S^ its the set
accepting the tax.
Conceptually, we sought the value A for which
(3.5) u(l, y-A; x) = u(0, y;x).
Combining equations (3.2) and (3.3) produces the result that when (3.5) holds,
A is defined as the following
A = [(a, . a0)(-7,) ]//»•
31
-------
Since ~n is random, so is A. To evaluate A we chose to take its expectation,
assuming «0i « i » and 0 to be constants, which yields
(3.6) E[A] = (at - «0)/0.
Thus, («i - «o)/0 is the expected (or average) tax that would make an
individual just indifferent to paying the tax in exchange for acceptable water
quality and not paying the tax but forgoing good water quality. Now we need
estimates of the parameters a i, «0» and ft to get a value for E[A],
Results ofAnalysis. BY Subgroups of Respondents
In developing the theory it was admitted that individual characteristics
(designated by the vector x) were likely to affect the utility function which in
turn would affect the parameters in (3.6). Some of these characteristics are
strictly idiosyncratic and not worth trying to model, but others may be
associated with identifiable subgroups of the population. Three means of
subdividing the population suggest themselves--by household income, by race
and by Bay user/nonuser. In the sample obtained in 1984 there was sufficient
correlation between race and income to make the separate treatment of these
infeasible. Additionally, it was difficult to subdivide the population by income
because income appears in the data set as a continuous variable and
arbitrarily dividing it into ranges did not prove useful.
After some preliminary logit analysis, a modification of the model shown in
(3.4) was estimated. One modification entailed making the (*1 - «o) depend on
whether someone in the household had used or intended to use the
Chesapeake Bay in 1984. A variable (D j ) was included to reflect use. This
approach allows us to test whether users value the change in the Bay's water
quality more than non-users, ceteria paribus. The other modification involved
making the bid coefficient, f, depend on the racial classification. Because
there is a wide disparity in income between whites and non-whites (average of
$40,000 annually vs. average of $25,000), the marginal utility of income, which
P represents, maY be different for the two groups. Use ofa binary variable
(D2> in conjunction with the tax variable permitted an examination of the
effect ofrace on the marginal utility of income.
The results of the estimation are reported in Table 3.4. The amount of
the tax significantly reduced the probability that a respondent would agree to
pay the annual tax increase. Also significant were the use/intercept
interaction variable and the tax/race interaction variable. Both users and
whitea were more likely to accept the tax increase.
32
-------
Table 3.4
Logistic Model Estimates Related to the Probability a Respondent
Will Accept a Tax Increase
Variable8
Constant (
-------
Table 3.5
Estimates of Utility Parameters by Income Group
Income Class
$0 - $20,000
$20,000- $50,000
$50,000 - $80,000
Over $80,000
Income not
reported
«i
Users
1.282
(1.89)"
1.652
(2.96)
1.695
(.98)
1.157
(2.81)
.533
(1.53)
Non-users
.833
(1.72)
.968
(2.04)
1.471
(2.80)
.543
(1.24)
.200
(.42)
ft
.028
(1.73)
.012
(.81)
.017
(.95)
.013
(.90)
.016
(1.11)
Sample
Size
99
200
101
22
93
Likelihood
Rat io
19.35
11.05
42.33 .
5'?.79
9.68
W-statistic in parenthesis
Table 3.6
Expected Value of Willingness to Pay
for Acceptable Water Quality for Swimming by Income Group and User Group,
1984.
Income Class
o - $20,000
$20,000-$50,000
$50,000-$60,000
over $80,000
not reported
Expected
Average for All
$ 38.54
108.60
95.16
66. '44
22.26
Value of Willingness to Pay
Average for Users Average
$ 415.94
134.25
101.20
89.00
32.64
for Non-users
$29.85
78.48
88.08
41.77
12.25
34
-------
Table 3.7
Estimated Willingness to Pay for Acceptable Water Quality
by Participation and Racial Composition of Household
1984.
Expected Willingness to Pay
Participation Status Racial Compos it ion
White Non-Whit e
User
Non-User
$183.63
(55.12)a
$48.13
(10.25)
$34.16
(10.40)
$ 8.95
(2.53)
a Standard deviation in parentheses
the range are less reliable. The results suggest that a wider range of tax
increases would have yielded more confidence in the estimates' accuracy.
Individual willingness to pay bids for water quality improvements appear to
have a larger range (i.e. take on larger values) than we anticipated when
constructing the survey.
In Table 3.8 the average willingness to pay for each subpopulation is
combined with estimates of the subpopulation size to project total willingness
to pay figures. The values are based on the telephone sample estimate that
57 percent of the population find Chesapeake Bay water quality unacceptable
and on the sample percentages of white users (27%), white non-users (35%),
non-white users (16%) and non-white non-users (21%).
Expected values as well as optimistic and pessimistic values are shown.
The optimistic (pessimistic) value is derived using the expected value of
willingness to pay plus (minus) one standard deviation. On the basis of these
estimates, we could argue a reasonable range of willingness to pay values of
$60 million to slightly over $100 million. Care must be exercised when
considering the standard deviation, as it is computed as an approximation and
is not associated with the normal distribution. The values shown, however,
represent an "order ofmagnitude" contingent valuation estimate of willingness
to pay for improved water quality.
35
-------
Table 3.8
Estimated Aggregate8Willingness to Pay for Water Quality Acceptables
for Swimming, by Classification and Scenarios
1984
Wingness to Pay
"Average*'
Scenario
"Optimistic"0
"Pessimistic"^
Thousand
User
White
Non-white
55,838
6,164
72,595
8,020
39,081
4,271
Non-user
White
Non-whit e
Aggregate
83,612
106,976
60,275
'Baltimore-Washington SMSA population
bfiased on expected willingness to pay
°Based on expected willingness to pay plus one standard deviation
^Based on expected willingness to pay minus one standard deviation
Regional Comparisons
Stretching the data somewhat further, one can also examine geographical
patterns of responses. The logistic model wae re-estimated using sub-samples
grouped by region: the Southeast region (Prince George's County, Charles
County, Anne Arundel County and the District of Columbia) , the Western region
(Northern Virginia, Montgomery County) and the Northern region (Baltimore
36
-------
City and County, Howard County, and Harford County) . The sub-samples
represent groups, each of which exhibits reasonable internal homogeneity, for
which we have at least one-hundred and fifty responses. Even with these
conditions, however, the statistical results are less significant than the earlier
ones because ofthe smaller sample size.
The results suggest regional similarities and differences OlcfcLe 3.9!. Some
consistency is evident as signs on all coefficients are the same for all regions.
Thus an increase in the hypothetical tax decreases the probability of
acceptance of the tax associated with water quality improvement. Additionally
the effect of participation and race on willingness to pay for the improvement
is consistent across all regions.
Table 3.9
Logistic Model Estimates Related to the Probability a Respondent
Will Accept a Tax Increase to Improve Chesapeake Bay Water Quality,
by Geographic Area
Variable
Constant
DI " constant
Amount of Tax
D2 . Tax
Southeast •
.334
(.94)d
.78
(2.36)
-.050
(3.33)
.041
(3.15)
West b
.71
(.46)
1.02
(2.49)
-.070
(3.04)
.060
(3.00)
North0
.12
(.30)
1.67
(4.77)
-.023
(1.77)
.015
(1.36)
Chi-squared for
Likelihood ratio
36.5
37.2
48.6
aDist. of Columbia and Counties of Prince George's, Charles and Anne Arundel
^Nort hem Virginia and Montgomery County
^Baltimore City and Count iee of Baitimore, Harford and Howard
^t-ratio in parentheses
There are, however, systematic differences across regions. Users from the
Northern region are willing to pay on average substantially more than those
from the southeast or western regions. The figures for nonusers are less
disparate across regions, with those for the West region somewhat larger. The
estimated willingness to pay figures are presented in Table 3.10.
37
-------
Table 3.10
Estimated Willingness to Pay for Acceptable Water Quality by Region,
Participation, and Racial Composition of Household, 1984.
Household Characteristic
White, User
Non-White, User
Non-White, Non-user
White, Non-user
Southeast
$124
22
7
3'?
Region
West
$133
25
10
55
North
$224
77
5
15
Existence Value
In the preceding contingent valuation experiment we present non-zero
willingness to pay estimates for non-users as well as users. There are a
number of reasons why non-users may be willing to pay for improved water
quality. One of these reasons has been labelled existence value by
non-market benefit analysts (Krutilla) and stems from early experiences
applying benefit cost analysis to water resources projects. Individuals who
never use a resource either directly or indirectly and never intend to uae it
may still be willing to pay to improve ita quality or assure its existence.
Formal studies of existence value are limited, but some empirical evidence
exists. Fisher and Raucher (1984) suggest that nonuse benefita (including
both option value and existence value) are some fraction of the use value of
water quality changes. Other research (e.g., Walsh et al., 1985; Schulze et al.,
1983 ) suggests that existence value may be greater than use value, and
sometimes substantially so.
Existence value is a frequently cited concept in the literature, and several
studies have attempted to derive explicit estimates of existence value
associated with water quality (Mitchell and Carson, 1981; Cronin, 1982; Walsh et
al., 1978; Desvousges et al., 1983). Nonetheless, no consensus exists on the
models which underlie the measurement. Behaviorally based methods of
welfare measurement are unsatisfactory because, by definition, existence value
is unconnected with behavior. Suspicion surrounds contingent valuation
estimates of existence value because these estimates are even less susceptible
to proof or disproof than contingent valuation estimates of use values. Even
moretothe point, the success of a contingent valuation approach depends on
well defined questions. Without a clear idea of the "motivations behind
existence value, properly focused questions are difficult to define.
38
-------
The Existence Value Experiment
In this section we present some preliminary results of an experiment
designed to shed some light on the motives behind existence value for the
Chesapeake Bay. The sample frame was derived from the phone survey
described above. The households contacted randomly by phone were asked if
they would complete an additional mail survey. Of the 1,044 contacted, 741
agreed to fill out and return a brief mail questionnaire regarding water
quality in the Chesapeake Bay, and of these 741 households, 282 actually
returned the questionnaires. Because only about 70 percent of those
contacted agreed to receive the mail questionnaire, and only 38 percent of
those who agreed actually returned these questionnaires, these results should
not be taken as representative of the population sampled but as useful for
gaining preliminary y insigh ts into willingness to pay motives.
The 282 respondents were grouped as users or non-users. Users were
defined as all respondents who currently use the Bay or thought they might
do so in the future. Respondents who felt certain that they would not use
the Bay for recreation at any time in the future were defined as non-users.
Non-users accounted for 16.3 percent of the respondents.
Respondents were asked to consider a series ofsituations concerning
public beaches surrounding the Chesapeake Bay. They were asked to assume
that water quality at these beaches had fallen below a level acceptable for
swimming, They were told that a project could be undertaken that would
clean the beaches so that a water quality level acceptable for swimming was
achieved and maintained. The respondents were then asked the following
question under four scenarios:
"Would you prefer that the clean-up project be undertaken?"
Scenario 1, No additional information.
Scenario 2. Access to the beaches by the public is permanently denied
so that even if clean, the beaches will not be used.
Scenario 3. If the project is undertaken, taxes would be raised so much
that nearly everyone prefers that the project is not
undertaken. These taxes would be paid by individuals other
than the respondent.
Scenario 4. If the project is not undertaken, funds would instead be
used to improve hospital services in selected communities
surrounding the Bay. Of all the people who care, half want
the beaches cleaned, and half want improved hospital
services, The respondent himself would never need to visit
any of the improved hospitals.
The proportion of "yes" responses for users and non-users under each
scenario is given in Table 3.11.
39
-------
Table 3.11
Summary Results of Contingent Valuation Experiment
on Existence Value
Scenario
Number
1
2
3
4
Proportion of
Yes Responses:
Users *
.96
.70
.71
.49
Standard
Error of
Difference
.032
.032
.035
Proportion of
Yes Responses:
Non-users^
.83
.69 .
.67
.37
Standard
Error of
Difference
.088
.088
.091
aThe number of users in the sample of respondents is 236.
blhe number ofnonusers in the sample ofrespondents is 46.
GThis is the standard error of the difference between the proportion in
Scenario 1 and the proportion in the given scenario.
Interpretation of the Results
In order to interpret the responses reported in Table 3.11, it is necessary
first to consider the 'potential motives for existence value. Two broad motives
may be discerned: intrinsic and altruistic. Existence value based on intrinsic
motives stems from a concern about the state of the world. Concern about the
order of things may cause people to suffer simply by learning about pollution
incidents. What has been called the "environmental ethic" is closely linked
with the intrinsic motive.
Of concern here is the second ofthe two motives: altruism. People can
gain value from the enhanced wellbeing ofothers (individualistic altruism).
An extensive discussion of these altruistic motivations can be found in
Madariaga and McConnell (1987) .
Responses to the question under Scenario 1 are used as a control to be
compared with responses under Scenarios 2 through 4, where Scenario 1 is
purposely ambiguous about 'project costs. As expected, moat respondents
preferred that the project be undertaken under Scenario 1. Interpreting
non-user responses of "yes" as evidence ofexistence value, the relatively
high number of non-users giving positive responses is consistent with the
results of previous studies that have found evidence of existence value.
40
-------
With access to beaches denied under Scenario 2, the number of "yes"
responses to the question predictably declined. Since the number of non-user
responses of l*yes" declined when access was denied, it appears that existence
value, at least to some individuals, is related to others' use. Thus, altruism
may , be one motive that underlies existence value. However, even with access
denied, most respondenta preferred that the project be undertaken. This may
reflect the presence of a number of motivational including an environmental
ethic. Finally, it is interesting to note the closeness of user and non-user
group responses under Scenario 2. Since with access denied there can be no
users, "yes" responses from the user group will also indicate positive
existence value. In this scenario, the proportion ofusers and non-users
exhibiting existence value was nearly identical.
Scenario 3 is similar to Scenario 2 in that both attempt to eliminate
altruistic motives. In this scenario, the Bay can be used after the cleanup,
but other individuals will be forced to pay more than what the improved water
quality is worth to them. The similarity ofproportions in Scenarios 2 and 3
supports the notion that the Chesapeake resource is valued for its own sake.
In Scenario 2, about 70 percent of the people support the project despite the
fact that there is no use and hence no direct use value. In Scenario 3,
roughly the same proportion supports the project even though there is no net
value to the users.
Under Scenario 4 the number of "yes" responses fell dramatically
compared with the responses under Scenario 1. Since less than half of the
non-users preferred that the cleanup project be undertaken, it appears that
the improved hospital services are on average at least as valuable as clean
water in the Bay.
The individuals were instructed that they would not need the services of
the hospital, themselves, so it is tempting to label their value for the
improved hospital services as existence value. However, the entire value of
the hospital services may be due to altruistic motives while individuals appear
to have motives beyond altruism for Chesapeake water quality improvements.
Conclusions
The underlying motives for existence value matter to the proper design
and interpretation of contingent valuation experiments. The preliminary
results concerning existence value associated with the Chesapeake Bay suggest
some ambiguity about its motivation. People are willing to pay for water
quality improvements in the Bay, but how much they are willing to pay
depends on the specific nature of the opportunities foregone by doing so.
Among other considerations, these suggest attention should be paid to the
methods for financing the cleanup ofthe Bay.
41
-------
Chapter 4
The Effect of Chesapeake Bay Water Quality on Beach Use
The previous chapter contains a range of benefits from improved
Chesapeake Bay water quality, based on a contingent valuation experiment.
Although there is substantial evidence to suggest that the responses to the
hypothetical questions were not random but rather associated with households'
use of the Bay and racial/income strata, problems still exist with the approach.
Follow-up questioning revealed households did not consider alternative uses of
tax increases (e.g. improvements in other public goods such as hospitals,
roads, etc.) . Moreover, the subjective nature of the water quality measure-
ment used in the contingent valuation question does not lend itself easily to
policy analysis, based as it is on objective (scientific) measures of water
quality. Finally, the values represent aggregate values, indistinguishable on
the basis of type ofrecreation or geographic location of pollution. Knowledge
of user group and geographical impacts of programs can provide a depth and
richness of understanding important in the political process.
The remaining chapters are devoted to providing analyses of the observed
behavior of households based on data gathered in previous studies which are
specific to different recreational activities. The analyses use cross-sectional
information on households to model beach use, boating and recreational
fishing. Once demand functions are estimated, benefits from access and from
changes in water quality are assessed for each of these activities.
This chapter contains a cross-sectional analysis ofbeach use on the
western shore of Maryland. It draws from the random telephone survey of
the Baltimore-Washington SMSA's and a stratified random survey of twelve
public beaches on Maryland's western shore. As such, the analysis is not
comprehensive of all beach use in Maryland but rather the use of the public
areas in one portion of the Bay by the citizens in the surrounding environs
ofthe two large metropolitan areas closest to the Bay.
A number of approaches to estimating recreational response to environ-
mental quality changes have evolved. Many of these depend first on the
estimation of demand for recreational activities which are closely linked to the
environmental resource in question. The recreational demand models currently
in use have grown out ofthe union of neoclassical demand theory and the
travel cost model proposed by Retelling and employed extensively by recrea-
tional economists for the paat several decades. The principal contribution of
the travel cost model is found in the use of the travel cost to the recreational
sita as the principal component in constructing a "price" for the recreational
commodity. The simple travel cost model can be derived from a neoclassical
utility maximization framework, as can more complex models which incorporate
added dimensions to the problem (see Bockstael, Strand and Hanemann, 1986).
One particularly important modification of the simple model is the
introduction of quality characteristics of recreational sites (see Volume I of
42
-------
this report for the theory). If the recreational demand model is to be used to
estimate the value ofenvironmental quality improvements then individuals'
behavioral responses to changes in quality must be modelled. This requires
observing behavior in the context of differing levels of environmental quality,
which can generally be done only by examining recreation behavior at a given
point in time over a group of sites which vary in quality.
The procedure can lead to a number of specific methods of analysis (see
Kling, Bockstael and Strand, 1985) , each imposing a different set of
restrictions/assumptions on recreational behavior. While there is no concensus
regarding the "correct" model, the two most prominent, models in the literature
can be categorized as the (modified) neoclassical model and the discrete choice
model. The neoclassical model has the form of the traditional demand system,
with quantities being a function of prices and water quality. The model is
modified in some way to facilitate the inclusion of water quality parameters
which do not tend to vary for a given site over the population, Additionally,
demands are generally treated independently. A common approach is the
varying parameter model (VPM), as put forth by Smitht Desvousges and
McGivney (1986) . Here, independent single-site models of recreational trip
demand are estimated, and the estimates of the intercept and price coefficient
are correlated with the site's water quality. Then, in policy analysis, changes
in water quality change the intercept and slope of the demand curve, thereby
influencing quantity consumed and the welfare derived from recreational
activities.
The discrete choice model (DCM) has also taken many forma (e. g., Caulkins;
Morey and Rowe) but the form employed by Bockatael, Hanemann and Strand is
representative. In this model, the individual is viewed as having a number of
choice occasions upon which to select a site. The selection of site is discrete
in the sense that only one site is chosen per choice occasion. Site
characteristics such as travel cost, water quality and facilities are used to
explain the choice of a site on any given occasion. A composite "value"
reflecting the desirability of available choices is computed from the discrete
choice estimation, and this is used with other factors to estimate the number
of choice occasions.
Although both models are behaviorally based, there are advantages and
disadvantages associated with both. The varying parameter model starts from
the assumption that the demand functions for trips to sites are interior
solutions to a utility maximization process. The discrete choice model,
however, starts from the viewpoint that, on any given occasion, an individual
chooses among a finite set of alternative sites. Neither approach is perfectly
satisfactory. In the DC model, the link between the number of choice
occasions and the site selection per choice occasion is ad hoc. With the VP
model, the demand for any one site does not adequately reflect the alterna-
tives. Additionally, the fact that individuals do not visit all sites is incon-
sistent with the implicit theory and must be handled econometrically. Kling
has employed Monte Carlo studies to examine the performance of these models.
Not surprisingly, her results suggest that the VP model excels when most
recreationalists tend to visit almost all alternative sites in a season, and the
DC model excels when most tend to visit one or only a few sites in a season.
43
-------
Most recreational data sets will be characterized by something between the two
extremes, however neither model has an obvious advantage, and no tractable
model is perfectly consistent with this situation.
In this chapter both the varying parameter and discrete choice models are
applied to western shore beach use in Maryland. In subsequent chapters only
the varying parameters model will be applied since neither the boating nor
fishing data can support the data intensive discrete choice model. The data
on beach use at western shore beaches is relatively rich, however. Both
types of models will utilize this same data set of Chesapeake Bay beach users
in the subsequent analysis. The results will be a range of values which
suggest orders of magnitude for welfare measures ofhypothetical changes in
water quality.
The Survey and the Data
This section is devoted to a description of the survey of Chesapeake
Beach Use conducted in 1984. Unlike the data used in analyses ofboating
and fishing in Chapters 5 and 6, the data used in this chapter were collected
during an earlier budget period of this cooperative agreement. Great care
was taken with the sampling frame to improve confidence in the results.
Because the survey itself is important to the project, the content and
procedures are described extensively in Appendix C. A copy of the survey
instrument can be found in Appendix D.
From May 26, 1984 to August 19, 1984, Research Triangle Institute (RTI)
interviewed individuals on the western shore beaches in Maryland. The study
population consisted of all residents of the Baltimore and Washington, D. C.,
SMSA'S, age 14 or older, that used these beaches for recreation in 1984. More
specifically, the population was limited to recreational users of the following 12
beaches:
Strata
Beach Geographic Size
1. Sandy Point north large
2. Point Lookout south large
3. Fort Small north small
4. Miami north small
5. Rocky Point north small
6. Elm' s Beach south small
7. Bay Ridge south large
8. Kurtz north small
9. Breezy Point south small
10. Rod & Reel south small
11. Morgantown south small
12. North Beach south small
44
-------
Four hundred and eight individuals were interviewed at the beach to
learn of their recreational patterns and perception of water quality at these
beaches. These individuals were randomly selected from sample beaches and
days. The sampling design can be described as a two-stage stratified sample
in which a probability y sample of beaches and days was selected , and a random
systematic sample of persons was interviewed at each sample site (day-beach
combination) .
The User Intercept Survey Questionnaire was designed to record and
collect the following:
Frequency of visits made to beaches on the western shore of the
Chesapeake
Activities that the respondent (and his/her family) participated in
when visiting beaches
Activities not participated in and the reason why they were not
Cast related to a typical trip to each beach that the respondent had
visited since January 1, 1984
The respondent's perception of the quality ofthe beach and the
beach facilities at each beach with which he/she was familiar
Factors that influenced a respondent's decision to visit or not visit a
beach
The respondent's willingness to continue to visit the sample site if
costs related to the use of the beach were to rise.
In addition, a series of demographic questions was included to enable analysts
to establish profiles of beach users.
The Data
Household Trips
Respondents were asked, on-site, how many trips they had taken in 1984
prior to the interview and how many they intended to take during the rest of
1984. Follow-up telephone interviews at the end of the seaeon obtained
complete 1984 trip information for 251 ofthe 408 beach users interviewed.
For the remaining households, information was obtained solely on-site.
To assure consistency in our trip measurement, the end of the season
information was compared with in-season response so that a correction factor
could be applied to households with only "in-season" information." Using data
from the largest beach (Sand y Point), a regression of end-of-season trips (xe )
on reported plus intended trips during the season (xr) yielded:
45
-------
(4.1) X = -686 + -632xr + e Ra • -89 (n = 148)
6 (1..40) (35.00)
where the t-statistics are in parentheses. Equation (4.1) was used to predict
total trips to a site from on-site information for households that did not
receive follow-up inteviews. The combination of a fairly small constant term
and a coefficient on Xr which is less than one suggests that households tend
to report intentions in excess of trips later realized.
Access Costs and Time Costs
Previous studies (e.g. Bockstael, Hanemann, and Strand, 1986) have
considered travel costs to include the person's (or household's) monetary
costs of travel as well as the opportunity costs of their time. Distance to a
site, transformed into transportation costs is a feature common to all visitors.
However, those individuals who forego income in order to take time to recreate
incur monetary expenses in excess of transportation costs. For these
individualist these costs can be measured as the foregone wage rate times the
time spent accessing the activity.
For households without employed persona or with employed persons with
fixed work schedules, there is no direct loss of income incurred when
recreation is undertaken. The opportunity cost of recreation time for these
individuals is the value of foregone alternative activities. Unfortunately,
opportunity costs will vary over individuals in ways which are not observable.
The only observable factor related to the total opportunity cost of the
recreation experience will be the time spent traveling and recreating. Even
this measurement is troublesome, however, since the on-site portion of this
time also measures the amount of the recreational good consumed. To avoid
many of these complications we employ only round-trip travel time as a
surrogate for opportunity costs in these cases.
In addition to these access costs, most western shore beaches have an
admittance fee which must be added to the other costs of traveling to the
site. Often the fee will vary depending on the day of the week and size of
party.
Water Quality
The Chessie System environmental quality data, maintained by EPA's
Chesapeake Bay program, were' used to construct the water quality measures.
Turbidity, bacteria counts, total suspended solids, total nitrogen, total
phosphorous, and total chlorophyll A are among the potential indicators of
water quality to which beach demand might be sensitive, Since these
exhibited a high degree of collinearity, two variables were extracted from the
data set to use in this analysis: total nitrogen and total phosphorus. A good
case can be made for using these variables. Studies of the Bay conducted by
the U. S. Environmental Protection Agency indicate that perhaps the most
significant problem facing Bay restoration and protection efforts is nutrient
46
-------
over-enrichment of Bay waters. Excessive nutrient levels may be the partial
cause of decreased submerged aquatic vegetation, which in turn has adverse
effects on the food chain and on the habitat for many fish species. Further,
over-enrichment leads to lower dissolved oxygen levels which have additional
adverse effects on fish stocks, degrading the appearance of the water as well.
High collinearity y between nitrogen and phosphorus readings prevented
separate inclusion of both variables in the analysis. The product of nitrogen
and phosphorus was used to avoid this problem and to capture the interactive
nature of these nutrients.
In each case, mean monthly water quality levels from April through
September 19?7 were calculated for areas of twelve counties contiguous to the
Bay, The summer months were chosen because they represent the peak of
recreational activity. Complete data over regions of the Bay were available for
only some years, 1977 being the closest to the survey year. The relative
water quality readings across the Bay are unlikely to be considerably
different between the two years, even if the absolute readings are different.
Additionally, individuals * decisions are unlikely to be related solely to water
quality in the current year, but will be based on a cumulative learning
process which includes past observations as well. Consequently, there is no
obvious correct choice, and the errors associated with using data from any
one year are unclear.
Other Variables
Additional factors are known to influence recreation activity, including the
ownership ofcertain types of household capital equipment. Boats, recreational
vehicles and swimming pools are the types ofcapital equipment which may
affect the use ofbeaches on the western shore. Some of these beaches have
boat-launch facilities, some camp sites, while others offer good swimming
possibilities. Years living in the area, previous recreational history, family
size and participation are some other factors which may be important.
The Varying Parameter Model
To formalize the model of behavior, the individual is assumed to maximize a
constrained utility function which is a function of number of trips taken to
each of nquality-differentiated sites, the quality characteristics of each site,
and a Hicksian good. Thus
(4.2) max u(x, q,z) s.t. px + z = y
where x is an n-dimensional vector of trips to the n sites, p is a
corresponding vector of costs of accessing the sites, q is a matrix of variables
qi-i, i = l,...,n and i = l,...,m, where qij is the level of the jth quality
characteristic at the f h site, z is the Hie sian good, and y is income. To
simplify notation in this section, we will assume that there is only one quality
47
-------
characteristic, and thus m = 1, and q represents the vector of values of the
one quality characteristic across sites.
Problem (4.2) defines n demand functions, each of which may be a function
of all nprices, n quality levels, and income:
(4.3) xi = g. (p, q/y) i = 1 ,...,n.
This model cannot be estimated with cross-section data. Imagine having
observations on S individuals who visit site i. Own price (pj) and substitute
prices ( Pi, k * i) will typically vary across individuals if they come to site i
from different geographical areas. However, there will be no variation in the
quality characteristic at site i (q ) ) across the S individuals, nor will there by
any variation across individual in the characteristics of other sites (q k, k *
i). With no variation in the q /a across observations, their coefficients cannot
be estimated, and nothing can be learned about behavioral response to quality
changes.
There are several methods for resolving the dilemma presented above.
Some of them build on the model presented in (4.3) (these are described in
Kling, Bockstael and Strand, 1985), while others rely on discrete choice models
(see Bockstael, Hanemann and Strand, 1986) . The method used in this section,
the varying parameters model, falls in the former category and follows similar
methods applied by Vaughan and Russell (1982) ; Smith, Desvousges and
McGivney (1983); and Smith and Desvousges (1985).
One way to motivate the varying parameters model is to consider a simple
linear form such as:
n
(4.4) x, = 00, + £ f
k=l
but to further assume that the parameters in site demand functions are
deterministic functions of the quality characteristics. For example, the 0*s
might be linear functions of the q's:
70 +
(4.5) 0itk~aok + «ik
-------
The model in (4.4) and (4.5) implies that variations in demand parameters
across sites (i.e., variations in the ft> i ls, ftt ( ls, etc. ) correspond to variations
in own- site attributes (qj ) and subatitute-site attributes (q^, j * i) .
Specifically, the above model implies a demand for trips to site i of the
following form:
(4.6) xf ^(70 +
K — 1
+ (6
Even though the model can be collapsed into one expression as in (4.6),
the estimation procedure usually involves two steps: the regression of trips
to each site on prices and income (e.g., n separate regressions) and the
regression of the coefficients from the" first n- regressions on the quality
characteristics of the sites. The second step requires the application of
generalized least squares because of the properties of the error structure
implicit in the estimation of (4.5) which must use estimated parameters (f 's) in
place of the true ft.
The first -stage estimation procedure is further complicated by the need to
correct for a censored sample bias. Moat consumer demand problems analyzed
with household data encounter this problem. A random sample of households
will reveal a. certain (often substantial) number of households that do not
consume the good in question and thus have zero as the value of their
dependent variable. In the sample, there will therefore be many observations
concentrated around zero. Neither omitting the zero observations, nor
including them in an OLS regression, will produce unbiased estimates.
Tobin analyzed this problem in 1958 and produced the first of several
approaches to handling the problem. His approach applied to the first stage
of our varying parameters model characterizes the problem in the following
way:
(4.7) x, = 0o
x( = 0 otherwise.
While e may be distributed as a normal, x will not be. The estimation of the
0's therefore requires maximum likelihood techniques, where the likelihood
function is given by
(4.8) L =
49
-------
where zf is the right-hand side of (4.4),
-------
Table 4.1
Average Values of Regression Variables for Visitors by Beach
vn
Beach
Tripe
Beach (#/yr)
Sandy Point
Fort Smallwood
Chesapeake Beach
Rod & Reel
Bay Ridge
Point Lookout
Rocky Point
Porter's New Beach
Miami
Morgantown
8.06
5.83
2.87
6.42
7.33
3.85
10.20
2.92
5.16
7.71
Access
coats
(S)
15.98
8.66
18.79
22.80
15.75
36.42
9.07
9.50
11.81
26.09
Access
Time
(hr)
1.29
1.11
1.45
1*58
1.14
2.73
.93
.72
.86
1.22
Substitute Beach
Access
coats
(S)
12.09
6.38
11.30
13.61
10.84
13.10
6.43
5.81
4.82
9.80
Access
Time
(hr)
.84
.69
.84
.87
.84
1.03
.55
.49
.59
1.02
Boat
(%)
15
14
17
13
12
17
13
19
19
17
Ownership
Recreation
Vehicle
(%)
18
20
28
33
28
30
12
11
17
35
Swimming
Pool
(%)
18
14
17
25
12
12
20
11
17
09
-------
Table 4.2
Tobit Estimates for Beach Demand Model, by Beach"
Beach
Beach
Sandy Paint
Fort Smallwood
Rod & Reel
Rocky Point
Chesapeake Beach
VI
to
Porter's New
Beach
Point Lookout.
Miami
Day Ridge
Constant
8.17
(2.83)b
.16
(.05)
10.44
(--2. 18)
10.29
(2.04)
-3.96
(-1.89)
-.70
( .31)
3.49
(-2.72)
-2.20
(-1.45)
-6.96
( 1.16)
Access
Costs
-.35
(-4.07)
-.53
(-2.86)
MO
(-.84)
-.47
( 1.45)
.18
(-2. 19)
-.29
(-2.21)
-.05
(-5.62)
.09
(-1.35)
-.78
(-4.90)
Access
Time
-4.85
(-3.61)
-4.24
(-2.58)
-1.51
(-1.28)
-5.63
(-2.38)
-1.19
(-1.76)
-1.28
(-1.28)
-1.72
( 4.72)
-1.27
(-1.18)
-9.63
(-3.50)
Substitute Beach
Access Access
costs Time
.24 2.47
(2.86) (1.15)
.34
(1.14)
.29
(1.25)
.19
(1.80)
.31
(1.10)
.12 4.55
(3.35) (5.41)
.83 7.40
(3.19) (1.96)
Ownership
Rec.
Boat Veh.
3.23
(2.58)
1.54
(1.32)
2.19 2.98
(1.69) (2.50)
4.37
(2.46)
-6.19 7.55
( -1.00) (1.50)
Swim.
Pool
3.55
(1.36)
2.04
( 1.31)
-1.76
(-1.21)
-5.67
(-1.13)
a
14.85
(57.59)
9.52
(11.61)
9.72
(5.47)
12.41
(19.00)
6.16
(10.00)
3.43
(5. 15)
5.96
(15.14)
7.42
(10. OG)
18.06
(17.56)
Nonlimit/
Limit
Observations
243/139
41/198
22/201
87/66
46/272
25/118
82/262
50/121
61/292
coefficients were significantly different from zero for Morgantown site.
Dt-ratiOS in parentheses
-------
The estimated coefficients on own-price (travel cost) were all ofthe
expected sign) and most were statistically significant from zero. Beaches for
which a reasonably large on-site sample was obtained yielded the most
significant estimates. Small sample effects of multicollinearity among the price
and time variables likely caused the large standard errors for Miami, Rod and
Reel, etc. In some instances, the multicollinearity was sufficiently troublesome
that only the own-price and own-time variables were considered. Obtaining
results for as many beaches as possible was critical because the sample size
in the second-stage estimation equals the number of beaches in the first
stage.
The results of the second-stage estimation, i.e. the estimation of equation
(4.5), were obtained from a weighted least squares procedure in which the
weights were l/up, ., the inverse of the standard error of the own-price
coefficient for each beach. The estimated equations are:
(4.10) ftij = -.0308- .00020 TNP.
(-.04) (-2.22) ]
= -2.66- .0016 TNPj
(-1.10) (-.001)
where TNP is the water quality variable defined earlier, and the values in
parentheses are t-ratios.
The results show no significant relationship between water quality and the
intercepts of the beach-use demand equations but a significant relationship
between water quality and the coefficients on travel cost. The poorer the
water quality (i.e. the higher the level of TNP), the larger the negative
response ofbeach users to travel costs. This results in a pivoting inward of
the demand curve as water quality deteriorates and a pivoting outward with
improvements. The results are in accordance with the proposition that poor
water quality lowers beach users willingness to pay for access to beaches.
Estimated Benefit Changes
The analysis above describes the behavioral response of the average
western shore beach user to the change in water quality. From this
information and information on the number of users of the beaches, we are
able to determine some estimates of benefits of hypothetical improvements in
water quality to the average user of each beach. We are also able to expand
to the total population of beach users.
Three hypothetical changes in the environmental variables are considered,
a 10 percent and a 20 percent decrease (environmental improvement) in the
environmental (pollution) variables, and a 20 percent increase (environmental
degradation) . Since we will want to assess the effects of the change, we will
want to calculate consumer surplus before and after the change. The formula
53
-------
for individual i's consumer surplus from site j is given by the following when
the demand function is linear
(4.11) Csi;j -
where x> • is its demand for trips to j and Pj i is the coefficient on cost of
access in the jfc" site demand function. For a given hypothetical change in
water quality at beach j, the weighted average change in consumer surplus
over the sample is calculated:
N.
(4.12) ACS = Sl(x,J(qJ))a/(-VJ1(qJ)) -(X|j(qJ))V(-20 j l (qj)) ]*k,/Nj
where k, is the weight, qj and qj are the levels of water quality before and
after the change, and the notation xl j (q. ) and P\ i (q^ ) implies that both
demand and the coefficient on travel cost are functions of the level of water
quality. Nd is the size of the sample of beach users used to estimate the
tobit equation for beach j . The sample includes all 408 observations minus
those for which information about beach j was unavailable.
Calculating consumer surplus for hypothetical environmental circumstances
(equation 4.12) thus requires values for X" = x(q°), 0°(q°), x1 = xfq1 ), and fl
= 0(ql)« The first step in assessing the hypothetical changes is to use the
results of model (4.10) to predict ft j i (qj ) , that is to predict the new travel
cost coefficient given the hypothetical change in water quality. The
coefficients 9 j \ (qj) and fj t (qj) are then used to determine values for demand,
i.e. xj and xj.
Prediction of the demand for trips is complicated because the demand
function was initially estimated using a Tobit procedure. Recall the model
underlying the Tobit,
x* = P'z + e e ~ N(0, 0
and
x = 0 otherwise.
Given the underlying model, the systematic portion of (4.4) cannot be used as
the expected value IOCUIY of x, The Tobit predicting equation given below
adjusts for the censored nature of the dependent variable:
54
-------
(4.13) B(x4j) = *
where z is a vector of explanatory variables and * and * are the density and
cumulative distribution functions for the standard normal, respectively, The
first term represents the conditional expectation of trips given that the
person participates times the probability that the person participates. The
second term corrects for non-normality because of potential truncation.
There are two ways of obtaining the "before" and "after" x's for the
consumer surplus functions. One way is to use the predicting equation (4.13)
to calculate both i° and xl values. The second method is to accept the
observed x as x° and then to adjust this x by x * - x° to reflect the
hypothetical change in water quality to obtain the estimated x '. (See
Bockstael and Strand, 1986, for details ofthe two approaches. )
Because there is no clear theoretical reason to choose one approach over
the other, we calculate the results both ways. Both methods "use formula
(4.11) to calculate the change in average consumer surplus. However, Method
A calculates trip values as
(4. 14)
and
M'ZI
(4.15) *»j = *|-V-^ h'z«j + ^*
where the z , , are the explanatory variables in the jth beach's regression (see
Table 4.2). Method B calculates the demand for trips in the following way:
xf j = observed value of x < j
and
where x} j and X? j are defined in (4. 14) and (4. 15).
Tables 4.3 - 4.5 summarize the average beach users benefits and losses
from the hypothetical changes in the nitrogen and phosphorus concentrations
in the Chesapeake Bay. The first and fourth columns in each table represent
the base line average consumer surplus over the entire sample of beach users
55
-------
Table 4.3
Annual Benefits per Beach User from a 20 Percent Decrease
in Pollutant, by Beach
1984
Calculation Method Aa
Beach
Sandy Point
Fort Smallwood
Chesapeake Beach
Rod & Reel Club
Porter's New Beach
Rocky Point
Point Lookout
Bay Ridge
Miami Beach
Consumer Surplus
Before After
133.94
.82
36.32
10.32
5.95
80.38
15.86
178.18
5.38
169.03
5.17
43.88
16.19
8.45
89.53
22.61
204.76
'10.27
Benefits
35.09
4.35
7.56
5.87
2.50
9.15
6.75
26.58
4.89
Calculation Method Bb
Consumer
Before
342.04
57.69
57.89
259.81
12.20
179.65
315.27
171.64
220.68
Surplus
After Benefits
379.33
73.13
60.77
284.08
12.34
191.02
415.06
178.98
304.99
37.06
15.44
2.88
24.27
1.14
11.34
99.79
7.34
84.31
• With Method A, the average consumer surplus for a change in quality at beach
j is taken over a sample which includes all beech users whether or not they
visited beach j.
"With Method B, the average consumer surplus for a change in quality at beach
j is taken over a sample which includes only users of beach j .
-------
Table 4.4
Annual Benefits per Beach User from a 10 Percent Decrease
in Pollutant, by Beach
1984
Calculation Method Aa
Beach
Sandy Point
Fort Smallwood
Chesapeake Beach
Rod & Reel Club
Porter's New Beach
Rocky Point
Point Lookout
Bay Ridge
Miami Beach
Consumer
Before
133.94
.82
36.32
10.32
5.95
80.38
15.86
178.18
5.38
Surplus
After
150.39
1.50
39.96
13.00
7.12
84.82
18.73
191.08
7.34
Benefits
16.45
.68
3.64
2.68
1.17
4.44
2.87
12.90
1.96
Calculation Method B^
Consumer
Before
342.04
57.69
57.88
259.81
12.20
179.65
315.27
171.46
220.68
Surplus
After
363.35
69.28
61.11
277.73
13.55
186.63
363.61
176.55
261.16
Benefits
21.31
11.59
3.22
17.92
1.35
6.98
48.34
5.09
40.48
awith Method A, the average consumer surplus for a change in quality at beach
j is taken over a sample which includes all beach users whether or not they
visited beach j .
Method B, the average consumer surplus for a change in quality at beach
j is taken over a sample-which includes only users of beach j .
57
-------
Table 4.5
Annual Losses per Beach User from a 2.0 Percent Increase
in Pollutant, by Beach
1984
Calculation Method Aa
Beach
Sandy Point
Fort Smallwood
Chesapeake Beach
Rod & Reel Club
Porter's New Beach
Rocky Point
Point Lookout
Bay Ridge
Miami Beach
Consumer
Before
133.94
.82
36.32
10.32
5.95
80.38
15.86
178.18
5.38
Surplus
After
106.54
.29
29.81
6.25
4.05
72.26
11.92
154.56
3.06
Losses
(27.40)
(.53)
(6.51)
(4.07)
(1.90)
(8.12)
(3.94)
(23.62)
(2.32)
Calculation Method Bb
Consumer
Before
342.04
57.69
57.88
259.81
12.20
179.65
315.27
171.64
220.68
Surplus
After
311.26
47.63
55.27
239.35
11.24
166.81
253.41
164.55
172.41
Losses
(30.78)
(10.06)
(2.62)
(20.46)
(.96)
(12.84)
(61.86)
(7.09)
(48.27)
'With Method A, the average consumer surplus for a change in quality at beach
j is taken over a sample which includes all beach users whether or not they"
visited beach j.
"with Method B, the average consumer surplus for a change in quality at beach
j is taken over a sample which includes only users of beach j.
58
-------
for use of each beach, calculated using each of the two methods mentioned
above. The second and fifth columns show the average consumer surplus per
beach user in the sample following a water quality change at each beach. The
third and sixth columns represent the change in surplus for the average
beach user associated with a water quality change at each beach.
The method of calculation makes a good deal ofdifference for some
beaches, especially Point Lookout, Miami Beach and Rod and Reel. Recalling
the econometric results in Table 4.2, the estimated demand equations for these
three are price inelastic relative to other beaches; that is, the absolute values
of their price coefficients are quite small. When demand is very inelastic, big
differences are likely between the mean consumer surplus and the consumer
surplus associated with the mean number of tripe (see Bockstael and Strand) .
The average consumer surplus values are expanded to the entire
Baltimore-Washington SMSA's in Table 4,6. This was accomplished by knowing
that the 1980 number ofregional households was 1,977,000 (census of the U. S.,
1980), by determining from a contemporaneous phone survey that 47 percent
of the regional population used western shore beaches" and by knowing the
percentage of western shore beach users who used each beach. Large
aggregate benefits are associated with Sandy Point (in both methods of
calculation) because ofthe very large number of households that visit that
beach. Whereas 21 percent of the population used western shore beaches,
over half used Sandy Point. When expanding to households, Sandy Point has
nearly twice as many users as any other beach.
The Discrete/Continuous Choice Model
The utility maximizing model in (4.2) and the resulting demand functions in
(4.3) are an apt description ofthe individual's decision problem only if he
chooses positive values for all x^ (i.e., if he is at interior solutions in all the
markets). It is not an adequate description if corner solutions arise (i.e., x i=
o). The discrete choice model is appropriate when an individual chooses one
from a finite set of alternatives, by comparing the available alternatives. The
discrete choice model presented here is amended to include a component which
describes the demand for trips as well as the discrete choice among trips on
any choice occasion.
The Choice Among Sites
The first part of the model involves the estimation of the household's
choice among sites. It will be important here to capture those elements which
vary over sites. McFadden (1976) provides a utility theoretic framework for
employing the multinominal logit model which is applicable to a discrete choice
problem of this sort. For further discussion of its application to recreation
demand, see Bockstael, Hanemann and Strand (1986) .
59
-------
Table 4.6
Aggregate Benefits/Losses to Users from Changes in Chesapeake Bay
Water Quality, by Beach
1984
Calculation Method A
Calculation Method B
Change Change
Improvement Degradation Improvement Degradation
Beach
Sandy Point
Fort Smallwood
Chesapeake Beach
Rod & Reel Club
Porter's New Beach
Rocky Point
Point Lookout
Bay Ridge
Miami Beach
20%
14,064
1,744
3,038
2,356
1,006
3,673
2,708
10,667
1,963
10%
6,602
275
1,462
1,075
468
1,781
1,153
5,176
788
20% 20%
. . . Thousands$
(11,001) 9,
(212) 1,
(2,612)
(1,632) 1,
(750)
(3,258)
(1,577) 12,
(9,484)
(931) 3,
967
576
680
316
52
923
484
823
975
10% ,
4,704
651
329
626
24
449
5,375
397
1,674
20%
(8,009)
(781)
(597)
(1,089)
(40)
(824)
(7,520)
(710)
(2,255)
60
-------
Suppose we call \^ a latent variable denoting the level of indirect utility
associated with the ith alternative. The observed variable Yi has the property
that
Yi = 1 if Vf = max(V*,VJ, . . .,V|)
Y = 0 otherwise.
Indirect utility associated with the ith alternative is some function of Zj_, a
vector of attributes of the ith alternative so that V£ = V^z^) + ej;. The
random component is generally attributed to the systematic? but unmeasurable,
variation in tastes and omitted variables. Thus, each household has a level of
error which, in a sense, remains with it over time. If the t's are
independently and identical y distributed with type I extreme value
distribution (Weibull), then it is well known that
Prob (Yi=l I z) =
(see Maddala 1983; McFadden, 1973; Domencich and McFadden, 1975) . The
likelihood function for the sample is
where gi = 1 if i is chosen, gi = 0 otherwise.
The multinominal logit has a property which in some circumstances is useful
but in others is unrealistic. The model presented above implicitly assumes
independence of irrelevant alternatives, i.e. the relative odds of choosing any
pair of alternatives remains constant no matter what happens in the remainder
of the choice set. Thus, this model allows for no specific pattern of
correlation among the errors associated with the alternatives; it denies—and
in fact is violated by--any particular similarities within groups ofalternatives.
McFadden (1978) has shown that a more general nested logit model
specifically incorporating varying correlations among the errors associated
with the alternatives can also be derived from a stochastic utility maximization
framework (see also Maddala, 1983) . If the e's have a generalized extreme
value distribution then a pattern of correlation among the choices can be
allowed. McFadden defines a probabilistic choice model
P, =
ev"
61
-------
here G1 is the patial of G with respect to the ith argument and
has certain properties which imply that
• '
= exp{-G(e
"£
is a multivariate extreme value distribution. When G(e * ..... e N) is defined
s £evi then the model reduces to the ordinary multinomial logit (MNL)
described above. However when
G(Y)
M
= La.
m=l
._
icS
where there are M subsets of the N alternatives and 0 »*m < 1, then a
general pattern of dependence among the alternatives is allowed. The
parameters, *m» can be interpreted as an index of the similarity within groups.
Suppose we were to classify the alternatives into these M groups where $m
denotes the set of alternatives in group m, and we were interested in the
probability of choosing some alternative i. Then
P =
where
r
2m
if i
Otherwise
and
l-«n
The above GEV model is useful in many applied discrete choice problems.
Frequently, alternatives group themselves in obvious patterns of
ability. If they do, it is both convenient and appropriate to estimate the GEV
model. It is appropriate because the results of an ordinary MNL will violate
62
-------
the independence of irrelevant alternatives asumption if such a pattern
actually exists. It is convenient because it reduces the number ofalternatives
included at each stage.
Let us make the estimation process explicit. In the problem at hand,
individuals are choosing among ten beaches. Two of these beaches are
qualitatively different. They are state parks, larger and providing more
services than the local beaches. Now we can view the choice problem as a
two-level nested one: the choice between state park or local beach (m = 1, 2)
and the choice among beaches within each group. Consider a redefinition of
Vj
i:
v. = »'Z.
im i
where the Z's denote attributes associated with all sites and the W's are
attributes associated with the state park and local beach choice. Also let us
assume that sm is identical within all groups and equal to 6
Now define a variable, Imi in the following way:
•
Then the probabilities above can be rewritten as
(4. 17)
and
(4.18) Pm =
The variable Im is sometimes termed an inclusive value (see McFadden, 1978)
and serves as an index of the relative value of the alternatives included in
subgroup m.
As expressed in (4.17) and (4.18), the probabilities of interest can be
estimated using MNL proced urea. First, the Pi | m are estimated with M
independent applications of the multinominal logit. Note that at this stage e is
not recoverable, but can be estimated only up to a scale factor of 1-4. From
63
-------
the results of (4.17), the inclusive prices (4.16) are calculated and
incorporated as variables in the second level of estimation (4.18). Here the fa
and the 6 are estimated.
A S outside the unit interval is inconsistent with the underlying utility
theoretic model and suggests misspecification (see McFadden) . The parameter
S is an index of similarity of alternatives within groups not present across
groups. A value of one for 15 indicates that alternatives within a group are
perfect substitutes. Thus, all relevant choice involves choice among groups.
A value of zero for <* implies there is no special similarity of alternatives
within groups and thus no particular gain from using a nested GEV model.
Two-step estimation, i.e. the estimation of (4.17) and (4.18) independently,
is not necessarily efficient. Amemiya (1973) explores this property of the
model and presents a correction factor. However, even Amemiya suggests that
the cost in computational complexity is probably not worth the gains. We
consider McFadden's estimation method adequate and use it to estimate a GEV
model in the next section.
Estimation of the Discrete Choice Among Beaches
The two-tiered discrete choice model considers the individual choosing
between two categories of sites (state park and local beach sites) and then
choosing among beaches within the desired category. The state park beaches
are located at Sandy Point (adjacent to the Chesapeake Bay Bridge) and Point
Lookout (at the mouth of the Potomac), whereas the local beaches are defined
to include Fort Smallwood, Bay Ridge, Kurtz's Pleasure Beach, Miami Beach,
Morgantown Beach, Porter's New Beach, Rod and Reel Club Beach, North Beach
and Chesapeake Beach.
In estimating the model, however, the decision among sites within each
category is dealt with first. In assessing the available sites within a
category on a choice occasion, the household chooses on the basis of certain
household attributes in combination with specific site characteristics. These
are denoted Z> ^and are defined for one model as:
Zn = access costs in $ to site i, calculated using distance (d ^ from the
household's origin to the site1 (Zx i= 1.088 + .049*^ - .000074d? )
plus the entrance fee plus the wages lost from traveling if the
individual had directly foregone income to visit the site;
Zal = access time (in minutes) to site i, calculated using distance from the
household'a origin to the sites(Za i= .7 + ,02d^ );
Z3 j = water pollution index for site i (see description page 18);
1 Exact formula was determined by regressing reported costs against distance.
2 Exact formula was determined by * regressing- reported travel time against
distance.
64
-------
z., = the availability of recreational vehicle facilities at site i (0 if not
available, 1 if available) times whether the household owned a
recreational vehicle (0 if not owned, 1 if owned);
Zsi = the availability of fishing facilities at site i times whether the
household owned fishing equipment;
Z*, = the availability of boat launch facilities at site i times whether the
household owned a boat.
The first stage of the estimation is reported in Table 4.7. The results
indicate that relatively large monetary and time costs negatively influence the
probability of choosing a beach. Water pollution also has a negative influence
as does fishing facilities. Presumably, fishing activity draws the household
members away from the beach. Boat facilities and recreational vehicle facilities
improve the probability that someone owning a boat or RV will attend beaches
with facilities for that equipment.
The second tier of the discrete decision involves whether individuals
select a state park (with many activities) or a local beach. The factors
hypothesized to be important in deciding to visit a state park were thought to
be the years the household had visited western shore beaches (WI), whether
the intercepted household had more than one family member in the party ( W2 ),
the size of the group intercepted (w, ) and the inclusive value (y derived
from the first-stage estimation. People with a larger history of beach use in
the area would be more likely to learn of the smaller beaches and hence be
less likely to use the state parks. On the other hand, the state parks usually
offer a greater variety of activities, and thus families and large parties might
be more likely to attend them.
The results of the estimations are presented in Table 4.8. The hypotheses
about the choice between state parks and local beaches were not rejected.
Signs of coefficients were as expected and coefficients statistically significant.
The estimated coefficient on the inclusive value term is .152 yielding an
estimate of .848 for S. This is significantly different from zero suggesting
that there are gains from using the nested model.. There is considerably more
similarity among beaches within the two categories than across the categories.
Had the nested model not been used, the independence ofirrelevant
alternative assumption would certainly have been violated. The estimate of $
is also significantly different from one, suggesting that beaches within groups,
although similar, are not perfect substitutes,
65
-------
Table 4.7
Logit Regression for Selection Among Sites
Variable
Access Access Water Recreational Fishing Boat
cost Time Pollution Vehicle Activity Facilities
Estimate Z-^ %2i z'( Zl i %i Z6i
Coefficient
(t-statistic)
-.072
(-5.30)
-*75
(-8.63)
-.00037
(-4.00)
1.06
(1.93)
-2.09
(-5.28)
1.14
(2.10)
= 311.2
Table 4.8
Logit Analysis for Selection Between State Parks and Local Beaches
Variable
Estimate
Coefficient
(t-statistic)
Inclusive
Value
:m
.152
(9.26)
Years Attending
Western Shore Beaches
w,
-.019
(-8.34)
Family
Members
W2
.261
(4.85)
Party
Size
w,
.024
( 12. 79)
Chi-squared = 28.07
66
-------
The Number of Trips Decision
There is no operational model that generally treats utility maximization
with non-negativity constraints (Bockstael et al., 1986, Chapter 8),. There
exists no utility theoretic means of linking the discrete choice model of site
choice on each choice occasion to a continuous choice model of demand for
trips (i.e., demand for choice occasions. )
When one decides to use a nonclassical continuous demand function for the
demand for recreational trips (irrespective of site), a problem immediately
arises in determining the appropriate choice of explanatory variables. Since
costs and quality vary across sites and since individuals are observed
choosing more than one site in a season, which site's price and quality should
be included?
The approach taken here is to consider the number of trips to western
shore beaches as a function of a number of explanatory variables and an
inclusive value type variable calculated from the second stage ofthe discrete
choice model. This was originally suggested by Hanemann (1978) and was used
in Bockstael et al. (1986) . Thus when water quality changes, the inclusive
value changes and influences the number of trips. In this sense, "the discrete
and continuous decisions are linked , although not in a utility theoretic way.
The discrete/continuous choice model has the advantage ofemphasizing the
substitutability ofsites but does appear to underestimate the response of
demand for trips to changes in cost and quality at one or more sites.
Based on the results of the discrete choice estimations, a new inclusive
value (In ) which includes the factors in the choice among sites and the choice
between state parks and local beaches is calculated. This value, along with
the individual's income (INC, income or full income if at interior in the labor
market ), discretional y time available (DT, if at corner in the labor market) and
the number of trips to western shore beaches in the previous year (x^ ) , is
used to estimate the 1984 total number of trips per household to western
shore beaches (xt) .
The higher an individual's inclusive value, the more attractive are his
beach alternatives (e.g. good beaches are cheaper to get to) and the more
trips he is likely to take. Additionally, beach use habits (as reflected in
previous trips) would likely lead to more trips. Whether income and
discretionary time positively or negatively affect the number of trips depends
on whether a day trip to western shore beaches is a normal (positive effect)
or inferior (negative effect)good.
The results ofan ordinary least squares regression are given in Table
4.9. The expected signs occur, and the results indicate a western shore trip
is an inferior good, both with respect to income and time. The predictive
powers of the equation are especially good considering the cross-sectional
nature of the data.
67
-------
Table 4.9
ordinary Least Squares Regression for "Choice Occasions" or Trips
Variable
constant
Estimate
Coefficient .442
(t-statistic)
Inclusive
Value
JN
4.06
(2.92)
Lagged
Trips
xt-l
.516
(8.70)
Income
INC
-.028
(-2.79)
Discretionary
Time
DT
-.981
(-4.80)
F-value (4,253) = 36.30
Estimated Benefit Chancres
The ultimate purpose of the modelling effort is to estimate the benefits
associated with improvements in water quality. Formulas for deriving welfare
measures in the context of discrete choice models of random utility
maximization have been developed by Hanemann (1982, 1984). It is generally
the compensating and/or equivalent variation of the quality change which is
taken as a useful measure of benefits. Selecting the compensating variation
(C) , this measure can be defined by the following expression:
v(p
°
y) = v(p'
y - c)
where again v is the indirect utility function, p and q are vectors of site
prices and qualities, and y is income.
The compensating variation is now defined by
v(p°,q°,y;O = v(p°,ql>y -C; O,
where « is random, and as a result C is now a random variable. Depending on
how one chooses to take account of this randomness, three different measures
of compensating variation can be defined. In the case of GEV models the
median value of C coincides with the C which equates the expected values of
the indirect utility functions (Hanemann, 1978) . It is this measure which we
calculate in the subsequent illustration.
In our problem, using the previous notation,
E e
-------
where J s is the set of state park sites and J 1 is the set of local beach sites;
where v = *'Z\m + ^'Wm; Zjm are factors which vary over sites; and Wm are
factors which vary between state park and local beach sites. Thus
, VM
G e , . . . .e = £ t
I J S=i
where
I. = in
£ e
its
Is
Then the expected value of the indirect utility function equals
v(w°, z°, y) = lnG(eV? evfi) + k,
where k is a constant.
Now consider a change in quality which causes w° and z° to change to W1
and z ^ The compensating variation measure (C') defined above is given by
v(W>,z<>,y) =v(w»,z»,jrC')'
or
in
= In
There is no closed-form solution for compensating variation in this case, but
Hanemann (1982) shows that the compensating variation per choice occasion of
this change can be approximated as:
(14) ACS =
jlj
69
-------
where the set J includes the two cases: state and local beaches, Y is the
element of the • vector which serves as the price coefficient, and vj = V w9 +
(1-Bythe
number of choice occasions estimated in the continuous choice model.
These equations were used to estimate the benefits from hypothetical
water quality changes. To be consistent with the varying parameters model
estimates, we considered a 20 percent reduction and a 20 percent increase in
water pollution. The values associated with the changes are $1.08 per trip
and $4,70 per household user of western shore beaches. Given that 20
percent of the households used western shore beaches (about 401,000
households), the total gains from a 20 percent improvement in water quality
were estimated to be nearly $2 million annually. The estimated loss for a 20
percent degradation was approximately the same.
Discussion
Reiterating, the purpose for our work was to offer benefit estimates based
on different methods so as to provide a range of reasonable values. The two
models derive from two different conceptualizations of the recreationalists1
decisions. The continuous, neoclassical model (represented here by the
varying parameters model) is strictly correct only if interior solutions
characterize demand for each site, with all individuals attending all sites,
Another drawback of this model is that, because of the econometric functions
estimated, total benefits cannot legitimately be added across sites. This sort
of aggregation provides upwardly biased results.
The discrete/continuous choice model, on the other hand, begins by
emphasizing the corner-solution nature of the decision on each choice
occasion. Thus, the substitutability among sites receives special attention.
The decision about number of trips per season is not well integrated into the
estimation process. These models tend to provide low estimates of aggregate
benefits because the effect of water quality improvements on demand for trips
is not well accounted for by the ad hoc inclusive value variables in the trips
equation.
The estimated benefit change resulting from changes in Chesapeake Bay
water quality at the western shore beaches is presented in Table 4.10 for the
two models. " Predictably, the varying parameter model offers the largest
change.
Table 4.10
Comparison of Benefits Based on a Varying Parameter Model
and Discrete/Continuous Choice
Change
Model 20 Percent Improvement 20 Percent Degradation
Varying Parameter "" "" ' ' (in thousands)
upper bound $26,160 - $25,839
Discrete/Centinuous Choice $ 1,885 - $1,884
70
-------
Chapter 5
Recreational Boating and the Benefits of Improved Water Quality
Boating is an especially important part of any study of Chesapeake Bay
recreation. As the second most popular recreational activity in our telephone
sample of Bay users (second only to beach use) ,34 percent of the area's
households participated in boating activities on the Chesapeake in 1984. This
represents nearly three-quarters of the houeeholdo who used the Bay for
recreational activities. Its importance is further supported by the large
number of registered boats in Maryland (as many as 134,000 in 1981) and by
the fact that of the approximately 15 million Person-trips taken on boats in
Maryland waters in 1979, 90 percent (or 13.5 million) were taken on estuarine
waters of the Bay or its tributaries (Harmon and Associates, 1983).
In this chapter we examine the behavior of Chesapeake boatetrs and
estimate the value to boaters from improved water quality. Since no new
survey could be initiated for this purpose, the data upon which the analysis
rests are drawn from a 1983 boat owners survey which waa made available by
the Sea Grant Program and the Department of Recreation at the University of
Maryland.
A Profile of Boaters and Boat Owners
The Boat Owners Survey
In 1983 a survey of boaters was sponsored by the University of Maryland
Sea Grant Program and the Maryland Coastal Zone Management Program. It
consisted of a mail survey of 2515 registered boat owners in Maryland. The
design of the sample provided equal representation to owners of boats kept in
slips and owners who trailered their boats. The questionnaire, which sought a
variety of information about the household and ita boating activities, achieved
a response rate of approximately 70 percent.
The boat owners' survey provides different but complementary information
to the telephone survey conducted by RTI and described in Chapter 3, as it
samples a different population and uses a different sampling scheme. The Sea
Grant survey draws only from the population of registered boat owners. From
the telephone survey, which is a random sample of the population, we can
identify not only those who own boats but also those who uae the Bay for
boating whether they own a boat or not. The telephone survey provides
information about non-boaters, as well. It does not, however, provide detailed
information about boating behavior. Consequently, it is the boat owner
survey which will provide most of the data for analysis.
Boaters and Boat Owner Characteristics
Information from both th boat owner survey and the random telephone
survey helps describe boating in Maryland. For example, a comparison of
71
-------
those households in the telephone survey who reported that they used the
Bay for boating with those who did not revealed no significant differences in
family size or years lired in the area but did suggest considerable differences
in income and race (ace Table 5.1). The average of boaters' incomes waa
significantly higher than the non-boaters' average, and a significantly higher
percentage of boaters were white. Interestingly, this significant difference in
income appears in the Prince Georges/Anne Arundel/Calvert Counties area
(southeastern region) but not in Northern Virginia/Montgomery Count y
(western region) or in the Baltimore area (northern region) . Likewise, when
broken down by subarea the difference in racial composition appears evident
in the southeastern and western subareas but not in the northern subarea.
Some information can be extracted from the telephone survey about boat
ownership as well. Once again average family size and years in the area were
not significant/ Different over the two groups, but average income and the
racial composition of the sample were (see Table 5.2).
The design ofthe boat owners survey permits a distinction to be drawn
between individuala who trailer their boats and those who keep their boats in
the water during the season (either at marinas, docks or moorings) . The
distinction between individual in these groups is important to establish, since
the decisions they face are quite different and their behavior must be
analyzed separately. Additionally, we shall see that profiles of both boats and
owners differ somewhat between the two groups. The sampling design was
stratified to contact approximately equal numbers from the two groups. Of
those who returned questionaires, 718 trailered their boats and 788 kept their
boata in the water during the season.
Some interesting features of these two groups are presented in Table 6.3.
By far the most common type of boat in the trailered boat sample is*
runabout. For obvious reasons, the sanple of trailered boats contains very
few with cabins (4 percent) . It also contains few sailboats (only 6 percent),
but this is not a representative figure, since sailboats which do not use
auxiliary motors are not required to register in Maryland and thus would not
be part of the population sampled. Of boats kept in the water, runabouts (at
33 percent) remain the single moat common class and sailboats (at 31 percent)
represent a close second. Once again sailboats with no auxiliary power are
likely to be under-repreaented in the sample. However, this distortion will
affect the trailered group more, as the boats kept in the water are larger and
more likely to have auxiliary engines. Combining cabin cruisers and cruising
sailboats, half the boats in the non-trailered sample are cruising boats
presumably outfitted for more than one-day trips.
Aa the difference in the type of boat inthe two groups suggests, the
average size and the average value of boats kept in the water are
significantly larger than the averages for trailered boats. Additionally, the
average income of boat owners inthe two groups is significantly different,
with trailered boat owners having on average lower incomes.
Returning to the telephone survey which provides data on boaters and not
just boat owners, we can learn something about the geographical distribution
-------
Table 5.1
Average Characteristics of Boaters and Non-Boaters
in BaMmore/Washington SMSA, 1984
from random telephone survey
Average Family Size
Average Years Lived in Area
Average Household Income8
By Area:
Northern MSL. /Montgomery Cty., Hd.
Prince George's, Anne Arundel,
Calvert Ctya. , Md.»
Baltimore County
Percent Whitea
By Area:
Northern Va . /Montgomery Cty. , Hd.*
Prince Georges, Anne Arundel,
Calvert Ctya. , Md.a
Baltimore County
Boaters
3.31
29
$46,858
$50,576
50, 2s8
41,824
88*
97%"
83*
87*
Non-boaters
3.47
26
$37,063
$41,471
28,083
38,211
74%
77*
64%
80*
aMeans of two samples are significantly different at 99% level
source : Telephone Survey, Research Triangle Institute, 1984
Table 5.2
Average Characteristics of Boat Owners and Non-Boat Owners
in Balttmore/Washington SMSA, 1984
Average E&mly Size
Average Years Lived in Area
Average Household Income3
Percent Whitea
Boat owners
3.67
28.6
$56,511
94
Non-boat Owners
3.27
28.2
$40,931
81
of two samples are significantly different at 99% level.
Source: Telephone Survey, Research Triangle Institute, 1984.
73
-------
Table 5.3
Characteristics of Boats and Boat Owners
from Boat Owners' Survey, 1983
Total number responding to
question of where boat kept
By boat type
Runabout
Cabin Cruiser
Cruising Sail
Day Sail
Workboat
Houseboat
Rowboat
Other
NA
Used boats for swimming
at least sometimes
Used boats for swimming
usually or always
Used boats for swimming
always
Used boats for fishing
at least sometimes
Used boats for fishing
usually or always
Used boats for fishing
always
Average Income of Owner'3
Average Current Boat Value'3
Average Boat Length"
Trailered Boats
718
445 (63%)
23 (3%)
4 (1%)
32 (5%)
124 (17%)
1 (<1%)
47 (7%)
35 (5%)
7
360 (51%)a
133 (19%)a
41 (6%)a
656 (94%)a
502 (72%)a
302 (43%)a
$38,000
$14,000
16 feet
Boats Kept in Water
788
251 (33%)
182 (24%)
199 (26%)
38 (5%)
38 (5%)
13 (2%)
25 (3%)
22 (3%)
20
555 (73%)a
235 (16%)a
41 (5%)a
582 (76%)a
290 (38%)a
107 (14%)'
$51,000
$25,000
23 feet
aNumbers in parentheses are percent of those answering question in each
. stratified sample who gave this response.
"Means of two stratified samples are significantly different at 99 percent level.
Source: Maryland Boat Owners Survey, 1983
74
-------
of households interested in this activity. The geographical distribution of
boaters within the state tells us something about the importance of thie
activity to different population subgroups. From the telephone survey it is
clear that boating is an important activity to Maryland residents throughout
the state; it is the first or second moat popular recreational activity on the
Bay in each of the geographical tubareaa of the study. However, as might be
expected, boaters are moat commonly residents of counties contiguous to the
Bay. A good example is the large proportion of Anne Arundel residents who
participate in boating. Extrapolating from our survey suggests that almost
half of the households in Anne Arundel County had at least one member who
went boating on the Bay in 1984.
Table 5.4 crntains residence data gleaned from the boat owners survey for
boat owners, by trailered and nontrailered classes. The values shown in this
table indicate the distribution of residence counties among those who
responded to this question in the boat owners survey. Of those who reported
residence, 64 percent ofeach group lived in counties contiguous to the Bay.
Approximately 60 percent of the respondents came from the four most
populated counties in the state - Baltimore, Anne Arundel, Montgomery and
Prince Georges. A final interesting feature of the sample is that about 5
percent of the respondents lived out-of-state even though they registered
their boats in Maryland.
The fact that boating is often the vehicle for other Bay recreational
activities makes the analysis of boating both important and complex. The
incidence of multiple activities is critical because participation in the
complementary activities of swimming and fishing may make boaters more
sensitive to water quality. Additionally, the overlap of activities complicates
benefit estimation when benefits are aggregated over activities.
From the boat owners' survey we can learn something about the
importance ofthese multiple activity trips. Table 5.5 reports frequencies of
responses to questions regarding these complementary activities. One striking
feature of these answers is that fishing is extremely important to those
registered boat owners who trailer their boats. In fact almosta]] (94 percent)
of the trailered boat owners who responded to the question indicated that
they fished at least occasionally from their boats, and almost half (43 percent)
claimed to fish on every trip. Fishing was less important among the
non-trailered boat owners, although three-quarters of them indicated that they
fished from their boats at least Occasionally. About the same percentage of
this group indicated they sometimes ueed their boats for swimming. Fewer,
about half, of the trailered boat owners sometimes used their boats for
swimming.
75
-------
Table 5.4
Number of Trailered Boats and Boats Kept in the Water,
By Residence, 1983.
Trailered Boats
Residence
Baltimore
Anne Arundel
Montgomery
Prince George's
Calvert
St. Mary's
Charles
Lower Western Shore -
Total
Cecil
Harford
Kent
Upper Bay -
Total
Dorcester
Queen Anne
Somerset
Talbot
Wicomico
Eastern Shore -
Total
Carol ine
Worcester
Carrel
Allegheny
Frederick
Garrett
Howard
Washington
Other - Total
Pennsylvania
Virginia
District of Cblvnbia
Not Identified
TOTAL
17
25
23
12
27
7
8
13
5
9
21
7
10
11
0
16
0?
19
20
162 (24%)a
99 (15%)
64 (9%)
71 (11%)
(3%)
(4%)
(3%)
65 (10%)
(2%)
(4%)
(1%)
46 (7%)
(1%)
(2X)
(1%)
(1%)
(3%)
56 (8%)
(1%)
(1%)
(2%)
(2%)
(3%)
(3%)
83 (12%)
13 (2%)
10 (1%)
6 (1%)
43
718
Boats
14
25
7
11
20
5
9
12
2
30
9
2
9
4
3
7
5
14
6
Kept in
(2%)
(3%)
(1%)
(1%)
(3%)
(1%)
(1%)
(2%)
(<1%)
(4%)
(1%)
(<1%)
(1%)
(1%)
(<1%)
(1%)
(1%)
(2%)
(1%)
144
184
60
28
46
36
62
50
67
36
18
57
788
Water
(20%)
( 25%)
(8%)
(4%)
(6%)
(5%)
(8%)
(8%)
(9%)
(5%)
(2%)
aNumbers in parentheses represent percent of those answering questions in each
stratified sample who gave this response.
Source: Maryland Boat Owners Survey, 1983
76
-------
Table 5.5
Percent of Boaters Who Fish or Swim While Boating
Boat Owners'Survey, 1983
Occasionally Usually
Always
At Least
Occasionally3
Trailered boats:
Fish while boating
Swim while boating
Non-trailered boats:
Fish while boating
Swim while boating
22%
33%
38%
42%
29%
13%
24%
25%
43%
6%
14%
5%
94%
52x
76%
72%
• Total of other columns
Source: Maryland Boat Owners Survey, 1883
The importance of Water Quality to Boaters
It is useful to consider the qualitative evidence that exists to support the
notion that water quality does in fact matter to boaters. Some evidence to
this effeot can be found in the 1983 boat owners survey. Tables 5.6 and 5.7
present a compilation of responses to a series of questions about factors
important in the selection of boating areas. As can be aeon from these tables,
boat owners who trailered their boats considered water quality to be the
moat important factor in choosing boating areas Water quality waaconsidered
"very important" or at least "moderately important" by the non-trailered boat
owners more often than any other factor except water depth. The latter is
often a physical constraint for the larger boats found in marinas. Comparing
the two subgroups, i.e. those who considered water quality "moderately" or
"very" important and those who did not, it is interesting to note that the
former had on average significantly higher incomes and more valuable boats.
The Behavior of Boat Owners Who Trailer Their boats
The General Model
We are interested in modelling two types of decisions that owners of
trailered boats make in a season. One of these is the commonly modelled
economic decision of how many trips the individual takes. The second haa to
do with the location to which the boat owner takes his boat. This subgroup
of boat owners is far more flexible in the short run than those who keep their
boats in the water during the season, because on a day to day basis they can
77
-------
Table 5.6 Factors Cited Most Important
in the Selection of a Boating Area in Maryland, 1983
Percent Response from718Boat Owners Who Trailer Their Boats
Factor
Water
Quality
Water
Depth
Natural
Beauty
Easy
Access
Lack of
Congestion
Not
Important
11.1
16.3
15.7
13.8
19.5
Slightly
Important
9.3
13.2
17.5
20.1
16.7
Moderately
Important
33.0
29.5
34.0
35.7
32.0
Very
Important
42.2
36.4
28.1
26.7
27.4
No
Response
3.1
3.3
4.6
3.8
4.3
Source: Maryland Boat Owners Survey, 1983
Table 5.7 Factors Cited Moat Important
in the Selection of a Boating Area in Maryland, 1983
Responses from 788 Boat Owners Who Keep Their Boats in Marinas
Factor
Water
Depth
Water.
Quality
Natural
Beauty
Protected
Anchorage
Lack of
Congestion
% Not
Important
7.7
9.3
9.8
23.5
19.9
% Slightly
Important
8.9
10.0
13.3
15.1
17.4
% Moderately
Important
24.4
35.7
41.0
31.5
37.2
% Very
Important
54.9
39.7
31.2
25.1
20.4
% No
Response
4.1
5 . 3
4.7
4.8
5.1
Source: Maryland Boat Owners Survey, 1983
78
-------
alter the boating area by trailering to different launch sites. Of course, the
farther they must trailer the boat, the more costly (in both time and money)
will be the trip. Responses to the survey indicate that factors besides costs
are important in choosing boating area. Specifically, water quality was
considered the moat important factor in this choice, Thus we analyze the
demand for trips to different sites with measurably different water quality.
This gives us some basis to deduce how demand might change if water quality
were to change at different sites.
In order to accomplish this, a model of multiple site demand is designed to
accommodate differing qualities across sites but to require no more data than is
available from the boat owners' survey and Chesapeake Bay water quality
data. Given these limitations and a desire for consistency of analysis across
recreational activities, the varying parameters model presented as the first
method of analysis in Chapter 4 is used.
The model takes the following form. The demand for trips to each site j
is estimated as a function of the coat to individual i of accessing the site
(c J j ), substitute site ace-a ccsts (s ^ j ), and other exogenous variables
associated with the individual (s ^ ) :
(5.1) xjj = fjicij, Sij, zj; fj) for all i,
where f j is the demand function for the jthsite, fj is the vector of
parameters in each of the site demand functions to be estimated.
Equation (5.1), which is the first stage of the varying parameters models,
can not be estimated using ordinary least squares methods. The sample upon
which the estimation is baaed includes a large number of zero values for the
dependent variable. As described earlier in this volume, ordinary least
squares applied to censored samples will produce biased estimators. As
described in the last chapter, Tobit estimation procedures are used to correct
for the problem.
The second stage of the model relates the set of f parameters in the site
demand functions to the site water quality characteristics (a vector w i), In
this way the demand for a site is implicitly modelled as a function 'of the
site's characteristics. The fecond stage model is of the form:
(5.2) fk. =gk(w, ) forall j,
where k indexes the Specific f coefficient within the demand equations and j
indexes the site demand. Again the application of ordinary least squares is
not optimal. The above model likely suffers from heteroskedasticity (see
Desvousges, Smith, and McGivney, 19S3) which will produce Inefficient
estimators. To correct for the expected heteroskedasticity, the entire equation
can be multiplied by the reciprocal of the standard error of the respective
estimated Coefficient Thus, if »k j is the standard error of the estimated
coefficient fk j from the Tobit estimation Of the demand for site j, then the
79
-------
second stage model corrected for heteroskedasticity can be written
(5.3)
for each of the k=l,...,K P coefficients from the first stage.
The Data
The arguments of equation (5.1) which were determined to be relevant for
the analysis include the coat of access to the site, the coat of access to a
substitute site, and the value of the boat. The latter variable waa chosen as
a surrogate for two very important factors. Boat value and income are highly
correlated , so boat value serves as a surrogate for income level, a variable
which is normally rather difficult to obtain with accuracy. Boat value alao will
be highly correlated with boat sise which haa an effect on the site choice.
Since boat length and income are correlated, including them separately in the
equation produces unresolvable multicollinearity.
The coat of access variable is the roundtrip cost, including time coat.
Costs vary across boaters for several reasons. Cbviously the county of origin
influences costs, for the further the county is from the launch site, the
higher the coat of travel. The cost variables include time costs, computed as
one-half the individual's average hourly wage (incorie/2080) time, the distance
traveled divided by 40 miles per hour. Finally the money cost of trailering a
boat depends on the size of the individual's boat. The coat per mile of
trailering a boat -was estimated, using coat and boat length information from
the data set, as -.78 + .08 *(boat length). The final coat variable includes the
toll for the Bay Bridge, when relevant.
To calculate the substitute cost, the above formula waa applied to the
closest cite not chosen. While a vector of costs to all alternative sites might
be considered preferable, there are several practical problems with including
such a vector in the regression, espicially severe multicollinearity.
Not all observations on boat owners who trailered their boats were ueed
because of heterogeneity of respondents and incomplete data. Observations
were deleted if the individual did not use his boat for any trip, in 1983 or if
he did not report launch sites or location of residence. Additionally, those
who reporeted their residence to be a distant state, precluding day trips to
the Bay, were deleted. Finally, to make the sample relatively homogeneous,
sailboats were excluded from the analysis. The latter accounted ultimately for
only seventeen deletions. The final sample included 408 observations.
The above information, waa obtained from the boat owners' survey.
However, this data source dose not include any information about either
perceived or measurable water quality at various sites. Once again the
Chessie system provided the environmental data, and the environmental quality
variable used was the product of nitrogen and phosphorus, as in Chapter 3.
80
-------
The Estimated Model
For the first stage of the model, the parameters in the following functional
form are estimated:
(5.4)
where x jj is trips to site j, is cost to site j , s cost to next best
alternative to site j,
jj is trips to site j, ^ j is cost to site j , s^ co
ive to site j, and b i is boat value, all for individual i.
The first stage results for each of the twelve sites are presented in Table
5.8, The results are remarkably consistent across sites. The own price
coefficients are all negative and significantly different from zero at the 99%
level of confidence. Substitute price coefficients are universally positive and
significantly different from zero for eight of the twelve sites. The coefficient
on boat value is significantly different from zero for seven of the sites and in
each of theaa cases haa a positive sign, suggesting that wealthier people
and/or people with bigger boats take more boating trips, ceteris paribus.
Demand is relatively inelastic with respect to substitute price and boat
value (i.e. a 1 percent change in either of these causes a less than 1 percent
change in the demand for trips), However, the demand for trips to a site is
quite elastic with respect to the cost of accessing the site, with own price
elasticity ranging from -1.5 to -7.0.
In the second stage, there are as many equations as there are parameter
from the first stage which we wish to allow to vary with the environmental
factors. Since we have no particular a priori information as to what
parameters might vary, we can. model each as a function of the environmental
variables and allow the test statistics to determine the outcome. *
The second stage model is given by
(5.5) JkJ = a0 + at TNPj + vj
and the results are presented in Table 5.9. The product of nitrogen and
phosphorous serves as the environmental variable. The regression of the own
coat coefficients from the linear first stage model on these environmental
variables produced good results. The coefficient is significantly different
from zero and negative. The negative sign suggests that the demand curve
becomes less steep with increasing levels of pollutant.
Neither the constant term nor the coefficient on boat value yielded
significant second stage results. Even though the coefficient of substitute
price waa associated with a significant negative coefficient on the
environmental variable, allowing this coefficient to vary had no appreciable
effect on the welfare results. As a consequence, in the remainder of the
analysis we allow only the coefficient on own price to vary wtth environmental
quality. The results suggest that an increase in pollution would tend to have
the effect ofpivoting inward the demand for trips to the site.
81
-------
TABLE 5.8
Estimated Tobit Demand Coefficients: Maryland Counties 1983
Estimated Coefficients (Variable)3
county
Anne Arundel
Baltimore
Calvert
Cecil
Charles
Dorchester
Harford
Kent
Queen Anne's
St. Marys
Somerset
Wicomico
,aEach equation
Pi
(cost/
trip)
-.13 .
(8.75)b
-.43
(9.21)
-.14
(4.14)
-.22
(4.84)
-.34 .
(8.41)
-.09
(2.86)
-.15
(5.55)
-.25
(4.94)
-.27
(6.17)
-.11
(6.40)
-.12
(4.76)
-.15
(6. 93)
is estimated
P,
(substitute
cost/trip)
.03
(3.42)
-.02
(1.13)
.08
(13.70)
.04
(1.54)
.07
(3.77)
.08
(2.69)
.05
(2.63)
. 1
(3.57)
.07
(2.88)
.05
(3.12)
-.03
(.58)
.05
(1.58)
with the 496
/»3 do
(boat value (constant)
1000's)
1.29
(5.94)
1.78
(4.01)
1.84
(3.45)
2.12
(3.09)
2.75
(6.79)
.66
(.78)
1.51
(1.67)
1 .14
(.14)
.12
(. 19)
1.25
(2.94)
2.81
(3.13)
1.02
(1.71)
boaters who
-2.21
(1.61)
-1.94
(.77)
-27.14
(7.21)
-16.44
(3.87)
.49
(• 19)
-34.38
(6.68)
-12.21
(3.74)
-18.25
(3.45)
-3.83
(1.03)
-9.46
(3.31)
-37.20
(6.64)
-7.03
(2.02)
trailer their
non-
limit
observa-
tions
142
75
44
17
38
30
36
28
36
67
24
26
boats.
^Absolute values of t-statistics in parentheses
-------
Table 5.9
Estimation Results from Second Stage
Regression of trip coat coefficient
Regression of substitute cost coefficient
Regression of constant
Constant
-.0887
(4.29)a
.0682
(7.47)
-19.414
(4.14)
TNP
-.000102
(3.54)
-.000016
(1.78)
.007338
(1.93)
• absolute values of t-statistics in parentheses
Modelling the Behavior of Boat Owners Who Do Not Trailer Their Boats
Boat owners who keep their boats in the water during the season are far
more restricted in the short run as to the quality of the water that they can
easily enjoy. To a/ferthe area in which they boat they must take long trips
in their boats or they must change mooring arrangement. We do not have
access to information that would allow us to model either of these decisions.
None ofour data allows us to observe these individual making trade-offs
between water quality and other goods ormoney. As a consequence we can
not deduce from observations on their behavior how they might value
improvements in water quality.
What we can do however is to learn something about their demand for
boating and their use of the Chesapeake, which in itself is useful information
for the policy maker. In this section we estimate the demand for boating
trip, by boaters who keep their boats in marinas. Since we are interested in
the short run decision of how many tripe to take, the relevant cost variable is
the variable coat of atrip. Given the information available, we can
approximate the money and time costs of travel to the marina. Explanatory
variables which may shift the demand function include income and the size (or
value) of the boat."
The boats in the marina subgroup are somewhat heterogeneous. For one
thing, about half are sail boats and half are motor boats. Consequently we
might wish to teat whether the demand functions for the two groups are
significantly different.
The results of thin analysis are reportad in Table 5.10. The coefficient on
coot is negative, as expected, and significantly different from zero. The
coefficient on boat value is positive, indicating that, all else equal, boaters
with more expensive boats take more trips. Income appears not to affect
systematically the demand for trips.
83
-------
Table 5.10
Estimated Demand for Boating Trip.
Boats Kept in Marinas
Explanation Variable
Cost
Boat Value ($1, 000)
Income ($1,000)
Sailboat Indexa
Cost Sailboat Index
Constant
R2 = .109
# of Observations = 240
Coefficient
-1.046
.357
.148
-23.596
.615
63.363
t-statistic
-3.93
1.94
.66
-2.27
2.16
6.66
•Variable equals 1 if boat is sailboat, 0 otherwise
Both the constant term and the oostcoefficient shift significantly for
individuals who own sail boats. The demand function for non-sail boat trips
is given by
TRPS = 63.36- 1.05 coat + .36 boat value + .15 income
where boat value and income are in thousands of dollars. The sail boat
demand for trips is given by
TRPS = 39.77- .43 cost + .35 boat value + .15 income.
At the same cost and boat value, sailboat owner, demand fewer trips, and
their demand for trips appears to be rim-e inelastic.
Because of the eventual need to aggregate behavior over recreational
activities, it would also be useful to know whether those boaters who •penal a
large portion of their time fishing have significantly different demands from
those who do not. These results are presented in Table 15.11. It appears
that the two groups demands are different. For any given cost and boat
value, fishermen demand more trips and their demand tends to be more elastic.
Fishermen}, demand is given by
84
-------
TOES = 73.61- 1.32 coat + .35 boat value + .15 income
and demand by non-fishermen is
TRPS = 43.00- .51 cost + .35 boat value + .15 income.
Table 5.11
Estimated Demand for Boating Trips, Fishing Behavior
Boats Kept in Marinas
Explanatory Variable
Trip Coat
Boat Value ($1,000)
Income ($1,000)
Fishing Indexa
Coat Fishing Index
Constant
Coefficient
-.506
.350
.148
30.535
-.815
43.085
t-statistic
-3.86
1.90
.88
2.8s
-1.82
3.81
R2 = .116
• Variable equals 1 if boater fishes on boating tripe "usually" or "always,1
0 otherwise.
calculating Estimates of the Benefits of Water Quality
Improvements for the Trailered Boat Samle
Because we have bean able to estimate the demand for boating trips by
boaters who trailer their boats to different areas as functions of costs and
the water quality in those areas, we can estimate welfare gains and losses
from water quality changes to this group. Unfortunately no observable
behavior of the boaters who keep their boats in the water allows us to deduce
anything about the value they place on improved water quality.
As in the previous chapter, three different changes in the environmental
variables are considered. In one caae we impose a 20 percent decrease
(environmental improvement) in the environmental (pollution) variables.
Subsequent experiments include a 10 percent decrease and a 20 percent
increase (environmental degradation).
85
-------
Additionally, two different calculation procedures are provided, identical to
those used With the varying parameters model in Chapter 4. The two methods
of calculating "before" and "after" trip demands yield two sets of average
consumer surplus estimate, before and after the environment change and,
consequently, two sets of benefit (or leas) estimates due to the environmental
change. (A description of these methods can be found in Chapter 4.) The
estimates are reported in Tables 5.12? 5.13, and 5.14 for the 10 percent
improvement 20 percent improvement and 20 percent deterioration in the
water quality variable, respectively. The average consumer surplus figures
are per boater ( trailered boats, only) per season for the designated site.
Thus the first entry in Table 5.12 is an estimate of the value of access to
sites in Anne Arundel County per boater in the sample.
In examining the consumer surplus figures, it is well to keep in mind that
these benefit estimates are affected by the probability that a boater will go to
a particular site, the number of trips taken, given that the boater goes to the
site, as well as the size of the own-price coefficient. The figures in columns
3 and 6 are estimates per boater of how the value of access to a site changes
with changes in the environmental variable. These surpluses are not additive
across sites. That is, if we want to consider the effects of a ten-percent
decrease in TNP throughout the Chesapeake Bay, we cannot add surplus
changes across sites. Each estimate of surplus per boater by site assumes
that the water quality at other sites remains fixed. The bias which would be
created by simple aggregation across sites depends on price and quality
elasticities and is of unknown size.
What do all these calculation say about the value of reductions in the
nitrogen/phosphorus variable? We can estimate the aggregate benefits of
changes in TNP from Tables 5.12, 5.13, and 5.14 by expanding these estimates
from the sample to the population of boat owners who trailer their boats.
Consider St. Mary's County. We have two estimates of the increase in Surplus
associated with a 10 percent decrease in TNP. These suggest a range of
between $1 and $5 per boater per season. If there are about 80,000 boaters
who trailer in Maryland (about the number estimated by Harmon and Associates
for 1983), we would estimate a change in total surplus in the range of $575,000
to $1,400,000 annually for a 20 percent reduction in TNP at the sites that
Anne Arundel County comprises. This calculation assumes that the original
sample from which the benefit estimate, were derived is representative of the
boater population as a whole.
The results in these tables eeem at least plausible. For example,
surpluses appear highest for western shore waters, those moat easily accessed
by the concentration of population in the state. The surplus figures for any
given site are not especially large, but this is to be expected since when
boaters have the ability to substitute relatively cheaply among sites, very
high surpluses at any one site would violate some prior expectations on the
size of benefits. While the magnitudes of returna from changes in TNP are
not extremely large on a per-boater basis, in the aggregate they are quite
substantial
86
-------
Table 5.12
Per Boater Annual Benefits from a 10% Decrease in Pollutant
in each Geographical Area
CJ
Calculation Method
County with
Water Quality
Change
Anne Arundel
Baltimore
Calvert
Ceci 1
Charles
Dorcester
Harford
Kent
Queen Anne's
St. Mary's
Somerset
Wicomico
A
Calculation Method B
Consumer Surplus Consumer Surplus Consumer Surplus
Before Change After Change Benefits Before Change
$30.01
15.07
17.50
4.80
38.79
1.61
3.45
24.08
26.17
14.80
7.17
7.87
$33.30
17.38
18.60
5.43
46.05
1.78
3.89
27.09
28.99
16.03
7.60
8.60
$3.29
2.32
1.11
.63
7.26
.17
.45
3.00
2.81
1.23
.44
.73
$119.05
49.83
108.57
18.55
38.34
75.72
47.24
73.71
51.74
139.22
99.18
32.82
Consumer Surplus
After Change
$127.46
54.95
111.38
19.51
43.40
78.15
49.63
76.86
53.98
"143.71
101.95
34.53
Benefits
$8.41
5.12
2.81
.96
5.05
2.43
2.38
3.15
2.24
4.49
2.77
1.71
-------
Table 5.13
Per Boater Annual Benefits from a 20% Decrease in Pollutant
in each Geographical Area
Calculation Method A
Count y with
Water Quality
Change
Anne Arundel
Baltimore
Calvert
Cecil
Charles
Dorcester
Harford
Kent
Queen Anne's
St. Mary's
Somerset
Wicomico
Consumer Surplus
Before Change
$30.01
15.07
17.50
4.80
38.79
1.61
3.45
24.08
26.17
14.80
7.17
7.87
Consumer Surplus
After Change
$37.17
20.33
19.79
6.17
55.54
1.98
4.41
30.51
32.18
17.40
8.08
9.46
Benefits
$7.16
5.26
2.30
1.38
16.75
.37
.97
6.42
6.00
2.61
.91
1.59
Calculation Method B
Consumer Surplus
Before Change
$119.05
49.83
108.57
18.55
38.34
75.72
47.24
73.71
51.74
139.22
99.18
32.82
Consumer Surplus
After Change
$137.05
61.20
114.35
20.60
50.13
80.74
52.27
80.38
56.51
148.54
104.88
36.44
Benefits
$18.01
11.37
5.78
2.06
11.79
5.03
5.03
6.67
4.76
9.32
5.70
3.62
-------
Table 5.14
Per Boater Annual Losses from a 20% Increase in Pollutant
in each Geographical Area
oo
V0
Calculation Method A
County with
Water Quality
Change
Anne Arundel
Baltimore
Calvert
Cecil
Charles
Dorcester
Harford
Kent
Queen Anne's
St. Mary's
Somerset
Wicomico
Consumer Surplus
Before Change
$30.01
15.07
17.50
4.80
38.79
1.61
3.45
24.08
26.17
14.80
7.17
7.87
Consumer Surplus
After Change
$24.76
11.68
15.51
3.80
28.78
1.33
2.74
19.12
21.48
12.70
6.40
6.68
Losses
$5.24
3.38
1.99
1.00
10.01
.28
.71
4.96
4.70
2.09
.77
1.19
Calculation Method B
Consumer Surplus
Before Change
$119.05
49.83
108.57
18.55
38.34
75.72
47.24
73.71
51.74
139.22
99.18
32.82
Consumer Surplus
After Change
$105.07
41.96
103.98
16.93
31.51
71.30
43.12
68.34
47.93
131.10
94.08
29.88
Losses
$13.98
7.87
5.19
1.62
6.83
4.42
4.13
5.37
3.81
8.12
5.10
2.94
-------
Chapter 6
The Benefits for Recreational Fishing: Striped Bass
This chapter provides come preliminary estimates of the increase in
benefits to sport anglers from increases in water quality. We use aportion of
the 1980 National Survey of Fishing, Hunting and Wildlife Related Recreation,
referred toas USFWS data, to estimate the demand for and value of fishing
for striped bass in Maryland. This survey, while not designed for these
purposes, is the only data aet currently available which enables us to
investigate the recreational fishing of the Chesapeake. Striped bass is the
only specie, important to the Chesapeake Bay recreational fishery forwhich
there is sufficiently detailed-catch information to link water quality changes to
the benefits of sportfishing.
The link between improved water quality and changes in recreational
fishing demand depends on the ecological connection between water quality
and catch rates and the behavioral connection betwean catch rates and fishing
activities. Descriptive and analytic studies of the Bay have focused on the
impacts of water pollution on the density and productivity offish stocks.
Lower dissolved oxygen, declines in SAV, and increases in water toxicants all
appear to have an impact on fish stocks. Further, where records are kept for
commercial fisheries, there haa been a substantial decline in landings per unit
effort, especially for those species which spawn in the Bay or ita tributaries
It is plausible to expect considerable benefits to recreational fishermen
from improvements in water quality. The number of recreational anglers is
quite large, baaed on information from the primary sources of data on
saltwater recreational fishing in Maryland. Estimates of saltwater fishing
participation in Maryland during 1980 range from 539,000 anglers over 16
years of age taking 4.1 million trips to somewhat over 800,000 anglers of all
age. taking 2.7 million trips (U. S. Fish and Wildlife Service and Bureau of
Census; U. S. National Marine Fisheries Service; William et al.). According to
NMFS and State of Maryland data, each saltwater angler took approximately
three trips, while USFWS estimates approximately 7.6 trip, and 9.0 days fished
per angler.
Data on striped bass fishing are somewhat more difficult to obtain.
According to the Maryland Department ofNatural Resources, roughly 203,000 of
the saltwater trips were for stiped bass. Our analyais of the USFWS data
indicates that 239,000 anglers (over 16 years of age) fished for striped bass
in Maryland and Sussex County, Delaware, fishing for approximately 2.1 million
days, or roughly 8.8 days per angler. Estimates of the striped baas
recreational catch in Maryland range from 211,000 to 377,000 fish, a total
weight of 200 to 474 metric tons. The USFWS data are not well suited for
estimating aggregate catch, because the survey used waa designed primarily
for other purpose., even though catches are self-reported by respondents for
come saltwater species, notably striped bass.
90
-------
Table 6.1 provides some descriptive information about the sample of
anglers which was analyzed in this portion of the study. The sample was
partitioned into two groups baaed on whether the individual fished for striped
bass or not. Individuals in the two subsamples are very similar in the amount
of fishing and hunting done and in their exprience, income, age, education,
and other demographic makeup, Striped baea fishermen, on average, showed a
slightly higher propensity to own a boat and to allocate moire money to
hunting and fishing activities, though these differences are not significantly
different from zero due to the high within-subsample variation.
Table 6.1
Characteristics of Striped Bass Fishermen
and Other Fishermen and/or Hunters in the Sample
Striped Baaa
Fishermen
Non-Striped Baaa
Fishermen
Number of Individual in Sample
Average Number of Days Fishing, Striped Baas
Average Number of Days Fishing, All Species
Percent Who Also Hunted
Average Number of Days, Hunting
Average Years of Fishing Experience
Average Age When First Fished
Percent Owning Inboard Boat
Percent Owing Outboard Boat
Percent owning Other Boat
Average Household Income
Average Fishing/Hunting Budget in 1980
Average A9ga
Average Years of Schooling
Percent Working in Job or Business
Percent fran Urban Areas
184
11 daya
28 daya
41%
17 daya
24 years
10
19%
42%
17%
$28,300
$982
38
13 years
70%
44%
576
0 days
2'7 daye
37%
15 days
24 years
12
7%
28%
12%
$27,600
$588
38
13 years
73%
38%
aThe sample is for individuals 16 yeara of age and over.
91
-------
A Description of the Data
The 1980 National Survey of Fishing, Hunting, and Wildlife-Associated
Recreation was the source data for analysis. Of the available data sets on
Chesapeake Bay sportfishing, the portions of this survey relating to saltwater
recreational fishing in Maryland, and by Maryland residents, offered the best
prospects for modelling the effects of water quality improvements. This data
set contained the essential variables for estimating recreational fishing demand
functions, including information on (a) trips taken by destination; (b) costs
incurred by recreationists for goods and services used in recreation; (c)
household income; and (d) catch rates reported by anglers for certain
species.
The survey consisted of two parts. The first was a telephone screening
of households, predominantly by telephone interviews, to collect demographic
characteristics and to determine the hunting, fishing, and non-consumptive
recreation activities ofhousehold members during 1980. The second part was
a detailed questionnaire administered (typically in person) to selected
individuals who indicated they had hunted or fished in 1880, collecting
information on activities and expenditures. Of the 30,300 fishermen and
hunters and 6,000 non-consumptive users interviewed nationwide, 760 pursued
some or all of these activities in Maryland. These 760 individuals were the
subject of this analysis.
Of the 760 who hunted, fished or Participate in non-consumptive
wildlife-related activities in Maryland, 456 indicated they participated in some
form of saltwater fishing. Catch rate estimates were only obtained for a
limited number of saltwater and estuarine species, with striped baas the only
recorded species relevant to Maryland. One hundred eighty-four individuals
indicated they fished for striped baas in 1960.
The survey waa designed to provide estimates of recreation activities and
expenditures at the state level, and states were divided into large subregions
for purpose, of identifying trip destinations. Maryland was divided into four
such regions, three of which border the Chesapeake and were the location of
striped bass fishing. Broadly defined, the four areas are: the Southeastern
Chesapeake region, Northern Chesapeake, Southwestern Chesapeake, and
Northwestern Maryland. Significant numbers of Maryland residents also fished
for striped baaa in Sussex County, Delaware. Of the 184 striped bass
fishermen in the sample, 16 reported fishing in Delaware, 46 indicated they
fished for striped baaa in the Northern Chesapeake, 59 fished in the
Southeastern Chesapeake region, and 86 in the Southwestern Chesapeake
(Table 6.2).
The data aet includes days fished for • trtped bass and other species,
rather than number of trips by specie., the latter being the preferable
measure for travel cost models. The survey did, however, include the total
number of trips to each region. Aggregating over all areas to get total trips
and all species to get total days fished, it was determined that anglers took
about 4.1 million trips and fished about 4.8 million days, yielding an average
of 1.17 days/trip. Thus, the two measures may not be bad approximations of
one another.
92
-------
Table 6.2
Sample Distribution Number ofFishermen, Days of Striped Bass Fishing
in 1880, and Catch Rate, By Regions
Region
Sussex DB
Northern Chesapeake3
Southeastern Chesapeake*3
Southwestern Chesapeake
Number
Who Visitad
16
46
59
88
Mean Striped
Baas Days
6.8
9.6
11.3
8.8
Mean Catch Rate
Fish/Day
4.4
4.9
3.3
2.8
'Baltimore City and Bait imore, Carroll , Cecil , Harford, Kent , and Queen Anne's
counties.
^Carol ine, Dorcester, Somerset, Talbot, Wicomico, and Worcester count ies.
cAime Arundel, Calvert, Charles, Howard, Montgomery, Prince George's, and St.
Mary's counties.
While detailed information was collected on costs of travel, lodging, food,
fees, and other expenses incurred during recreation trip., these costs were
not area specific; instead total expenditures over all saltwater fishing trips to
all areas (regardless of species sought) were collected for each cost category.
The variable cost of trips to a single area could be determined only by
prorating total variable costs according to distance travelled. The method
ueed in this analysis waa to determine the total miles travelled by the
individual for all saltwater fishing trip, in 1980, aa the •m of products of
round trip miles travelled to (the usual fishing location in) each area and the
number of trips taken to each area, The fraction of total variable fishing
expenses prorated for eaoh trip to each site waa the round trip miles
travelled to the site divided by total miles travelled. The money coat of a
trip to each site waa this fraction times the reported total variable costs for
saltwater fishing. Espressed as a formula,
n
B
where MC j j is the money coat of a trip by individual i to area j, Mj_ j is the
round trip miles travelled by individual i to area j, •< i is the number of trips
individual i takes to area j, VC ( is individual i'. reported saltwater fishing
variable costs, and there are n areaa.
The coat of time spent in recreation is also an important determinant of
demand. The • urvey data were not ideal for determining this coat because no
93
-------
information was collected on the time spent in travel or at the site on each
trip. However, miles traveled is a reasonably good proxy for time spent in
travel. The procedure used here waa to assume an average rate of speed
during travel of 40 miles per hour and that the annual household income
divided by the number of hours in the average full-time work year (2,040)
was a suitable approximation for the wage rate. Then, the value of time
travelled was determined as the product of the amount of time spent in travel
and 40 percent of the wage rate. Expressed as a formula,
TCJ -
y«
2,040
204,000
where TC i j is the time coat for individual i travelling to area j, y. is the
household of person i, and M « * is, again, round-trip miles. Of course, this is
a rather arbitrary formulation for time cost based on a series of restrictive
assumptions, but preferable ways of treating the value of time were not possi-
ble given the available data. The full price of a trip is then calculated as
the sum of the time and money prices for each individual 1P1 = MCj + TC^.
In the survey, respondents were asked to estimate their average catch
rate per day for selected species. Unfortunately, there was a lag of up to a
year or more between the time the fishing trip was taken and the time the
questionnaire was answered. There is evidence (e.g. Deuel, Hiett and Worrall)
that fishermen do not accurately remember numbers of fish caught or their
sizes well beyond a period of a few months. A comparison of the USFWS data
and data collected by the State of Maryland suggests that the USFWS data
might contain an upward bias in reported catch rates. The sample and
population average catch rates were both somewhat over three striped bass
per day, which is considerably higher than the State of Maryland data which
suggests a catch rate for the came period of one striped bass per day. When
the sample catch rates were extrapolated to estimate total 1980 catch, the
estimate was an order ofmagnitude or more larger than the published
estimates noted in the introduction, although some of this difference may be
attributable to difference, in estimates baaed on total trips versus total days.
The fact that sample catch rates do not predict aggregate catch well does not
invalidate their uae as quality indicators, however. As indicators of the
quality factors which signal individuals' fishing decisions, sample catch rates
may perform quite well.
The survey data contained a categorical variable measure of household
income. A second measure was also calculated: total budget for fishing and
hunting recreation, the sum of all fishing and hunting-related expenditures in
1980. If the individual has a weakly separable utility function and determines
first the total amount of income to allocate to hunting and fising recreation,
94
-------
the fishing and hunting budget to a more relevant income constraint than
overall household income. The fishing/hunting budget measures, however, is
subject to potential errors Of measurement, both from faulty recall by
respondents and from year-to-year fluctuations due to purchase of major
durable goods.
The Basic Model
For a variety ofreasons, the model we estimate for recreational fishing is
different from the recreational beach uee and boating model. In the beach
use and boating estimations , trip data existed for a number of
quality-differentiated sites. In the fishing data, trips are available only by
region. There are only four of these regions, and each is large so there can
easily be as much variation within any region M there in among the regions.
Further, 164 of the 184 striped bass anglers in the sample visited only one
region. Rather than estimating four demand curves, we have estimated a
single equation where the dependent variable is the sum of the trips to all
sites.
The handling of the quality variables differs al- For recreational
boating, w. ueed a varying Parameter model because the quality variable,
scientific measures of water quality, varied across sites but not across
individuals. The quality variable in recreational fishting, catch rate, varies
across individuals. Consequently, we need not use a varying parameter
model. Instead we uee the observation on the reported catch in arena where
the individual took his trips. The data set includes many individuals who did
not fish for striped bass. For these individual, costs and catch rates were
inferred.
The fishing model estimated waa
(6.1) x, = f9 + ftTCi + 0aCH, + 03IB, + 04CB1 + 0SBD1
where xi is the number of days taken by the i^ individual, TCd is the
individual', full coat (in dollars per trip) of striped bass fishing, CRi is the
catch rate (fish per day), IB ^ andOB j_are (0,1) variables denoting availability
of an inboard or outboard boat for fishing, respectively; and BD1 is the
individual'. fishing/hunting budget in dollar, per year.
No eubatitute sites were specified in the model because the regions were
so broadly defined that they might not in fact act as substitutes for each
other. There is probably extensive substitution among sites within each
region that cannot be captured at all given the level ofaggregation we face;
and the sample data indicates that only about 10 percent of respondents
visited more than one region. Instead, the price and catch rate for
Participant who visited more than one site were calculated as the mean of
price, and catch rate, at each region visited, weighted by the day. fished.
Only slightly more than one-quarter of the respondents who either hunted
or fished in Maryland reported having fished for striped bass. This level of
95
-------
non-participation implies a serious censored variables problem. There are
several ways of handling this problem in recreational demand models. We
explore these approaches in Chapter 4 of Volume I of this report. For the
current task of estimating fisheries demand models, we choose the following
simple Tobit formulation:
(P'Z + e P'z + « > 0
(6-2) x- lo ft'z + « * 0
where z is the vector of explanatory variable.. The Tobit model imposes come
rather extreme restrictions on individual behavior that more general sample
selection models avoid. But for preliminary results, we accept these
restrictions for the sake of simplicity.
Determining the relevant price and catch rate for non-participants was
problematic. For these individuals, it was not known which of the four
price-quality combinations were moat relevant to their decisionto go/not go
striped bass fishing. In the application we used the minimum price to access
a striped bass "site" and its corresponding catch rate.
Welfare measures are calculated, in principle, the " same way as for the
varying parameter, model. That is, the benefits of an increase in catch rates
are given by the change in consumer 'a • mplua which, for the linear model
above, is
(6.3,
-2ft, -2ft,
where t\ is the own-price coefficient, and x is the individual's trip level.
Empirical Results
The model in equations (6.1) and (6.2) was estimated using the maximum
likelihood method of LIMDBP. Table 6.3 gives the results which will be used
for preliminary benefit estimation, along with the sample means of the
variables. The results in Table 6.3 are fora model in which actual catch
rates reported were used for participant, and a predicted catch rate was
used for non -participants. We also estimated a model in which predicted catch
rates were used for every individual In the latter estimation, the coefficient
estimates remained basically unchanged, but the standard error on the catch
rate coefficient increased resulting in a t-statistic of about 1.3.
The coefficient estimates all have intuitively correct signs, and they are
different from zero at better than the 5 percent significance level Having an
inboard motorboat seems to induce more striped bass tripa than having an
outboard motorboat. The own-price elasticity for Participant is about minus
one, while the catch rata elasticity for participant is about .10,
96
-------
Table 6.3
Tobit Estimation of the Demand for Striped Bass Fishing
Explanatory
Variable
Constant (C)
Own price ('IT)
Catch rate (CR)
Inboard Motor ( IB)
Outboard Motor (OB)
Budget (FHB)
** = 18.3
N= 760
Coefficient
Estimate
-10.6
-.336
.337
12.65
6.66
1.40
t-statistic
-5.79
-7.52
2.13
4.49
3.47
3.04
Mean of
Variable
1.00
$27.2
3.2 fish/day
.10
.31
. 70($000)
We can use the estimated coefficients in Table 6.3 to estimate welfare
effects of increases in catch rates. As in Chapters 4 and 5, two estimates of
consumer surplus are provided. Method A employs predicted trips plus
changes in predictions whereas Method B uses actual trips plus changes in
predictions.
It is rather eaay to expand sample results to the population, since the
Fish and Wildlife Survey includes sample weight or sample expansion factors,
These weights account for the fact that different population strata are
sampled disproportionately. Consumer's • rplua for the population is simply
the weighted sum of the surpluses of the sample observation:
(6.4)
Cs =
where s is the sample size and f ( is the expansion factor.
Table 6.4 gives the estimates of aggregate surplus. The first column is
the estimate of the value of access to striped baas fishing as it was perceived
in 1880, baaed on 1980 prices. The actual and predicted estimates differ
substantially, with the actual being more than three time, larger than the
predicted.
97
-------
Table 6.4
Aggregate Consumers' Surplus for Striped Bass Fishing:
Effect ofChanging Catch Rates, 1960
(Population of Maryland Hunters and Fishermen)
Aggregate
Consumer Surplus Change with Change in Catch Rate
Surplus
for Access 20% Decrease 10% Increase 20% Increase
Thousands of Dollars
Predicted (Method A) 14,652 -572 314 1,501
Actual (Method B) 54,196 -422 231 4sl
The second, third and fourth columns in Table 6.4 give the net impact of
a 20 percent reduction, 10 percent increase and 20 percent increase,
respectively, in the striped baa. catch rate compared with the level perceived
in 1960. Here the actual and predicted results are closer, especially for the
10 percent changes.
The numbers that are most interesting for environmental policy on the
Chesapeake are found in the third and fourth columns. These figures are
rough estimate, of the dollar amount people who currently fish or hunt in
Maryland might gain annually from improving striped baas fishing.
There are a number of complicating factors which cannot be integrated
into our preliminary calculations ofbenefit estimates. First, consider how long
it would take for environmental policy to produce a • ubdantid, sustainable
increase in catch rate. Reduction in effluents for one year will have only a
small effect. To improve ambient water quality enough to bring about better
striped bass reproduction and survival could take many years.
The second question relates to the role of expectations regarding catch.
Aaide from the likely bias and high noise in the catch rate estimate, what
respondents report is the ox post realisation of catch rates, while their
decisions regarding whether, when, and how frequently to go are baaed on
expectation about the catch rate, ex ante. Consequently, while recalled ex
post catch rate ds the best quality variable we could obtain for striped baas
96
-------
fishing, we need to be skeptical about its implications for the relationship
between days taken and expected catch rate.
The second question concerning these benefit estimates is whether, given
sufficiently improved ambient water quality, the catch rates are sustainable.
The answer is no. Better catch rates induce more fishing and hence more
harvest. Since there is some evidence that overharvesting is partly
responsible for the decline in fish populations to begin with, it is likely that
healthier stocks will induce more harvesting. The long run equilibrium will
result in higher than current benefits, but smaller than the benefits which
implicitly assume that the increase in fishing effort will have no long run
effects on fish stocks.
Last, it is worth remembering that the benefit estimates are baaed on a
sample of households that hunted or fished in Maryland in 1980. If there are
people who currently do not hunt or fish, but would go striped bass fishing
if the fishing improved sufficiently, then the annual benefit estimates are an
underestimate.
99
-------
chapter 7
Conclusions
Restorationof Chesapeake Bay water qulity rewires substantial
resources on the part of public agencies, private firms and households. There
are many choices to be made in implementing programs to clean up the Bay.
This report haa described acme of the activities which would benefit from the
enhancement of Bay water quality. Chapters 2 and 3 described ways in which
people think about the Bay and benefit from better water quality. Chapters 4
through 6 contain descriptions of come recreational activities which would gain
from improved water quality. All of these chapters provide estimated
willingness to pay from potential improvements.
In deriving benefits, sometimes we 1000 eight ofthe informational content
of the models behind the benefit estimates, the estimated demand functions
themselves. Chapters 2 through 6 contain substantial new information about
the structure of demand for recreational activities associated with the
Chesapeake Bay. In nearly every instance where sufficient data were
available, recreators responded to travel and time costs in a manner consistent
with our theoretical model. They were also observed to be responsive to even
the crudest of water quality measures. Additionally, demographic variables
such as income, race, and boat ownership were observed to influence behavior.
As we turn to the benefit estimates, the reader is reminded not to consider
the "bottom-he" benefit figures as the only value of this report.
Demand for Chesapeake Bay Recreational Activities
The data and modelling exercises described in Chapters 3 through 6
provide a good picture of the recreational use of the Chesapeake Bay.
Chapter 3 includes an overall picture of Chesapeake recreational activities
derived from a random sample of all households in the Baltimore/Washington
SMSA's (BWSMSA) . This survey revealed that a full 43 percent of the BWSMSA
population used the Bay or intended to use it for recreation in 19S4.
Geographical distribution of users showed Anne Arundel County reaidenta (69
percent) moat likely to be Bay users and District of Columbia residents (21
percent) to be least likely (ace Table 3.1). The moat common recreational
activities were fishing, swimming and boating, with about a third of the Bay
users participating in all three activities. Of these activities, swimming was
enjoyed by more people than either of the other two, with 77 percent of users
participating. In the remaining chapters, each of these activities was looked
at in greater detail using specific surveys of subsamples of the population.
In Chaptar 4, wa provide two types of demand models for western shore
beach uae activity. Each draw, on an on-site sample of beach users at
western shore beaches in the summer of 19S4. The varying parameter model
is a modification of traditional demand models where the demands for tripe to
each cite are treated largely independently, but the difference in parameters
across sites are attributed in part to site characteristic. The discrete choice
model explains the choice among cites directly, as a function of site
characteristics, but does not handle the total number of trips well. Each type
ofmodel gives a good description of one aspect ofthe recreational decision.
100
-------
From the results in Chapter 4 it is char that both money and time
access costs are extremely important in determining demand for trips to any
given beach, as are the costs of accessing alternative beaches. Whether or
not an individual owns a boat or recreational vehicle also affects demand fora
subset of beaches, those which have facilities for these capital goods. Demand
functions for trips to a site tend to pivot inward, becoming more elastic, with
declining water quality.
The results of the nested multinominal logit or discrete choice model of
beach use suggested acme similar and acme additional characteristics of
demand for this activity. Once again, money and time costs of access were
important this time in explaining the choice among sites. Additionally, the
availability of boating and recreational vehicle facilities increased the
likelihood of a boat or recreational vehicle owner to choose a site. An
interesting hypothesis was tested regarding the differential substitutability
among local beaches and among state beaches vi«-A-vi« the substitutability
between local and state beaches. Beach users seemed to consider local
beaches closer substitutes for one another than for state beaches. Individual
with larger parties or families were more likely to attend state beaches where
a variety of activities were available. The longer an individual had attended
western shore beaches, the more likely he was to use local rather than state
beaches.
Chapter 5 provides a rather extensive profile of boaters and boat owners
derived from a survey of boaters sponsored by Maryland Sea Grant and
Maryland Coastal Zone Management and from the BWSMSA telephone survey.
The boater survey subsample includes registered boat owners in Maryland.
The profile includes an analysis of characteristics which distinguish boat
owners from others and looks at these distinguishing characteristics b y
geographical area. Average household income, for example, is higher for boat
owners than non-owners, but this difference is only • tatiatically significant in
Prince Georges, Anne Arundel and Calvert counties.
Considering the boats themselves, a different profile characterizes those
which are kept in the water all season (in marinas, moored, etc. ) than
characterizes boats which are trailered. As would be expected, trailered boats
are significantly smaller and less valuable, they are more likely to be
runabouts or workboats and their owners are likely to have less income than
the owners of boats kept in the water. Almost all trailered boats were used
for fishing at least Occasionally. About three-quarters of the non-trailered
boats were used for swimming at least Occasionally.
Table 5.4 summarizes the boat owners survey by county of residence,
revealing more about the geographical distribution of Bay users. Residents of
Baltimore and Anne Arundel counties accounted for 39 pecent of the trailered
boats and 45 percent of the non-trailered boats with Prince Georges County
and Montgomery County residents accounting for another 20 percent of
trailered and 12 percent of non-trailered boats.
The last of the descriptive information suggests the importance of water
quality to boaters. Water quality was considered either moderately or very
important in the selection of a boating area by 75 percent of the trailered
boat owners and by 76 percent of the non-trailered boat owners,
101
-------
A varying parameters model similar to the one used in Chapter 4 revealed
that trailered boat owners' demand for trips from launch sites was affected by
access costs to the launch site and costs of accessing alternative sites. In
general, the demand for trips to any given site was Positively affected by the
value of the boat; i.e., individuals with more valuable boats took more trips.
The demand function fo Bny given site tended to pivot inward and become
more elastic as water quality declined.
Owners of boats kept in the water do not choose launch sites when they
take a trip, and consequently we have no way of knowing where they boat.
As a result we cannot model their decisions in responnse to varying water
quality. For these individuals? simple demand functions were estimated.
Factors which significantly affected their demand for boating trip, included
the coat of a trip (negatively) and the value of the boat (positively).
Additionally it was determined that sailboat owners tend to take fewer trips
and their demand for trips ds more price inelastic. Finally, boat owners who
fish while boating tend to demand more trips and their demand tend, to be
more price elastic.
In Chapter 6 information about • portfishing on the Bay is presented.
Estimates of sportfishing activity vary by data source and range from 539,000
to 900,000 anglers in 19SO and from 2.7 million to 4.1 million trips for that
came year. The two prominent sources efinformation on sportfishing are the
U. S. Fish and Wildlife Hunting and Fishing Survey and the U. S. National
Marine Fisheries Survey.
Our analysis in this chapter concentrated on • tripod bass fishing since
this was the only species important to Chesapeake recreational fishing for
which Sufficiently detailed data existed. One source (U. S. Fish and Wildlife)
reports that in 1960, 239,000 anglers fished for striped bass in Maryland and
Sussex County, Delaware and fished 2.1 million days in total. Table 6.1
presents come descriptive statistics of striped bass fishermen and other
Chesapeake Bay fishermen.
In the analytical section of Chapter 6, demand for sportfishing trips was
modelled as a function of the individual'. trip costs, catch rates, his annual
fishing/hunting budget and indices of types of boat ownership. All variables
affected the demand for trip, in the expected direction, with owners of
inboard motorboat likely to take more trips than those with outboard
(presumably smaller) motorboat.
3_ "? — of Benefits from Water Quality Improvements
While the analysis of the demand for recreational activities is worthwhile
in ita own right, more information about the size of rewards from Bay
restoration can be obtained, There are several reasons for computing
aggregate willingness to pay rather than simply providing descriptive
measures such as recreational use days. Obviously such measures cannot be
compared to the costs of restoration; they cannot even be added across
activities. A day of fishing is different from a day of swimming, and changes
in water quality have different effects on the benefits derived from the two
activities. Further, as we observed in Chapter 3, there is some willingness to
pay for clean water by people who do not use the Bay. If we limit ourselves
to descriptive measures such as user days, we ignore the returns to people
102
-------
who value cleaner water but do not uae it. Consequently, we have taken a
first step toward the logical, albeit venturesome, task of estimating the
aggregate benefits of improving the Bay's water quality.
Caveats
The aggregation of benefits across activities and for the population at
large is venturesome because it is so filled with known difficulties. We can
take a systematic view of these potential errors by recalling the links between
environmental policies designed to reduce effluent pollution and the benefits
of environmental improvements. Policies influence effluents directly through
regulations and indirectly through changes in incentives. Reductions in
effluents will eventually improve the ambient water quality. Improvements in
ambient quality when perceived by individuals eventually lead to changes in
behavior toward the Bay, implying benefits. Further, when non-users per-
ceive improvements in the ambient water quality, they too will be better off.
There is potential for errors in the measurement of each link in this process.
The analysis of the previous chapters has concentrated on the connection
between ambient quality and economic benefits. It rests, however, on the
relationship between environmental policy, effluents, and ambient quality. The
considerable debate regarding the connection between effluents and ambient
quality suggests the potential for honest differences of opinion on the nature
of the ecological links. Similar uncertainty over the behavioral and perception
links exists.
While a complete catalog of the sources of potential error would take an
entire chapter, we describe broadly what we think the major difficulties are.
If the problems inherent in explaining the link between policy and ambient
quality are ignored, the foremost uncertainty is between ambient quality and
behavior. Recall briefly how this link waa estimated. For boating and beach
uae we used a varying parameters model to estimate the relationship between
the product of total phosphorus and nitrogen readings in 1977 and trips in
19S4. There is clearly substantial room for error in this relationship.
First, since people cannot perceive nitrogen and phosphorus, we must
assume that the nitrogen and phosphorus are approximate measures of the
ambient quality. It is not unreasonable to expect such a relationship to hold
in principle. Chapter 2 describes ways in which individuals form perceptions
of water quality. Some of the deductive and media-baaed means by which
individuals form quality perceptions may be directly related to effluent
discharges. Others, such as stimulants of sensory perceptions, may be highly
correlated with, or even caused by, nitrogen and phosphorus levels. Previous
studies which have attempted to link behavior to individual ambient water
quality indicator. (e.g. Binckley and Hanemann) have detected a
correspondence. Chapter 2 describes acme evidence which supports this
hypothesized link derived from our telephone survey of the BWSMSA and the
field survey of western shore beaches. Through the telephone a significant
relationship waa detected between a household perception of the water quality
in the Bay and ita likelihood to quit using the Bay. Additionally, a significant
relationship appeared between objective measures ofthe Bay's water quality
over time and the proportion of households who atopped using the Bay for
recreation because they perceived the Bay's water quality to be unacceptable.
Finally, the user (field ) survey showed a positive correlation between
103
-------
measures of fecal coliform at each of nine beaches and the proportion of
households that found each beach unacceptable.
A further difficulty is the seven years which separate the
nitrogen/phosphorus readings and the recreational behavior, since 1977 _ was
the latest year for which complete information was available. While this is
clearly a source of potential error, there are a few reasons why it might not
be as bad as it seems. For one thing, the relative levels across different
regions of the Bay may have remained approximately constant even if absolute
levels have changed. Additionally, it is not clear what year or combinations of
years would be correct in signaling the recreational behavior stimulated by
water quality because behavior is probably largely affected by prior
experiences.
Since we are really explaining choices among sites of different quality,
our behavioral models depend more on the relative levels of ambient water
quality rather than on absolute levels; and if relative levels have remained
fairly constant, our behavioral models are likely to be quite good. Extracting
benefit measures from these models, however, must be done with caution since
the absolute levels of nitrogen and phosphorus readings used may not be
trustworthy y.
For recreational fishing the problem is in some waya a little simpler.
Here we use the catch rate experienced by the individual for 1980, the year
the trips were taken. There is of course a complex and uncertain chain of
relationships between improvements in ambient quality and growth in the
density of fish stocks. There is further uncertainty in the connection
between fish stocks and catch rates. These are largely, although not
completely, problems of biology and are not addressed here, but nonetheless
remain as imperfectly understood links in the system.
Restricting our comments entirely to the behavioral realm does not
eliminate these uncertainties and potential sources of modelling error. In what
sense is the catch rate in the year the trip, were taken a good measure of
quality? Fishermen may value higher catch rates but their demand (behavior)
fortripa this year may be baaed on catch rates experienced in previous
years. When the quality of the good is uncertain to the consumer, there may
be one eat of quality indicators that stimulate demand and another which
affect the benefits derived from consumption. Further, there is no guarantee
that catch rate is the only (or moat important) variable which determines the
enjoyment of trips to catch fish. For example, catching one five-pound
striped baaa may be batter than catching two two-pound stripers.
In addition to the severe difficulties in inferring the relationships
between ambient quality, there are two other significant sources of error in
computing aggregate benefits. First, there is the problem of sampling and
non-sampling error associated with the measurement of the number of trips
per participant and the number of participant in each activity, M well as
measurements of exogenous variables such as costs per trip. The boating
survey is a good example of non-sampling error for trips This survey waa a
mail survey, so in a sense the respondents are volunteers. The return rate
was 70 percent. We have no way of knowing whether those who competed
their questionnaire, were representative of the boating population as a whole
or if there is a built-in sample selection bias.
104
-------
We have also used only segments of the total population in our analysis
of benefits. The boaters were limited to those who trailer their boats, the
fishermen to those who fish for striped bass and the beach _ users to _those
who use public-accees western shore beaches. In the boating and fishing
analysis we have excluded non-Maryland households. In the contingent
valuation and beach uae analysis, only 20 percent of Virginia's population wae
included and about 80 percent of Maryland's households. In every instance, a
major portion of users is excluded so any estimates derived will be lower
bounda.
Another source of error in aggregating benefits across activities is
aggregation bias. This comes in two forms: simple doublecounting and
conceptual aggregation bias. Doublecounting occurs because a 0ubatantial
number of boaters also fish, and many fishermen have boats. The conceptual
aggregation biaa occurs because of the jointness of choice among sites for a
given activity and among activities. For example, the choice of visiting Sandy
Point versus Point Lookout may depend in part on water quality. Enhancing
water quality at both sites may only increase attendance at one sites making
the addition of benefits across sites incorrect. A discussion of this problem is
offered in Chapter 3, but both forms of aggregation bias are treated in detail
in Chapter 5 of the conceptual volume of this report.
Finally, we must remember that we have only three activities: boating,
fishing, and swimming. There are many other recreational and commercial uses
of the Bay whose value is enhanced by cleaner water. For example, our
analysis of fishing rovers only striped bass; fishing for species besides
striped bass (e.g. crabbing) is widespread and not covered by our analysis.
And our analysis of the effect of changes in water quality covers only
trailered boats, not boats at marinas. Many other, especially more casual,
activities are omitted. We have limited our analysis to boating, fishing, and
swimming because we could obtain data of adequate quality only for these
activities
Estimates
With these difficulties firmly in mind, we are prepared to hazard some
judgments on the magnitude of the aggregate benefits of improving the Bay's
water quality. Cur approach is to present low, middle and high benefits for
the beach use (Chapter 4), boating (Chapter 5), and fishing (Chapter 6) and
qualitatively compare those benefits with the total benefits derived from
Chapter 3. Comparing the ranges of these independent sources of benefits
will help us to form a judgment, but nothing more, of the magnitude of
aggregate benefits.
Chapters 4 through 6 give benefit estimates for activities conditioned on
the computational method and the proportionate change in ambient quality and
catch rate. We adopt the convention of analyzing a 20 percent reduction in
nitrogen and phosphorus for boating and beach use and a 20 percent increase
in the catch rata for striped bass fishing. These changes should be
interpreted loosely as considerable improvements in the quality of the Bay
without attaching much significance to the absolute change in ambient
readings which would be implied. In particular, one should not interpret the
estimated effect of nitrogen and phosphorus as an "all else equal" effect. The
change in nitrogen and phosphorus is a proxy for changes in moat ambient
105
-------
determinantts of water quality so that the implicit assumption is that a range
of ambient factors may be improving. Further, to counteract the problem of
aggregating across sites for a given activity, we select as a pessimistic
estimate the lowest estimate of the benefits of improving the quality by 20
percent at the one moat important site.
Table 7.1 summarizes some of the estimates of aggregate benefita for our
groups of boaters, sportfishermen and beach users, translated into 1987
dollars. The variation from pessimistic to optimistic is provided by two
sources: variation induced by the method of calculating benefits (i.e., using
actual trips versus predicted trips) and variation caused by choosing one site
rather than the sum over all sites. Recall that because each site's benefits
are calculated assuming other sites' quality remains unchanged aggregating
these measures over sites will produce an upwardly biased aggregate benefit.
The pessimistic estimates for beach use and for boating are the lower of the
two estimates of the benefits for a 20 percent improvement in water quality
from Sand y Point for beach uae and Anne Arundel County for boating. One
site was chosen as a lower bound because wtth only one site all (upward)
aggregation bias is eliminated. The average estimates for beach use and
boating are the lower of the two calculation methods for sums across all sites.
The optimistic estimates are the higher of the two calculation methods for the
sums across all sites. For striped bass fishing, the pessimistic estimate is the
lower of the two methods of calculation. The sites have already been
aggregated for the fishing case, and as we show in Chapter 5 of the
accompanying volume, the nature of the aggregation bias in this case is not
obvious. The optimistic estimate is the higher of the two calculation arithmetic
methods and the average is the mean of the pessimistic and optimistic.
Table 7.1
Aggregate Benefits for Three Water-related Activities from a
"20%" Improvement in the Chesapeake Bay's Water Quality
in 1967 dollars
Benefit Estimate
Activity
Public Western Shore Beach Usea
Boating with Trailered Boat"
Striped Bass Sportf ishingc
Pessimistic
16, 853
664
664
"Averaqe"
, . .
1 (? Thousand) "
34,658
4,717
1,366
(Mimistic
44,960
8,129
2,071
• From Table 4.6. Pessimistic estimate is the Method B value for Sandy Point,
the average is the sum of Method B values over all ten sites, and the
optimistic is the sum of Method A values over all sites.
bFrom Table 5.13. All per boater estimates expanded to 60,000 boaters
trailering boats. Pessimistic estimate is the low value (Method A) for Anne
Arundel County, the average estimate in the sum of low values (Method A)
across all counties and the optimistic value is the sum of high values (Method
B) across all counties.
GFrom Table 6.4. Pessimistic value is the value using Method B, the "average"
value is the average of the pessimistic and optimistic value, the optimistic
value is the value using Method A.
106
-------
The aggregate meaaurea of willingness to pay for water quality
improvements are revealing for several reasons. First, regardless of which
benefit measure we use (pessimistic, average, or optimistic), the returns to
beach use are the greatest. This ia primarily because a larger proportion of
the population engages in come beach-going during the year than boating Or
fishing. Additionally, this group may be more sensitive to changes in water
quality than the boating-fishing group.
A second interesting implication of the results, although not obvious from
looking at Table 7.1, is the importance of regional variation in water quality.
If we were able to clean up the water around Anne Arundel County only, we
would still go a long way towards satisfying some of the human needs for
using the Bay. While we realize that confining a water qualtty improvement
program to a particular locality may not be technically or ecological y feasible,
any clean-up strategies which result in significant improvements in this region
of the Bay will yield substantial benefits.
A comparison of the behaviorally based measures of benefits presented in
Table 7.1 with benefit estimates derived from contingent valuation (ace Table
7.2) is interesting even though the valuation questions driving the two
analyses are different. All of the estimates in Table 7,1 are partial estimates
in that they account for only one activity and involve only fubaeta of the
population. Table 7.2 presents contingent valuation produced benefit estimates
associated with a broader but less precise hypothetical improvement:
improving water quality to an "acceptable" level. The subset of the
population includes those in the BWSMSA who found water quality unacceptable
for swimming or related uses.
Table 7.2
Aggregate" Benefits from Water Quality Improvements-
Contingent Valuation
in 1984 dollars
Willingness to Pav for Improved Water Quality"
Group Pessimistic0 Average0 Optimistic0
User
Non-User
47,254
18,446
67,582
23.556
87,870
28.733
Total 65,700 91,137 116,603
•Population is the Washington, D. C. and Baltimore SMSA's
Willingness to acceept tax increase to raise Chesapeake Bay Water quality from
a level unacceptable for swimming end/or other related activities to a level
acceptable fo
The average will
See Table 3.8.
acceptable for swimming.
GTheaveragewillingnesstopay plus or minus one standard error in est imate.
107
-------
The numbers in Tables 701 and 7.2 give likely magnitudes for the annual
benefits of dirproving water quality in the Bay. The numbers suggest a range
of from $10 million to over $100 million, We know that there are numerous
random elements in all estimates. Further, we know that several activities and
populations have been omitted. But based on these numbers, it seems
plausible to estimate that the annual returns to cleaning up the Chesapeake
are at least of this order of magnitude. We have only the evidence presented
herein to make this judgment.
In conclusion, we recapitulate the premise. Society haa undertaken an
investment program. The nature of the program is the cleanup of the
Chesapeake Bay. The costs of the program include such things as sewage
treatment plants, funding of government programs to regulate and monitor
agricultural effluents, installation of industrial waste disposal systems,
restrictions on housing development, etc. The annual returns on the
investment program are measured by what people are willing to pay for the
improved services of the Bay. TMa is the dividend yielded by the public's
investment program.
For several reasons, we think that the long-run benefits are higher than
the figures Tables 7.1 and 7.2 indicate. First, as people learn that the Bay
haa become clamor, they will adjust their preferences toward Bay recreation.
This is especially true of people who do not currently use the Bay and are
largely excluded from the analyaia. Second, the population and income of the
area have grown since 1984, and both are likely to grow more, increasing the
demand for and value ofimprovements in water quality. Finally, we have
ignored the value (both uae and existence value) which households outaide the
BWSMSA may have for the Bay. The Chesapeake Bay is a nationally prominent
resource. Its improved health is of value to many who will never use it.
10s
-------
REFERENCES
Bell, F. W. 1968. "The Pope and the Price of Fish. " American Economic
Review 58142-147.
Binkley, Clark S. and W. Michael Hanemann. 1978. The Recreation Benefits of
Water Quality Improvements; An Analysis of Day Trips in an Urban
Setting. Washington, B.C. E. P.A.-600/5-78-010.
Bishop, R.C. and T.A. Heberlein. "Simulated Markets, Hypothetical Markets and
Travel Cost Analysis: Alternative Method, of Estimating Outdoor
Recreation Demand." Working paper no. 187.
Bishop, R. C. and K. Sample. 1886. "Estimating the Value of Variations in
Anglers' Success Rate.: An Application of the Multiple-Site Travel Cost
Method." Marine Resource Economics 2:1:55-74
Blackburn, A.J. 1870. "An Alternative Approach to Aggregation and Estimation
in the Non-Linear Model." Ch. 9 in The Damand for Travel; Theory and
Measurement. Edited by R E. Quandt. B.C. Heath, Lexington.
Bockstael, Nancy, W. Michael Hanemann and Catherine Kling. 1887. "Modeling
Recreational Demand in a Multiple Site Framework." Water Resources
Research. Forthcoming.
Bockstael, Nancy, W. Michael Banemann and Ivar Strand. 18SS. Measuring the
Benefits of Water Quality Improvments Using Itocreaticn Demand Models.
Report presented to the B.P.A. under Cooperative Agreement
CR-81 143-01-1, Volume II.
Bockstael, N. E., and K. E. McCormell. 1981. "Theory and Estimation of the
Household Production Function for Wildlife Recreation." Journal of
Environmental Economics and Management 8:188-214.
Bockstael, N. E., and K. E. McConnell. 1983. "Welfare Measurement in the
Household Production Framework." American Economic Review 73:806-814.
Bockstael, N. B., and J. J. Opaluch. 1983. "Discrete Modelling of Supply
Response under Uncertainty: The Case of the Fishery." Journal of
Environmental Economics and Management 10:125-137.
Bockstael, N. B., I. B. Strand, and W. M. Hanemann. 1987. "Time and The
Recreational Demand Model." American Journal of Agricultural Economics
69213-302
Bockstael, N. E. and I. E. Strand. 1987. "The Effect of Common Sources of
Regression Error on Benefit Estimates." Land Economics 63:11-20.
109
-------
Bocks tael, N. E., I. E. Strand, and A. Graefe. Economic Analysis of Artificial
Reefs: A Pilot Study of Selected Valuation Methodologies. Technical
Report #6. Artificial Reef Development Center, Sportfishing Institute.
1986.
Brookshire, D.S., L.S. Eubanks and A. Randall. . 1983. "Estimating Option
Prices and Existing Values for Wildlife Resources. " Land Economics
58:1-15.
Brown, Gard nor and Robert Mendelsohn. 1984. "The Hedonic Travel Cost
Method." Review of Economics and Statistics 66:427-33.
Brown, W. G., C. Sorhus, B. Chow-Yang, S. Richards. 1883. "Using Individual
Observations to Estimate Recreation Demand Functions." " American Journal
of Agricultural Economics 65:154-157.
Brown, W. G., C. N. Sorhus, and D. A. Meyer. 1882. "Estimation of Oregon
Sport-Caught Salmon and Steelhead Values Via the Travel Cost Method."
Mimeo, ASSA Meetings, 25 ppa.
Brown, W. G., A. Singh and E. N. Castle. "Net Economic Value of the Oregon
Salmon-Steelhead Sport Fishery." Journal of Wildlife Management
29(2):266-279.
Burt, Oscar and Durward Brewer. 1872. "Estimation of Net Social Benefits
From Outdoor Recreation." Econometrica 39:813.
Carson, R. T. and R C. Mitchell 188sO "The Value of Clean Water: The
Public'. Willingness to Pay for Beatable, Fishable, and Swimmable Quality
Water." Discussion paper No. QE85-08 revised, (Washington, D. C.,
Resources for the Future).
Castle, Emery N. and W. G. Brown. 1884. "The Economic Value of a
Recreational Resource: A Case Study of the Oregon Salmon-Steelhead
Sport Fishery. Paper presented before the Committee on the Economics
of Water Resource* Development of the Western Agricultural Economics
Research Council, San Francisco, California.
Caulkins, Peter, Richard Bishop and Nicolass Bouwes. 1982. "The Travel Coat
Model for Lake Recreation A Comparison of Two Methods for
Incorporating Site Quality and Substitution Effects." Mimeo, ABA annual
meetings,
Caulkins, Peter P., Richard C. Bishop and Nicolass W. Bouwes, Sr. 1888. "The
Travel Cost Model for Lake Recreation: A Comparison of Two Methods for
Incorporating Site Quality and Substitution Effects." Amer. J. Air. Econ.
68(2):291-297.
Caulkins, Peter, Richard Bishop and Nicolass Bouwes. 1985. "Omitted
Cross-Price Variable Biases in the Linear Travel Coat Model: Correcting
Common Misperceptions. " Land Economics 61:182-85.
110
-------
&aario, Frank J. 1969. "Operations Research in Outdoor Recreation. "
Journal of Leisure Research 1:33-52.
Cesario, Frank J. 1973. "A Generalized Trip Distribution Model." Journal of
Regional Science 13:233-48.
Cesario, Frank J. 1974. "More on the Generalized Trip Distribution Model,"
Journal of Regional Science 14:389-97
Cesario, Frank J. and Jack Knetach. 1976. "A Recreation Site Demand and
Benefit Estimation Model. " Regional Studies 10:97-104.
Christensen, L. R., Jorgenson, D.W. and Lawrence Lau. 1975. "Transcendental
Logarithmic Utility Functions." American Economic Review 65:367-83.
Cicchetti, Charles, Anthony Fisher and V. Kerry Smith. 1976. "An Econometric
Evaluation of a Generalized Consumer Surplus Measure: The Mineral King
Controversy." Econometrics 44;1268-76.
Cicchetti, C.J., J.J. Seneca and Paul Davidson. 1869. The Demand and Supply
of Outdoor Recreation: An Econometric Analysis. Rutgers. 1869.
Clawson, MartOn. 1859. Methods for Measuring the Demand for and Value of
Outdoor Recreation. Resources for the Future, Inc., Reprint no. 10.
Clawson, Marion and Jaok Knetach. 1866. Economics of Outdoor Recreation.
Resources for the Future, Washington D.C.
Cox, D.R. 1861. "Tests of Separate Families of Hypotheses." Proceedings of
the Fourth Berkeley Symposium on Mathematical Statistics and Probability.
Vol 1. University of California Press.
Cronin, F. J. 1882. "Valuing Non-Market Goods Through Contingent Markets. "
Rep. PNL-42S5. Pac. Northw. Lab., Richland, Wash.
Cropper, M. L., L.B. Deck, K.B. McConnell and T.T. Phipps. 1886. "Should the
Rosen Model be Used to Value Environmental Amenities?" Papar
presented at the conference on the Economics of Chesapeake Bay
Management 11, Annapolis, MD.
Cummings,R G. , D. S. Brookshire and W. D. Schulze. 1984. valuing
Environmental Goods: A State of the Arts Assessmentof the Contingent
Valuation Method. Draft report to EPA. Universities of New Mexico and
Wyoming. "
Daubert, J. T. and R A. Young. 1981. "Recreational Demands for Maintaining
Instream Flows: A Contingent Vaulation Approach." Amer. J. Apr. Econ.
63(4):666-676.
Beaton, A.S. 1974, "The Analysis of Consumer Demand in the United Kingdom,
1800-1970." Econometrics 42:341-67.
Ill
-------
Beaton, Henry. 1978. "Specification and Testing in Applied Demand Analysis. "
Economic Journal 88:524-36.
Desvousges, W. H. , V. K. Smith and M. P. McGivney. 1983. A Comparison of
Alternative Approaches for Estimation of Recreation and RelatedBenefits
of Water Quality Improvements. Washington, D. C.: Environmental
Protection Agency, Office of Policy Analysis.
Deuel, D. G. 1978. "National Marine Recreational Fishery Statistics Program. "
Paper presented at Canadian Sport Fisheries Conference, New Brunswick,
Canada.
Domencich, Thomas and Daniel McFadden. 1975. Urban Travel Demand; A
Behavioral Analysis. North-Holland, Amsterdam.
Dwyer, J. F. and M. D. Bowes. 1978. "Concepts of Value for Marine
Recreational Fishing." American Journal of Agricultural Economics 60(5);
1008-10120
Feenberg, Daniel and Edwin Mills. 1980. Measuring the Benefits of Water
Pollution Abatement. New York, Academic Press.
Fisher, Ann and William H. Desvousges. 1881. "Alternative Contingent Market
Approaches to Valuation of Recreation and Related Benefits of Improved
Water Quality." Paper presented at the Eastern Economics Association
Meetings, Washington B.C.
Fisher, Am and Robert Raucher. 1982. "Comparison of Alternative Methods of
Evaluating the Intrinsic Benefits of Improved Water Qualiy." Paper
presented at ASSA meetings, New York City.
Fisher, F. and K. Shell. 1871. "Taste and Quality Change in the Pure Theory
ofthe True Cost-of-Living Index." in Price Indexes and Quality Change.
Edited by Zvi Griliches. Harvard University Press, Massachusetts.
Freeman, A. Myrick III. 1979. The Benefits of Environmental Improvement:
Theory and Practice. Johns Hopkins Press for Resources for the Future,
Inc., Baltimore. 1979.
Freeman, A. Myrick III. 1882. Air and Water Pollution Control: A Benefit-Coat
Assessment. John Wiley and Sons.
Freund, R. J. and R R Wilson. 1974. "AnExample of a Gravity Model to
Estimate Recreation Travel." Journal of Leisure Research 6:241-56.
German, W.M. 1976. "Tricks with Utility Function." in Essays in Econmic
Analysis: Proceddings of the 1975 AUTE Conference. Sheffeld. Edited by
M.J. Artis and A.R. Nobay. Cambridge University Press, Cambridge.
112
-------
Greenley, Douglas A., Richard G. Walsh and Robert A. Young. 1981. "Option
Value: Empirical Evidence from a Case Study of Recreation and Water
Quality." Ouartelv J. Econ. 96(41:657-673.
Gum, R.L. and W.E. Martin. 1975. "Problems and Solutions in Estimating the
Demand for and Value of Rural Outdoor Recreation. " American Journal of
Agricultural Economics 57:558-66.
Hanemann, W. Michael. 1980. "Quality Changes, Consumer's Surplus, and
Hedonic Price Indices." California Experiment Station working paper
number 116.
Hanemann, W. Michael. 1981. "Quality and Demand Analysis."^. 3 in New
Directions in Econometric Modeling and Forecasting in U.S. Agriculture.
Edited by Gordon Rausser, Elsevier/North-Holland, New York.
Hanemann, W. Michael. 1882. "Applied Welfare Analysis with Qualitative
Response Models." California Agricultural Experiment Station working
paper no. 241.
Hanemann, W. Michael. 1984a. "Multiple Site Demand: ARew'ewofExisting
Models and New Development", Chapter 9 in Measuring the Benefits of
Water Quality Improvements Using Recreation Demand Mod«=3g- Draft
Report presented to E.P.A. under Cooperative Agreement CR-80043-01-0,
VOL II.
Hanemann, W. Michael. 1984b. "Discrete/Continuous Models of Consumer
Demand." Econometrics 52:541-61.
Hanemann, W. M. 1978. "A Methodological and Empirical Study of Recreation
Benefits from Water Quality Improvement." Ph. D. dissertation,
Department of Economics, Harvard University.
Hanemann, W. M. 1984. "Welfare Evaluations with Market Data." American
Journal of Agricultural Economics. 66:332-41.
Hanemann, W. M. 1984. "Welfare Evaluation in Contingent Valuation
Experiment with Discrete Responses." American Journal of Agricultural
Economics. VOL 66(3) :332-341.
Harmon, R. J. and Associates. 1883. Recreational Resources at Risk. Taak 2
Working Paper. Washington, DC.
Harrison, D. and D. Rubinfeld. 1878. "Hedonic Housing Prices and the Demand
for Clean Air." Journal of Environmental Economics and Management
5:81-102.
Hausman, J.A. 1886. "Exact Consumer's Surplus and Deadweight Loss."
American Economic Review 71:662-76.
113
-------
Heintz, H.T. Jr., A. Hershaft and G.C. Horak. 1976. National Damages of Air
and Water Pollution. A Report to the EPA. Enviro Control, Inc. Rockville,
MD.
Hiett, R L. and J. W. Worrall. 1977. "Marine Recreational Fisheries Ability to
Estimate Catch and to Recall Catch and Effort over Time." Human
Sciences Research, Inc., Report HSR-RR-77/18, Washington, D.C.
Horowitz, Joel. 1983. "Statistical Comparison ofNon-nested Probabilistic
Discrete Choice Models. " Transportation Science 17:319-50.
Hotelling, H. 1948. Letter to the National Park Service in "Economics of
Outdoor Recreation" - The Prewitt Report.
Huppert, D.D. 1983. NMFS Guideline, on Economic Valuation of Marine
Recreational _ Fishing. NOAA Technical Memorandum
NOAA-TM-NMFS-SWFC-32, Department of Commerce, Washington, D.C.
Johnston, J. 1972. Econometric Methods. Second Edition McGraw-Hill, New
York.
Just, Richard, Darrell Hueth and Andrew Schmits. 1982. Applied Welfare
Economics and Public Policy. Englewood Cliffs, N.J. Prentice-Hall inc.
Just, R E., and D. L. Hueth. 1979. "Multimarket Welfare Measurement."
American Economic Review 69947-64.
Kealy, Mary Jo, and Richard C. Bishop. 1983. "Theoretical and Empirical
Specification Issues in Travel Cost Demand Studies. "Unpublished.
Kendall, M. and Alan Stuart. 1977. The Advanced Theory of shaHgHrw. 4th
Edition, (Vol. 1), New York: MacMillan.
Kling, Catherine. 1986. *A Simulation Approach to Comparing Multiple Site
Recreation Demand Model. Using Chesapeake Bay Survey Data." Paper
presented at the Conference on the Economics of the Chesapeake Bay
Management II, Annapolis, MD.
Kling, Catherine, Nancy Bockstael and Ivar Strand. 1985. "An Evaluation of
Multiple Site Model Development." Paper presented at the AEA meet ings,
New York, N.Y.
Kling, Catherine. 1886. Meaaurin* the Recreational Benefits of Environmental
AmeniUa* Uain* Multiple Site Model*: An Evaluation of Techniauaa. Ph.D.
Dissertation, University of Maryland.
Krutilla, J. V. 1967. "Conservation Revisited." American Economic Review.
57( Sept. ):777-86
Lewbel, Arthur. 1975. "A Unified Approach to Incorporating Demographic or
Other Effects into Demand Systems." Review of Economic Studies 42.
114
-------
Loomis, J. 1986. "An Examination of the Variation in Empirical Estimates of
Site Specific Marginal Values for Recreational Fishermen." Working paper,
Division of Environmental Studies, University of California, Davis.
Maddala, G. S. 1983. Limited Dependent and Qualitative Variables in
Econometrics. Cambridge University Press New York.
Maddala, G. S. 1977. Econometric McGraw-Hill, Inc., N. Y. 516 pp.
Maler, Karl-Goran. 1974. Environmental Economics; A Theoretical Inquiry,
Baltimore, Johns Hopkins University Press for Resources for the Future.
Mathews, S. B. and G. S. Brown. 1970. "Economic Evaluation of the 1967
Sport Salmon Fisheries of Washington*" Washington Department of
Fisheries, Technical Report #2.
McConnell, K. E. and N. E. Bockstael. 1984. "Aggregation in Recreation
Economics: Issues of Estimation and Benefit Measurement." Northeastern
Journal of Agricultural and Resource Economics 13:181-186.
McConnell, K. E. and I. E. Strand. 1981. 'Measuring the Cost of Time in
Recreation Demand Analysis." American Journal of Agricultural Economics
63{ Feb):153-156.
McConnelltK. B. 1986. Identification of preference, in Hedonic Models.
Report presented to Environmental Protection Agency under Cooperative
Agreement CR-811043-01-0, Vol I.
McConnell, K. E. 1879. "Values of Marine Recreational Fishing: Measurement
and Impact of Measurement." American Journal of Agricultural Economics
61(5):921-923.
McConnellrK. E. and T. P. Smith. 1979. "Marine Recreational Fishing in Rhode
Island," University of Rhode Island Marine Memorandum #62.
McFadden, Daniel. 1874. "Conditional Logit Analysis of Qualitative Choice
Behavior." in Frontiers in Econometrics. Edited by P. Zarembka.
McFadden, D. 1978. "Modelling the Choice of Residential Location. "
In Karlqjist, Lundquist, Snickars, and Weibull (eds) Spatial Interaction
Theory and Planning Models. New York North-Holland Publishing
company.
Mendelsohn, Robert. 1984. "Estimating the Structural Equations of Implicit
Markets and Household Production Functions." Review of Economics and
Statistics 66:673-77.
Meyer, P. A. 1982. Net Economic Values for Salmon and Steelhead from the
Columbia River Sustem. NOM Technical Memorandum NMFS F/NWR-3, U.
S. Department of Commerce, Washington, D.C.
115
-------
Miller, Jon. 1982. "Valuation Methods for Fish and Wildlife Associated
Recreation: An Empirical Comparison of the Travel Coat and Contingent
Valuation Method-." Paper presented at the ASSA meetings, New York,
N.Y.
Mitchell, R C. and Richard T. Carson. 1984. "National Fresh Water Quality
Benefits: Findings from a New National Contingent Valuation Survey."
Presented at the annual AERE meeting, Cornell University.
Morey, Edward. 1981. "The Demand for Site-Specific Recreational Activities: A
Charactaristics Approach." Journal of Environmental Economics and
Management 8:345-71.
Morey, Edward. 1984a. "Characteristics, Consumer Surplus and New Activities:
A Proposed Ski Area." Paper presented at the ASSA meetings, Dallas,
TX.
Morey, Edward. 1984b. "The Choice of Ski Areas: Estimation of a Generalized
CBS Preference Ordering with Characteristics." Review ofEconomics and
Statistics 66:584-90.
Muellbauer, J. "Household Composition, Engel Curvea and Welfare Comparisons
Between Households A Duality Approach." European Economic Review
5(1974):103-22.
JVorton, V. J., T. P. Smith, and I. E. Strand. 1889. Stripers: The Economic
Value ofthe Atlantic Coast Commercial and Recreational Striped Bass
Fisheries. Maryland Sea Grant Publication #UM-SG-TS-83-12.
Pacific Fisheries Management Council. 1886. Review of 1985 Ocean Salmon
Fisheries. Portland, Oregon.
Parks, RS. 1969, "Systems of Demand Equations: An Enpirical Comparison of
Alternative Functional Forma." Econometrics 37:629-65
Pesaran, M.H. and A.S. Deaton. 1978. "Testing Non-nested Non-linear
Regression Models." Econometricca 46:677-94.
Pindyck, Robert and Daniel Rubinfeld. 1981. Econometric Models and Economic
Forecasts. New York, McGraw-Hill Book Company.
Pow, Robert. 1971. "Additive Utility Functions and Linear Engel Curves. "
Review of Bconoaic Studies 38401-14.
Pollak, Robert and Terence Wales. 19790 "Estimation of the Linear
Expenditure System." Econometrics 37:611-29.
Pollak, Robert and Terence Wales. 1881. "Demographic Variables in Demand
Analysis." Econometrics 4*1533-51.
116
-------
Reiling, S. D., K.C. Gibbs and H.H. Stoevner. 1973. Economic Benefits from an
Improvement in Water Quality. Washington,D.C., U.S. EPA.
Ribaudo, Marc 0. and Donald J. Epp. 1983. "The Importance of Sample
Discrimination in Using the Travel Cost Method to Estimate the Benefits of
Improved Water Quality." Paper presented at the annual AAEA meeting,
Purdue, Indiana.
Rosen, Sherwin. 1974. "Hedonic Prices and Implicit Markets: Product
Differentiation in Pure Competition." Journal ofPolitical Economy
82:34-55.
Rowe, Robert, Edward Morey, Arthur Ross and W. Douglass Shaw. 1985.
Valuing Marine Recreation Fishing on the Pacific Coast. Energy and
Resurce Consultants, Report prepared for National Marine Fisheries
Service, NOAA.
Russell, C. S., and W. A. Vaughan. 1982. "The National Recreational Fishing
Benefits of Water Pollution Control." Journal of Environmental Economics
and Management 9:328-354.
Samples, K. C., and R. C. Bishop. (Undated). "Estimating Social Value, of
Sport Caught Fish A Suggested Approach." Draft, 12 pps.
Samuelson, Paul A. 1947. "Some Implications of Linearity," Review of
Economic Studies 15:88-90.
Saxonhouse, G.R. 1977. "Regressions from Samples Having Different
Character!-tics." Review of Economics and Statistics 59294-7.
Schulse, W. D., D. S. Brookshire, B. G. Walther, M. K. McFarland, M. A. Thayer,
R. L. Whitworth, S. Ben-David, W. Maim, and J. Molenar. 19S3. "The
Economic Benefits of Preserving Visibility in the National Parklands of the
southwest." Natural Resources Journal 23:149-173.
Sellar, Christine, John Stoll and Jean-Paul Chavas. 196s. "Validation of
Empirical Measures of Welfare Change: A Comparison of Nonmarket
Techniques. " Land Economics 61:156-75.
Sellar, C., J. P. Chavas and J. R. Stoll. 1966. "Specification of the Logit
Model: The Case of Valuation of Nonmarket Goods." Journal 'of
Environmental Economics and Manaoment Vol. 13:382-390.
Smith, V. Kerry and William H. Deavousge. 1985. "The Generalized Travel
Cost Model and Water Quality Benefits: A Reconsideration." Southern
Economic Journal 62371-81.
Smith, V. K., W. H. Desvousges and Ann Fisher. 1986. "A Comparison of Direct
and Indirect Methods for Estimating Environmental Benefits." Amer. J.
Aar. Econ. 68 (2) :280-290.
117
-------
Smith, V. Kerry, William Ho Desvousges and M.P. McGivney. 1983. "Estimating
Water Quality Benefits: An Econometric Analysis. " Southern Economic
-50:422-37"
Smith, V. K., W. H. Desvousges and M. P. McGivney. 1983. "The Opportunity
Cost of Travel Time in Recreation Demand. Models." Land Economics.
59:3:259-78.
Smith, V. Kerry and Yoshiaki Kaoru. 1985. "The Hedonic Travel Cast Model: A
View from the Trenches." Mimeo.
Stevens, J. B. 1966. "Recreation Benefits from Water Pollution Control. "
Water Resources Research 2:167-182.
Stevens, Joe B. 1866. "Angler Success as a Quality Determinant of Sport
Fishery Recreational Values. " Transactions of the American Fisheries
Society. 95:357-362.
Stoevener, H. H. 1863. "An Economic Evaluation of Water Pollution Control
Alternatives: A Progress Report." Water Resources and Economic
Development of the West, No. 12, Committee on the Economics of Water
Resources Development of the Western Agricultural Economic. Research
Council, Salt Lake City, 47-59.
Stoevener, H. H. 1864. "Water uee Relationships as Affected by Water Quality
on the Yaquina Bay." New Horizons for Resources Research: Issues and
Methodology. Western Resources Conference, University of Colorado Press,
Boulder, 87-86.
Stone, J.R.M. 1854. The Measurement of Consumers' Expenditure and
Behaviour in the United Kingdom. 1920-1938, Vol L Cambridge University
Press, Cambridge.
Strand, I. E. and N. E. Bockstael. 1985. "The Effect of Common Sources of
Regression Error on Benefit Estimation." Working paper, University of
Maryland.
Sutherland, R.J. 1881. "A Regional Approach to Estimating Recreation Benefits
of Improved Water Quality. " Journal of Environmental Economics and
Management 9:229-47.
Thayer, Mark. 1981. "Contingent Valuation Techniques for Assessing
Environmental Impacts: Further Evidence." Journal of Environmental
Economic* and Management 8:27-44.
Theil, Henry. 1976. Theory and Measurement of Consumer Demand. Vols 1 and
2. Amsterdam.
Tobin, James. 1985. "Estimation of Relationships for Limited Dependent
Variables." Econometrics 26:24-36.
118
-------
u. S. EPA. 1983. Water Quality Standards Handbook. office ofWater
Regulation, and Standards. Washington, DC.
U. S. Department of Commerce. 1984. "Marine Recreational Fishery Statistics
Survey, Atlantic and Gulf Coasts, 1979 (Revised) - 1980, Current Fishery
Statistics 8322." Washington, DC.
u, S. Department of the Interior. "1980 National Survey of Fishing, Hunting,
and Wildlife-Associated Recreation." U. S. Fish and Wildlife Service.
Washington, DC. Undated.
Vaug ban, W. J. and C. S. Russell. 1982. "Valuing a Fishing Day: An
Application of a Systematic Varying Parameter Model." Land Economics
58:450-62.
Vaughan, W. J. and C. S. Russell. 1982. Fresh Water Recretional Fishing; The
National Benefits of Water Pollution Control. Washington D. C.:RFF, Inc.
Vaughan, W. J., C. S. Russell, L. P. Gianessi and L. A. Nielsen. "Measuring and
Predicting Water Quality in Recreation Related Terms." J." Env. Mornt.
15:363-380.
Violette, Daniel. 188s. "A Model to Estimate the Economic Impacts on
Recreational Fishing in the Adirondacks from Current Levels of
Acidification." Paper presented at the AERE Workshop on Recreation
Demand Modeling, Boulder, CO.
Wales, Terence J. 1971. "A Generalised Linear Expenditure Model of the
Demand forNon-Durable Goods in Canada." Canadian Journal of Eonomics
4:471-84.
Wales, Terence and A. D. Woodland. 1883. "Estimation of Consumer Demand
Systems with Binding Non-Negativity Constraints." Journal of
Econometrics 21:263-85.
Walsh, Richard G. Undated. '*An Extension of the Contingent Valuation
Approach to Estimate the Effect oflnstream Flow on Recreational
Demand." Unpublished.
Walsh, R G., D. A. Greenley, R A. Young, J. RMcKean, anc&. A. Pinto. 1978.
"Option Values and Recreational Benefits of Improved Water Quality: A
Case Study of the South Platte River Basin, Colorado." Rep.
EPA-600/5-78-100. U. S. Environmental Protection Agency, Research
Triangle Park, N. C.
Walsh, R G., L. D. Sanders, and J. B. Loomis. 1985. Wild and Scenic River
Economics: Recreation Use and Preservation Values. American Wilderness
Alliance, Englewood, Colorado.
119
-------
Wegge, Thomas, W. Michael Hanemann and Ivar Strand. 1985. "An Economic
Assessment of Marine Recreational Fishing in Southern California."
Report subritbed to National Marine Fisheries Service Southwest Regional
Office.
Wennergren, E. B. and D. B. Nielson. 1970. ' "Probability Estimates of
Recreation Demands." Journal of Leisure Research 2:112-22.
Williams, J. B., T. P. Smith, M. J. Speir, and S. Early. 1983. "1980 Maryland
Saltwater Sport Fishing Survey. " Maryland Department of Natural
Resources Tidewater Administration Report No. TA-CRD-83-1. Annapolis,
MD.
Willig, R.D. 1976. "Consumer's Surplus without Apology." American Economic
Review 66:589-97.
Willig, R. D. 1978. "Increment Consumer's Surplus and Hedonic Price
Adjustments." Journal o/Economic Theory 17:227-253.
Wilman, Elizabeth A. 1884. External Costa of Coastal Beach Pollution.
Resources for the Future, Washington, DC.
Wilman, Elizabeth. 1985. "Valuing Quality Changes in Recreation Resources.™
Paper presented at the AERE Workshop on Recreation Demand Modeling,
Boulder, CO.
Zellner, A . and Soo-Bin Park. 1879. "Minimum Expected Lose (MELO)
Estimators for Functions of Parameters and Structural Coeffiecients of
Econometric Models." Joimial of the American Statistical Association.
74:365:185-193.
120
-------
REFERENCES
Bell, F. W. 1968. "The Pope and the Price of Fish. " American Economic
Review 58142-147.
Binkley, Clark S. and W. Michael Hanemann. 1978. The Recreation Benefits of
Water Quality Improvements; An Analysis of Day Trips in an Urban
Setting. Washington, B.C. E. P.A.-600/5-78-010.
Bishop, R.C. and T.A. Heberlein. "Simulated Markets, Hypothetical Markets and
Travel Cost Analysis: Alternative Method, of Estimating Outdoor
Recreation Demand." Working paper no. 187.
Bishop, R. C. and K. Sample. 1886. "Estimating the Value of Variations in
Anglers' Success Rate.: An Application of the Multiple-Site Travel Cost
Method." Marine Resource Economics 2:1:55-74
Blackburn, A.J. 1870. "An Alternative Approach to Aggregation and Estimation
in the Non-Linear Model." Ch. 9 in The Damand for Travel; Theory and
Measurement. Edited by R E. Quandt. B.C. Heath, Lexington.
Bockstael, Nancy, W. Michael Hanemann and Catherine Kling. 1887. "Modeling
Recreational Demand in a Multiple Site Framework." Water Resources
Research. Forthcoming.
Bockstael, Nancy, W. Michael Banemann and Ivar Strand. 18SS. Measuring the
Benefits of Water Quality Improvments Using Itocreaticn Demand Models.
Report presented to the B.P.A. under Cooperative Agreement
CR-81 143-01-1, Volume II.
Bockstael, N. E., and K. E. McCormell. 1981. "Theory and Estimation of the
Household Production Function for Wildlife Recreation." Journal of
Environmental Economics and Management 8:188-214.
Bockstael, N. E., and K. E. McConnell. 1983. "Welfare Measurement in the
Household Production Framework." American Economic Review 73:806-814.
Bockstael, N. B., and J. J. Opaluch. 1983. "Discrete Modelling of Supply
Response under Uncertainty: The Case of the Fishery." Journal of
Environmental Economics and Management 10:125-137.
Bockstael, N. B., I. B. Strand, and W. M. Hanemann. 1987. "Time and The
Recreational Demand Model." American Journal of Agricultural Economics
69213-302
Bockstael, N. E. and I. E. Strand. 1987. "The Effect of Common Sources of
Regression Error on Benefit Estimates." Land Economics 63:11-20.
109
-------
Bocks tael, N. E., I. E. Strand, and A. Graefe. Economic Analysis of Artificial
Reefs: A Pilot Study of Selected Valuation Methodologies. Technical
Report #6. Artificial Reef Development Center, Sportfishing Institute.
1986.
Brookshire, D.S., L.S. Eubanks and A. Randall. . 1983. "Estimating Option
Prices and Existing Values for Wildlife Resources. " Land Economics
58:1-15.
Brown, Gard nor and Robert Mendelsohn. 1984. "The Hedonic Travel Cost
Method." Review of Economics and Statistics 66:427-33.
Brown, W. G., C. Sorhus, B. Chow-Yang, S. Richards. 1883. "Using Individual
Observations to Estimate Recreation Demand Functions." " American Journal
of Agricultural Economics 65:154-157.
Brown, W. G., C. N. Sorhus, and D. A. Meyer. 1882. "Estimation of Oregon
Sport-Caught Salmon and Steelhead Values Via the Travel Cost Method."
Mimeo, ASSA Meetings, 25 ppa.
Brown, W. G., A. Singh and E. N. Castle. "Net Economic Value of the Oregon
Salmon-Steelhead Sport Fishery." Journal of Wildlife Management
29(2):266-279.
Burt, Oscar and Durward Brewer. 1872. "Estimation of Net Social Benefits
From Outdoor Recreation." Econometrica 39:813.
Carson, R. T. and R C. Mitchell 188sO "The Value of Clean Water: The
Public'. Willingness to Pay for Beatable, Fishable, and Swimmable Quality
Water." Discussion paper No. QE85-08 revised, (Washington, D. C.,
Resources for the Future).
Castle, Emery N. and W. G. Brown. 1884. "The Economic Value of a
Recreational Resource: A Case Study of the Oregon Salmon-Steelhead
Sport Fishery. Paper presented before the Committee on the Economics
of Water Resource* Development of the Western Agricultural Economics
Research Council, San Francisco, California.
Caulkins, Peter, Richard Bishop and Nicolass Bouwes. 1982. "The Travel Coat
Model for Lake Recreation A Comparison of Two Methods for
Incorporating Site Quality and Substitution Effects." Mimeo, ABA annual
meetings,
Caulkins, Peter P., Richard C. Bishop and Nicolass W. Bouwes, Sr. 1888. "The
Travel Cost Model for Lake Recreation: A Comparison of Two Methods for
Incorporating Site Quality and Substitution Effects." Amer. J. Air. Econ.
68(2):291-297.
Caulkins, Peter, Richard Bishop and Nicolass Bouwes. 1985. "Omitted
Cross-Price Variable Biases in the Linear Travel Coat Model: Correcting
Common Misperceptions. " Land Economics 61:182-85.
110
-------
&aario, Frank J. 1969. "Operations Research in Outdoor Recreation. "
Journal of Leisure Research 1:33-52.
Cesario, Frank J. 1973. "A Generalized Trip Distribution Model." Journal of
Regional Science 13:233-48.
Cesario, Frank J. 1974. "More on the Generalized Trip Distribution Model,"
Journal of Regional Science 14:389-97
Cesario, Frank J. and Jack Knetach. 1976. "A Recreation Site Demand and
Benefit Estimation Model. " Regional Studies 10:97-104.
Christensen, L. R., Jorgenson, D.W. and Lawrence Lau. 1975. "Transcendental
Logarithmic Utility Functions." American Economic Review 65:367-83.
Cicchetti, Charles, Anthony Fisher and V. Kerry Smith. 1976. "An Econometric
Evaluation of a Generalized Consumer Surplus Measure: The Mineral King
Controversy." Econometrics 44;1268-76.
Cicchetti, C.J., J.J. Seneca and Paul Davidson. 1869. The Demand and Supply
of Outdoor Recreation: An Econometric Analysis. Rutgers. 1869.
Clawson, MartOn. 1859. Methods for Measuring the Demand for and Value of
Outdoor Recreation. Resources for the Future, Inc., Reprint no. 10.
Clawson, Marion and Jaok Knetach. 1866. Economics of Outdoor Recreation.
Resources for the Future, Washington D.C.
Cox, D.R. 1861. "Tests of Separate Families of Hypotheses." Proceedings of
the Fourth Berkeley Symposium on Mathematical Statistics and Probability.
Vol 1. University of California Press.
Cronin, F. J. 1882. "Valuing Non-Market Goods Through Contingent Markets. "
Rep. PNL-42S5. Pac. Northw. Lab., Richland, Wash.
Cropper, M. L., L.B. Deck, K.B. McConnell and T.T. Phipps. 1886. "Should the
Rosen Model be Used to Value Environmental Amenities?" Papar
presented at the conference on the Economics of Chesapeake Bay
Management 11, Annapolis, MD.
Cummings,R G. , D. S. Brookshire and W. D. Schulze. 1984. valuing
Environmental Goods: A State of the Arts Assessmentof the Contingent
Valuation Method. Draft report to EPA. Universities of New Mexico and
Wyoming. "
Daubert, J. T. and R A. Young. 1981. "Recreational Demands for Maintaining
Instream Flows: A Contingent Vaulation Approach." Amer. J. Apr. Econ.
63(4):666-676.
Beaton, A.S. 1974, "The Analysis of Consumer Demand in the United Kingdom,
1800-1970." Econometrics 42:341-67.
Ill
-------
Beaton, Henry. 1978. "Specification and Testing in Applied Demand Analysis. "
Economic Journal 88:524-36.
Desvousges, W. H. , V. K. Smith and M. P. McGivney. 1983. A Comparison of
Alternative Approaches for Estimation of Recreation and RelatedBenefits
of Water Quality Improvements. Washington, D. C.: Environmental
Protection Agency, Office of Policy Analysis.
Deuel, D. G. 1978. "National Marine Recreational Fishery Statistics Program. "
Paper presented at Canadian Sport Fisheries Conference, New Brunswick,
Canada.
Domencich, Thomas and Daniel McFadden. 1975. Urban Travel Demand; A
Behavioral Analysis. North-Holland, Amsterdam.
Dwyer, J. F. and M. D. Bowes. 1978. "Concepts of Value for Marine
Recreational Fishing." American Journal of Agricultural Economics 60(5);
1008-10120
Feenberg, Daniel and Edwin Mills. 1980. Measuring the Benefits of Water
Pollution Abatement. New York, Academic Press.
Fisher, Ann and William H. Desvousges. 1881. "Alternative Contingent Market
Approaches to Valuation of Recreation and Related Benefits of Improved
Water Quality." Paper presented at the Eastern Economics Association
Meetings, Washington B.C.
Fisher, Am and Robert Raucher. 1982. "Comparison of Alternative Methods of
Evaluating the Intrinsic Benefits of Improved Water Qualiy." Paper
presented at ASSA meetings, New York City.
Fisher, F. and K. Shell. 1871. "Taste and Quality Change in the Pure Theory
ofthe True Cost-of-Living Index." in Price Indexes and Quality Change.
Edited by Zvi Griliches. Harvard University Press, Massachusetts.
Freeman, A. Myrick III. 1979. The Benefits of Environmental Improvement:
Theory and Practice. Johns Hopkins Press for Resources for the Future,
Inc., Baltimore. 1979.
Freeman, A. Myrick III. 1882. Air and Water Pollution Control: A Benefit-Coat
Assessment. John Wiley and Sons.
Freund, R. J. and R R Wilson. 1974. "AnExample of a Gravity Model to
Estimate Recreation Travel." Journal of Leisure Research 6:241-56.
German, W.M. 1976. "Tricks with Utility Function." in Essays in Econmic
Analysis: Proceddings of the 1975 AUTE Conference. Sheffeld. Edited by
M.J. Artis and A.R. Nobay. Cambridge University Press, Cambridge.
112
-------
Greenley, Douglas A., Richard G. Walsh and Robert A. Young. 1981. "Option
Value: Empirical Evidence from a Case Study of Recreation and Water
Quality." Ouartelv J. Econ. 96(41:657-673.
Gum, R.L. and W.E. Martin. 1975. "Problems and Solutions in Estimating the
Demand for and Value of Rural Outdoor Recreation. " American Journal of
Agricultural Economics 57:558-66.
Hanemann, W. Michael. 1980. "Quality Changes, Consumer's Surplus, and
Hedonic Price Indices." California Experiment Station working paper
number 116.
Hanemann, W. Michael. 1981. "Quality and Demand Analysis."^. 3 in New
Directions in Econometric Modeling and Forecasting in U.S. Agriculture.
Edited by Gordon Rausser, Elsevier/North-Holland, New York.
Hanemann, W. Michael. 1882. "Applied Welfare Analysis with Qualitative
Response Models." California Agricultural Experiment Station working
paper no. 241.
Hanemann, W. Michael. 1984a. "Multiple Site Demand: ARew'ewofExisting
Models and New Development", Chapter 9 in Measuring the Benefits of
Water Quality Improvements Using Recreation Demand Mod«=3g- Draft
Report presented to E.P.A. under Cooperative Agreement CR-80043-01-0,
VOL II.
Hanemann, W. Michael. 1984b. "Discrete/Continuous Models of Consumer
Demand." Econometrics 52:541-61.
Hanemann, W. M. 1978. "A Methodological and Empirical Study of Recreation
Benefits from Water Quality Improvement." Ph. D. dissertation,
Department of Economics, Harvard University.
Hanemann, W. M. 1984. "Welfare Evaluations with Market Data." American
Journal of Agricultural Economics. 66:332-41.
Hanemann, W. M. 1984. "Welfare Evaluation in Contingent Valuation
Experiment with Discrete Responses." American Journal of Agricultural
Economics. VOL 66(3) :332-341.
Harmon, R. J. and Associates. 1883. Recreational Resources at Risk. Taak 2
Working Paper. Washington, DC.
Harrison, D. and D. Rubinfeld. 1878. "Hedonic Housing Prices and the Demand
for Clean Air." Journal of Environmental Economics and Management
5:81-102.
Hausman, J.A. 1886. "Exact Consumer's Surplus and Deadweight Loss."
American Economic Review 71:662-76.
113
-------
Heintz, H.T. Jr., A. Hershaft and G.C. Horak. 1976. National Damages of Air
and Water Pollution. A Report to the EPA. Enviro Control, Inc. Rockville,
MD.
Hiett, R L. and J. W. Worrall. 1977. "Marine Recreational Fisheries Ability to
Estimate Catch and to Recall Catch and Effort over Time." Human
Sciences Research, Inc., Report HSR-RR-77/18, Washington, D.C.
Horowitz, Joel. 1983. "Statistical Comparison ofNon-nested Probabilistic
Discrete Choice Models. " Transportation Science 17:319-50.
Hotelling, H. 1948. Letter to the National Park Service in "Economics of
Outdoor Recreation" - The Prewitt Report.
Huppert, D.D. 1983. NMFS Guideline, on Economic Valuation of Marine
Recreational _ Fishing. NOAA Technical Memorandum
NOAA-TM-NMFS-SWFC-32, Department of Commerce, Washington, D.C.
Johnston, J. 1972. Econometric Methods. Second Edition McGraw-Hill, New
York.
Just, Richard, Darrell Hueth and Andrew Schmits. 1982. Applied Welfare
Economics and Public Policy. Englewood Cliffs, N.J. Prentice-Hall inc.
Just, R E., and D. L. Hueth. 1979. "Multimarket Welfare Measurement."
American Economic Review 69947-64.
Kealy, Mary Jo, and Richard C. Bishop. 1983. "Theoretical and Empirical
Specification Issues in Travel Cost Demand Studies. "Unpublished.
Kendall, M. and Alan Stuart. 1977. The Advanced Theory of shaHgHrw. 4th
Edition, (Vol. 1), New York: MacMillan.
Kling, Catherine. 1986. *A Simulation Approach to Comparing Multiple Site
Recreation Demand Model. Using Chesapeake Bay Survey Data." Paper
presented at the Conference on the Economics of the Chesapeake Bay
Management II, Annapolis, MD.
Kling, Catherine, Nancy Bockstael and Ivar Strand. 1985. "An Evaluation of
Multiple Site Model Development." Paper presented at the AEA meet ings,
New York, N.Y.
Kling, Catherine. 1886. Meaaurin* the Recreational Benefits of Environmental
AmeniUa* Uain* Multiple Site Model*: An Evaluation of Techniauaa. Ph.D.
Dissertation, University of Maryland.
Krutilla, J. V. 1967. "Conservation Revisited." American Economic Review.
57( Sept. ):777-86
Lewbel, Arthur. 1975. "A Unified Approach to Incorporating Demographic or
Other Effects into Demand Systems." Review of Economic Studies 42.
114
-------
Loomis, J. 1986. "An Examination of the Variation in Empirical Estimates of
Site Specific Marginal Values for Recreational Fishermen." Working paper,
Division of Environmental Studies, University of California, Davis.
Maddala, G. S. 1983. Limited Dependent and Qualitative Variables in
Econometrics. Cambridge University Press New York.
Maddala, G. S. 1977. Econometric McGraw-Hill, Inc., N. Y. 516 pp.
Maler, Karl-Goran. 1974. Environmental Economics; A Theoretical Inquiry,
Baltimore, Johns Hopkins University Press for Resources for the Future.
Mathews, S. B. and G. S. Brown. 1970. "Economic Evaluation of the 1967
Sport Salmon Fisheries of Washington*" Washington Department of
Fisheries, Technical Report #2.
McConnell, K. E. and N. E. Bockstael. 1984. "Aggregation in Recreation
Economics: Issues of Estimation and Benefit Measurement." Northeastern
Journal of Agricultural and Resource Economics 13:181-186.
McConnell, K. E. and I. E. Strand. 1981. 'Measuring the Cost of Time in
Recreation Demand Analysis." American Journal of Agricultural Economics
63{ Feb):153-156.
McConnelltK. B. 1986. Identification of preference, in Hedonic Models.
Report presented to Environmental Protection Agency under Cooperative
Agreement CR-811043-01-0, Vol I.
McConnell, K. E. 1879. "Values of Marine Recreational Fishing: Measurement
and Impact of Measurement." American Journal of Agricultural Economics
61(5):921-923.
McConnellrK. E. and T. P. Smith. 1979. "Marine Recreational Fishing in Rhode
Island," University of Rhode Island Marine Memorandum #62.
McFadden, Daniel. 1874. "Conditional Logit Analysis of Qualitative Choice
Behavior." in Frontiers in Econometrics. Edited by P. Zarembka.
McFadden, D. 1978. "Modelling the Choice of Residential Location. "
In Karlqjist, Lundquist, Snickars, and Weibull (eds) Spatial Interaction
Theory and Planning Models. New York North-Holland Publishing
company.
Mendelsohn, Robert. 1984. "Estimating the Structural Equations of Implicit
Markets and Household Production Functions." Review of Economics and
Statistics 66:673-77.
Meyer, P. A. 1982. Net Economic Values for Salmon and Steelhead from the
Columbia River Sustem. NOM Technical Memorandum NMFS F/NWR-3, U.
S. Department of Commerce, Washington, D.C.
115
-------
Miller, Jon. 1982. "Valuation Methods for Fish and Wildlife Associated
Recreation: An Empirical Comparison of the Travel Coat and Contingent
Valuation Method-." Paper presented at the ASSA meetings, New York,
N.Y.
Mitchell, R C. and Richard T. Carson. 1984. "National Fresh Water Quality
Benefits: Findings from a New National Contingent Valuation Survey."
Presented at the annual AERE meeting, Cornell University.
Morey, Edward. 1981. "The Demand for Site-Specific Recreational Activities: A
Charactaristics Approach." Journal of Environmental Economics and
Management 8:345-71.
Morey, Edward. 1984a. "Characteristics, Consumer Surplus and New Activities:
A Proposed Ski Area." Paper presented at the ASSA meetings, Dallas,
TX.
Morey, Edward. 1984b. "The Choice of Ski Areas: Estimation of a Generalized
CBS Preference Ordering with Characteristics." Review ofEconomics and
Statistics 66:584-90.
Muellbauer, J. "Household Composition, Engel Curvea and Welfare Comparisons
Between Households A Duality Approach." European Economic Review
5(1974):103-22.
JVorton, V. J., T. P. Smith, and I. E. Strand. 1889. Stripers: The Economic
Value ofthe Atlantic Coast Commercial and Recreational Striped Bass
Fisheries. Maryland Sea Grant Publication #UM-SG-TS-83-12.
Pacific Fisheries Management Council. 1886. Review of 1985 Ocean Salmon
Fisheries. Portland, Oregon.
Parks, RS. 1969, "Systems of Demand Equations: An Enpirical Comparison of
Alternative Functional Forma." Econometrics 37:629-65
Pesaran, M.H. and A.S. Deaton. 1978. "Testing Non-nested Non-linear
Regression Models." Econometricca 46:677-94.
Pindyck, Robert and Daniel Rubinfeld. 1981. Econometric Models and Economic
Forecasts. New York, McGraw-Hill Book Company.
Pow, Robert. 1971. "Additive Utility Functions and Linear Engel Curves. "
Review of Bconoaic Studies 38401-14.
Pollak, Robert and Terence Wales. 19790 "Estimation of the Linear
Expenditure System." Econometrics 37:611-29.
Pollak, Robert and Terence Wales. 1881. "Demographic Variables in Demand
Analysis." Econometrics 4*1533-51.
116
-------
Reiling, S. D., K.C. Gibbs and H.H. Stoevner. 1973. Economic Benefits from an
Improvement in Water Quality. Washington,D.C., U.S. EPA.
Ribaudo, Marc 0. and Donald J. Epp. 1983. "The Importance of Sample
Discrimination in Using the Travel Cost Method to Estimate the Benefits of
Improved Water Quality." Paper presented at the annual AAEA meeting,
Purdue, Indiana.
Rosen, Sherwin. 1974. "Hedonic Prices and Implicit Markets: Product
Differentiation in Pure Competition." Journal ofPolitical Economy
82:34-55.
Rowe, Robert, Edward Morey, Arthur Ross and W. Douglass Shaw. 1985.
Valuing Marine Recreation Fishing on the Pacific Coast. Energy and
Resurce Consultants, Report prepared for National Marine Fisheries
Service, NOAA.
Russell, C. S., and W. A. Vaughan. 1982. "The National Recreational Fishing
Benefits of Water Pollution Control." Journal of Environmental Economics
and Management 9:328-354.
Samples, K. C., and R. C. Bishop. (Undated). "Estimating Social Value, of
Sport Caught Fish A Suggested Approach." Draft, 12 pps.
Samuelson, Paul A. 1947. "Some Implications of Linearity," Review of
Economic Studies 15:88-90.
Saxonhouse, G.R. 1977. "Regressions from Samples Having Different
Character!-tics." Review of Economics and Statistics 59294-7.
Schulse, W. D., D. S. Brookshire, B. G. Walther, M. K. McFarland, M. A. Thayer,
R. L. Whitworth, S. Ben-David, W. Maim, and J. Molenar. 19S3. "The
Economic Benefits of Preserving Visibility in the National Parklands of the
southwest." Natural Resources Journal 23:149-173.
Sellar, Christine, John Stoll and Jean-Paul Chavas. 196s. "Validation of
Empirical Measures of Welfare Change: A Comparison of Nonmarket
Techniques. " Land Economics 61:156-75.
Sellar, C., J. P. Chavas and J. R. Stoll. 1966. "Specification of the Logit
Model: The Case of Valuation of Nonmarket Goods." Journal 'of
Environmental Economics and Manaoment Vol. 13:382-390.
Smith, V. Kerry and William H. Deavousge. 1985. "The Generalized Travel
Cost Model and Water Quality Benefits: A Reconsideration." Southern
Economic Journal 62371-81.
Smith, V. K., W. H. Desvousges and Ann Fisher. 1986. "A Comparison of Direct
and Indirect Methods for Estimating Environmental Benefits." Amer. J.
Aar. Econ. 68 (2) :280-290.
117
-------
Smith, V. Kerry, William Ho Desvousges and M.P. McGivney. 1983. "Estimating
Water Quality Benefits: An Econometric Analysis. " Southern Economic
-50:422-37"
Smith, V. K., W. H. Desvousges and M. P. McGivney. 1983. "The Opportunity
Cost of Travel Time in Recreation Demand. Models." Land Economics.
59:3:259-78.
Smith, V. Kerry and Yoshiaki Kaoru. 1985. "The Hedonic Travel Cast Model: A
View from the Trenches." Mimeo.
Stevens, J. B. 1966. "Recreation Benefits from Water Pollution Control. "
Water Resources Research 2:167-182.
Stevens, Joe B. 1866. "Angler Success as a Quality Determinant of Sport
Fishery Recreational Values. " Transactions of the American Fisheries
Society. 95:357-362.
Stoevener, H. H. 1863. "An Economic Evaluation of Water Pollution Control
Alternatives: A Progress Report." Water Resources and Economic
Development of the West, No. 12, Committee on the Economics of Water
Resources Development of the Western Agricultural Economic. Research
Council, Salt Lake City, 47-59.
Stoevener, H. H. 1864. "Water uee Relationships as Affected by Water Quality
on the Yaquina Bay." New Horizons for Resources Research: Issues and
Methodology. Western Resources Conference, University of Colorado Press,
Boulder, 87-86.
Stone, J.R.M. 1854. The Measurement of Consumers' Expenditure and
Behaviour in the United Kingdom. 1920-1938, Vol L Cambridge University
Press, Cambridge.
Strand, I. E. and N. E. Bockstael. 1985. "The Effect of Common Sources of
Regression Error on Benefit Estimation." Working paper, University of
Maryland.
Sutherland, R.J. 1881. "A Regional Approach to Estimating Recreation Benefits
of Improved Water Quality. " Journal of Environmental Economics and
Management 9:229-47.
Thayer, Mark. 1981. "Contingent Valuation Techniques for Assessing
Environmental Impacts: Further Evidence." Journal of Environmental
Economic* and Management 8:27-44.
Theil, Henry. 1976. Theory and Measurement of Consumer Demand. Vols 1 and
2. Amsterdam.
Tobin, James. 1985. "Estimation of Relationships for Limited Dependent
Variables." Econometrics 26:24-36.
118
-------
u. S. EPA. 1983. Water Quality Standards Handbook. office ofWater
Regulation, and Standards. Washington, DC.
U. S. Department of Commerce. 1984. "Marine Recreational Fishery Statistics
Survey, Atlantic and Gulf Coasts, 1979 (Revised) - 1980, Current Fishery
Statistics 8322." Washington, DC.
u, S. Department of the Interior. "1980 National Survey of Fishing, Hunting,
and Wildlife-Associated Recreation." U. S. Fish and Wildlife Service.
Washington, DC. Undated.
Vaug ban, W. J. and C. S. Russell. 1982. "Valuing a Fishing Day: An
Application of a Systematic Varying Parameter Model." Land Economics
58:450-62.
Vaughan, W. J. and C. S. Russell. 1982. Fresh Water Recretional Fishing; The
National Benefits of Water Pollution Control. Washington D. C.:RFF, Inc.
Vaughan, W. J., C. S. Russell, L. P. Gianessi and L. A. Nielsen. "Measuring and
Predicting Water Quality in Recreation Related Terms." J." Env. Mornt.
15:363-380.
Violette, Daniel. 188s. "A Model to Estimate the Economic Impacts on
Recreational Fishing in the Adirondacks from Current Levels of
Acidification." Paper presented at the AERE Workshop on Recreation
Demand Modeling, Boulder, CO.
Wales, Terence J. 1971. "A Generalised Linear Expenditure Model of the
Demand forNon-Durable Goods in Canada." Canadian Journal of Eonomics
4:471-84.
Wales, Terence and A. D. Woodland. 1883. "Estimation of Consumer Demand
Systems with Binding Non-Negativity Constraints." Journal of
Econometrics 21:263-85.
Walsh, Richard G. Undated. '*An Extension of the Contingent Valuation
Approach to Estimate the Effect oflnstream Flow on Recreational
Demand." Unpublished.
Walsh, R G., D. A. Greenley, R A. Young, J. RMcKean, anc&. A. Pinto. 1978.
"Option Values and Recreational Benefits of Improved Water Quality: A
Case Study of the South Platte River Basin, Colorado." Rep.
EPA-600/5-78-100. U. S. Environmental Protection Agency, Research
Triangle Park, N. C.
Walsh, R G., L. D. Sanders, and J. B. Loomis. 1985. Wild and Scenic River
Economics: Recreation Use and Preservation Values. American Wilderness
Alliance, Englewood, Colorado.
119
-------
Wegge, Thomas, W. Michael Hanemann and Ivar Strand. 1985. "An Economic
Assessment of Marine Recreational Fishing in Southern California."
Report subritbed to National Marine Fisheries Service Southwest Regional
Office.
Wennergren, E. B. and D. B. Nielson. 1970. ' "Probability Estimates of
Recreation Demands." Journal of Leisure Research 2:112-22.
Williams, J. B., T. P. Smith, M. J. Speir, and S. Early. 1983. "1980 Maryland
Saltwater Sport Fishing Survey. " Maryland Department of Natural
Resources Tidewater Administration Report No. TA-CRD-83-1. Annapolis,
MD.
Willig, R.D. 1976. "Consumer's Surplus without Apology." American Economic
Review 66:589-97.
Willig, R. D. 1978. "Increment Consumer's Surplus and Hedonic Price
Adjustments." Journal o/Economic Theory 17:227-253.
Wilman, Elizabeth A. 1884. External Costa of Coastal Beach Pollution.
Resources for the Future, Washington, DC.
Wilman, Elizabeth. 1985. "Valuing Quality Changes in Recreation Resources.™
Paper presented at the AERE Workshop on Recreation Demand Modeling,
Boulder, CO.
Zellner, A . and Soo-Bin Park. 1879. "Minimum Expected Lose (MELO)
Estimators for Functions of Parameters and Structural Coeffiecients of
Econometric Models." Joimial of the American Statistical Association.
74:365:185-193.
120
-------
Appendix A
The Random Digit Dialing Survey Telephone Procedures
For the Random Digit Dialing Survey, three instruments (copies included in
the Appendix) were developed: a two-page screening form, an 1 1-page,
28-item questionnaire, and a Record of Calls sheet.
1. Screening Form
The screening form was intended to determine the eligibility of the
location served by a randomly generated telephone number. The number was
printed on a label affixed to the top of the screening form. Pay phones and
phones used only by businesses were not considered eligible, since people
answering ouch telephones would be eligible at their residence phone.. In
addition, if the household served by the phone waa not located within the
counties/cities making up the selected SMSA's then that phone (residence) was
not eligible. Once an eligible phone (residence) waa identified, a member of
the household who waa 18 years of age or older waa required. If all
residents were under 18, the screening waa completed with •member of the
household who waa 14 or older.
The screening form waa composed of five sections: an Identification
sectdoB consisting of an area code, telephone number, and five-digit case
identification number, all printed on the aforementioned label; »brief
introduction to be read by the interviewer which explained the study; a
screening section which waa used to eliminate pay phones, businesses without
living accommodations, and residences not located in certain specific SMSA's;. a
screening status section to record the screening eligibility of the location; and
a questionnaire status section to record whether or not a questionnaire wafl
completed with an rfigiHlft person.
2. Questionnaire
The Random Digit Dialing Survey Questionnaire was intended to determine
the following:
Uae or intent to use the Chesapeake Bay for recreation during 1984;
Reasons for nonuse;
Activities that the respondent (and his/her family) participated in
while visiting beaches;
Reasons the respondent or other members of his/her family do not go
in the water during visits to the western shore beaches;
Changes in swimming participation in the Chesapeake brought about
by change, in the water quality;
The respondent's perception of the water quality in the Chesapeake;
The value respondents place on the Bay and how they visualize that
improvement should be made and financed.
121
-------
AS in the User Intercept Survey, a series of demographic questions
which will enable analysts to establish profiles of beach users and nonuser
were included in the questionnaire.
The Random Digit Dialing Survey Questionnaire waa also divided into
sections. The first, Identification, had space provided for recording the case
ID number from the screening forms the telephone interviewer's initials, and
the date the interview was completed. The second section, as in the User
Intercept Survey, was a lengthier introduction to be read by the telephone
interviewer, which went into greater detail regarding the purpose of the
survey and contained statement informing the respondent ofthe voluntary
nature of his/her participation in the study and assurances of the
confidentiality of the data collected. The third section, Recreational Uae of
the Chesapeake Bay, sought specific responses which would: (1) enable
analysts to determine if and how the beaches were used and (2) what the
overall perception of the water quality waa. Thie waa followed by a fourth
and final section consisting of some 18 demographic questions.
Data Collection Methods
Two field interviewers were trained in Baltimore for the data collection of
the Ueer Intercept Survey on May 25, 1884. A Field Interviewer Manual was
developed (which is available upon request) and included quemtion-by-question
specification, probing techniques, confidentiality procedures, refusal
conversion strategies, and other measure*necessary to assure the collection
of standardized, quality data during the course of the field survey process.
Alao covered in the manual were: background information, assignment
information, sampling procedures and administrative procedures.
The final day of field work on the ueer survey was August 16, 1984. The
confirmation portion of the user survey was completed on September 1. The
following represents the response rates forthe field work:
Table A.I
Response Rates for Beach User Survey
Sample
Individual
Successfully
screened
Eligible
Individuals
Eligible Individuals
Interviewed
468
100%)
463
(98.79%)
414
100%)
408
(88.55%)
122
-------
Of the 468 individuals screened, 60 were not administered questionnaires for
the fclawing reasons:
Ineligible because of residence 39
Refused screening 7
Language barrier-screening 6
Other 2
Refused questionnaire 3
Language -barrier - questionnaire 3
Regarding the confirmation portion of the ueer survey, 340 of the people
interviewed gave telephone numbers or come other piece of information
through which contact could be made to conduct a confirmation/intention
interview. Approximately 240 (71 percent) of these individuals were
successfully contacted during the time period allowed.
Training of telephone interviewers for the Random Digit Dialing Survey
started on July 23. A total of 11 telephone interviewer, were hired with three
of these spending the majority of their time making confirmation/intention
calls to participants in the User Intercept Survey.
As in the User Intercept Surveyt each interviewer received a copy of a
Telephone Interviewer Manual specifically developed for this phase of the
project, as well as copies of the Random Digit Dialing instruments. The
Telephone Interviewer Manual (available upon request) included question-by-
question specification, probing techniques, confidentiality porcedures, refusal
conversion • trategiea, and other measures necessary to assure the collection
of standardized, quality data during the course of the telephone •nvey
process.
Approximately 192 telephone interview, were completed with western shore
beach users. The remainder consisted of approximately 804 nonusers and 48
intended users. The following two tables represent questionnaire completions
per strata and final totals for screening and questionnaire status codes.
123
-------
Table A.2
Questionnaire Completions Per Strata
Stratum
Number
1
2
3
4
5
6
Totals
Cases
Avail .
1,230
1,100
408
1,014
820
1.560
6,132
Cases
Assigned
1,060
1,000
408
1,014
820
1.560
5,962
Quest.
Complete
155
225
70
96
111
293
1,010
Quest.
Partial
10
7
0
4
6
7
34
Total
Quest.
165
232
70
100
111
300
1,044
Quest .
Needed
138
220
" 77
112
158
295
1,000
Diff.
+ 27
+ 12
- 7
- 12
+ 19
+ 5
-44
Table A.3
Final Telephone Result Totals for Screening and Questionnaire Status Codes
Screening Status Cod»»
Eligible Identified/Screener Completed 1,108
Nota Working Telephone Number 2,866
13
843
897
11
10
203
Pay Telephone
Business Telephone
No Answer After Repeated Calls
Telephone Not Located in Bait./Wash. SMSA
No Eligible Respondent Available After Repeated Calls . . .
Refused to Answer Screening Questions
Language Barrier 0
Other
Questionnaire Status Codes
3
Questionnaire Completed 1,010
Questionnaire Partially Completed 34
Language Barrier . ...' 0
Questionnaire Refused 63
Other 1
124
-------
Appendix B
Telephone Survey instrument
125
-------
CHESAPEAKE BAT BEACH USE SURVEY
TELEPHONE SCREENING FORM
A. IDENTIFICATION
TsuPllGNs NUMBER ( ) .= - CASE ID NUMBER
B. INTRODUCTION
I an calling fro* the Research Triangle Institute near Raleigh, North Carol in* to
conduct • telephone survey • bout the Chesapeake Soy for the University of Maryland. To
find out if I'vt contacted the proper type of place, I need to ask • few sieple questions.
First,
c. SCREENING
1. Is chi* telephone niaabcr (READ THE 7 DIGIT NUMBER ABOVE) in • rea cede
(READ 3 DIGIT AREA CODE ABOVE)?
Me JT1 (RANG UP AND DIAL NUMBER AGAIN) (j;
TOS m
2. It this • pay phone?
*« m (2i
Yes |1"| (CO TO *)
3. • . Is tblt telephone located in • private residence or a business?
Private residence JTJ (GO TO 4a) (3)
Business iTl
b. Aro thereby living aeconodatians* t this place of business?
No Qj (GO TO 6) (4)
Yes GO
6. Do the people living there US. this phone for their calls?
No CD (00 TO 6) (s)
Tes Qj
d. Whoa can I apeak to one of the people who lives • t this business
location and uses this telephone for personal calls?
ANSWER :
(CALL BACK IF NECESSARY TO COMPLETE SCREENING)
4.e. Are you • neaber of the household serviced by this telephone?
TesQFI1 (m TO So)
b. Unen will* •caber of the household bea> liable to talk tone?
ANSHP;
__ _________ _
(CALL BACK IT MECISSAKI TO COMPLETE SCREENING)
S. a. Is this residence located in Maryland, Virginia, Vathin|ton, DC
or son* other place?
Maryland ITI (GO TO b) <6>
Virginia ITi (GO TO b)
Washington, DC ITI (GO TO 7)
Sosw other place PH (CO TO 6)
b. In whet county is this residence located?
ANSWER : _ (RECORD ANSWER AND CODS BELOW)
MARYLAND l?>
(Anne Aruadel, Baltinore including city, Carroll,
Charles, Harford, Howard, Nontgonery,
Prince Georges ................................... J3 (GOTO 7)
VIRGINIA
(Arlington, Fairfax. Loudon, • od Prince William.
Also include Alexandria city, Fairfax city and __
Falli Church city .................................. ITI (CO TO 7)
-------
6. I'm sorry but we are not interested in talking to people (at Pay telephones/at
business telephones/in th« • rea where YOU live). Think You for • svermg my
question* • nd could I hove your OOOO in case my supervisor want* to check my
work?
NAME (COMPLETE SCREENING STATUS IN PART D ONLY)
7. • This telephone is located • tsresidence in • • rea where we* re interested
in talking to people . Are you under 18 years of • ge or over 18 years of
• ge?
under 18 JT] (GO T0 b)
over 18 FT1 (GO TO c)
b. Whoa will I be • bLo to talk to SOMOQC over 18 years of • ge who lives in
this household? ANSUER:
(CALLBACK IF NECESSARY TO VERIFY ANSWERS AND COMPLETE QUESTIONNAIRE. IF
ALL RESIDENTS ARE UNDER 18 PROCEED WITH ANT RESIDENT 14 OR OLDER.)
c. How swny telephones with different nuabers, not • xtensions, service this
household? , , | (8-9)
d. Hay I have your oaaM in case sqr supervisor wants to check eiy work?
NAME _____ (PROCEED TO QUESTIONNAIRE BUT RETURN
TO AND CODE SCREENING STATUS AND QUESTIONNAIRE STATUS AFTER QUESTIONNAIRE
ADMINISTRATION.)
D. y<^fprfi»c STATUS COPES
Eligible Identified/Screener Completed ........................ l"oT!
Not a Working Telephone Nuaber ................................ ("oTi
Pay Telephone ................................................. I 03 i
felines* Telephone ............................................ CoTI (20-11.
No Answer After Repeated Cello ................................ foTl
Telephone Not Located In l*lt. /Wash. S. M.S. A .................. FSTi
lo IligibU toistadiat Anlliblt Adit Itptattd CtUi ......... llTi
Refused To Answer Sereenint Questions , ........................ 1 97 i
Ut|U|i brtiit .............................................. I II i
Other [[[ m I
E. QUESTIONNAIRE STATUS CODES
Questionnaire Coexisted (No Mil followup) .................... L2U
Questionnaire Completed (Mail followup) ....................... I °2 '
Questionnaire Partially Completed ............................. l"03j
-^— ^ (2 w» " ~
Lan|uage Sorrier .............................................. I 04 i
Questionnaire Refused ......................................... I *1 i
Other .................................................. - ...... I 061
COLS. U'73 'blank
-------
CHESAPEAKE BAY TELEPHONE SURVEY QUESTIONNAIRE
Conducted by Research Triangle Institute
for the University of Maryland
A. IDENTIFICATION
CASE ID
Interviewer Initials
Date Interview Completed I I I | |
Month Date
B. INTRODUCTION
As I said Earlier, researchers «t the University of Maryland «re current-
ly studying citizens' use of the Chesapeake Bay. I will • sk you some
questions regarding your recreational use of the Chesapeake Bay, parti-
cularly tt the beaches. I flso have tofsk some questions which will
enable the researchers to fstablish profiles of typical users fnd non"
users of the Bay. There is no direct benefit from taking part in this
study tnd you have the right to refuse to taswer fay or til of the ques~
tions or discontinue your participation at any time. The information
that you provide will be combiaed with that provided by other people who
participate in the survey to •ssure complete confidentiality• nd your
same will not be released or revealed to 0nyone other than authorized
project staff. The results of this survey may be helpful in Effectively
Allocating mosey to cleaning up the Bay.
-------
RECREATIONAL USE OF THE CHESAPEAKE BAY
1. Have you orfnymembers of your household used the Chesapeake Bay
for recreation in 1984?
No FT! (GO TO Q. 2.)
Yes ..... JT| (GO TO Q. 4.)
2. Do you or any members of your household intend to use the Chesapeake
Bay for recreation during the rest of 1984?
No Qj (GO TO Q.3. )
Yes jTj (GO TO Q.4)
3. What tre the reasons you tnd members of your household have not used
and do not intend to use the Chesapeake Bay for recreation during
1984? (CODE ALL RESPONSES GIVEN INTO-THE CATEGORIES BELOW)
CODE IF GIVEN
•\ Not interested in water related recreation I 2 I
b. Unable for health reasons I 2 I
c. Coiti too auch I 2 I
d. Takes too much time to get there (too far to
travel) I 2 I
•\ Unacceptable water quality I 2 |
f. Too atny jellyfiih fTj
g. Too crowded ( 2 I
h. Have not had • chance (too busy) I 2 I
i. Other l~2~l
(GO TO Q. 8.)
-------
What* ctivities did or will you (and/or members of your household)
participate in while using the Chesaoeake? (READ EACH OF THE FOLLOW-
ING AND INDICATE PARTICIPATION FOR EACH ACTIVITY.)
NO YES
a. Fishing 0 . • • TJ] |T"i
b. Swinaning . . . TT] ITl
c. Boating Qj [Tj
d. Hunting • • «.LLJ IT"!
». Beach Activities •...-...• LJj I 2 I
f. Sightseeing o= ...CD QQ
g. Other CD
5. This next question 01SO pertains to YOU and •«•!»*" Of
household. During 1984 did 0ny of you or will 0ny of you... (READ
THE FOLLOWING. )
NO YES
» Visit beaches, on the Eastern Shore of the
Chesapeake, for^xample shores close to
Cambridge, Salisbury or Chestertown? I 1 I I 2 j
b. Visit beaches on the ocean, such »s
Ocean City? « •.. « | 1 I I 2 \
c. Go iwianing from •boat in the Chesapeake? .... I 1 I I 2. I
d. Go swiming in public or private stunning pools? . I 1 I I 2 I
e. Visit beaches on the Western Shore of the
Chesapeake, for •xample beaches near Baltimore,
Annapolis, Prince Frederick or Lexington Park? . . . (JJ I 2 1
(IF YES TO PART e., ASK f. IF NO TO PART t, GO TO QUESTION 8.)
f. During visits to Western Shore beaches did
nr will anyone tttend but not go , . ,
in the water fflt reMOn? m Qj
for any reason?
(IF YES TO PART f, GOTO Q. 6. IF NO TO PART f, GO TO Q. 7.
-------
What* re the reasons you or others do not go in the water during
visits to the Western Shore beaches?
Do (READ EACH PART AND CODE NO OR YES.)
NO YES
a. You or they believe the water is
b.
c.
Can
You or they believe there t-re too many
jellyfish ,
You or they have some other reason
you tell me which Western Shore beaches you (and
have visited in 1984 or plan to visit during the rest
. . i_u 23
. . Cu Qu
your family)
of this year?
(CHECK A/0 OR YES FOR EACH BEACH LISTED. )
a.
b.
c.
d.
e.
f.
8"
h.
i.
j.
k.
1.
m.
n.
0.
P-
q-
r.
BEACH
Sandy Point St. Park
Fort Siallvood
Bay Bridge Beach
Herrington Harbor *
Kurtz Pleasure Beach
Caop Merrick
Breezy Point Beach
Chesapeake Beach
North Beach
Rod and Reel Dock
Point lookout St. Park
Elm'f Beach
Horgantown Beach
Miaii Beach (Baltimore)
RockyPointPark
Conrad' sRuthVilla
PorterNewPark
Other (SPECIFY)
VISITED
NO YEs
..mm
..mm
. .QJ m
. . LU UJ
..mm
. -Lll LLJ
. . m LLJ
. -LU LLJ
. tix: m
• LU LLJ
. CD m
. . m m
. m m
..mm
..mm
..mm
..mm
QJ CD
-------
Have you (or members of your family who live with your) »ver changed
your swimming participation in the Chesapeake because of changes in
the Bay's water quality?
No |T"l (GO TO Q. 9.) . (4f
Yes (stopped) . . f2~|
Yes (started). . FT]
a. In what year did you (or members of your family) last change
your swimming habits in the Chesapeake because of changes in
the Bay's water quality?
Year.
(4
We would like to find out how people currently perceive the water
quality in the Chesapeake Bay.
• \ Lto you consider the water quality in the Chesapeake to be
w cceptable or unacceptable for swioBini • rid/or other water
tctivities.
Acceptable ....... jTj (4
Unacceptable ..... ^ m
b. Do you believe the water quality varies ft different beaches
• ions the Western Shore of the Chesapeake?
No ........... [Tj (GO TO CHECKPOINT I.) (5
Yes . . . '. ...... JTl
c. (IF YES, SAY:) In general, which statement best describes your
beliefs?
The water quality is better North of Annapolis . . !JJ , c
< «r
The water quality is better South of Annapolis . . I 2 |
INTERVIEWER CHECKPOINT I
REFER TO QUESTION 9.A. .
WAS THE WATER QUALITY IN THE CHESAPEAKE RATED AS UNACCEPTABLE?
No Qj (GO TO Q. 11. )
Yes [Tj (GO TO CHECKPOINT TABLE.)
-------
CHECKPOINT TABLE
ENTER THE LAST DIGIT OF THE CASE ID NUMBER HERE.
CIRCLE AND
IF THE LAST DIGIT IN USE THIS AMOUNT IN
THE CASE ID NUMBER IS QUESTION 10
1 $5.00
2 $10.00
3 $15.00
4 $20.00
5 $25.00
6 $30.00
7 .... . . . . . $35.00 "
8 ....:::::..:::::::..::::. $40.00
9 ' $45.00
0 0 ., . . ..$50.00
10. You indicated that in your opinion the water quality in the Chesa-
peake is unacceptable for (visaing. Would you be willing to pay
(AMOUNT FROM CP TABLE) in txtra state or federal taxes per year, if
the water quality were improved so that you found it • cceptable to
•via in the Chesapeake?
No. . . .".
Kes . . .
Don't know
D. BACKGROUND INFORMATION
11. The next few questions tre tbout you tnd your household. How many
of tach of the following types of people live in your household?
(READ EACH OF THE FOLLOWING AND ENTER THE NUMBER OF EACH TYPE. )
*. Adults (age 18 tnd older)
b. Children between the tjges ofl4*)nd 18
c. Children under age of 14
12. What best describes your status in the household?
a. Grandparent I 1 j
b. Pirent jTj
c. Child. . . . , fTi
d. Other relative GQ
e. I live flone or with unrelated individuals . . .15!
-------
13. How many years have you (and your family) lived in either Maryland,
Virginia, or Washington, DC?
Number of years . . I I |
14. Do you or any other members of your household own (READ THE FOLLOW-
ING) ... '
• \
b.
c.
d.
'e.
f.
Are
• ar
stuc
»
b.
c.
d.
»
f.
NO
a boat? « i Qj "|
a boat trailer? 01 1
fishing tackle (rod, reel,
tackle box, etc.)? i 01 i
a recreational vehicle (RV)?. . . ; 01 j
• •wining pool? | 01 t
Other recreational items (SPECIFY)
! Q.
ron
YES
Q)2J
DDL
Pol"
COLS.
ron
ron
PbT",
you one of the principal waie tamers in your
ner but not the principal wage earner, or «re yo
lent or retired?
One of the principal wagetarners in the family
A wage earner but not the principal wage earner
Homemaker . . .
Retired
Student
Other (SPECIFY)
DK RF
|~98^ |TH
- rw i~9Ti
rgTi r?7j
ilFi p97~i
73-80 - » CASE Oil
OD OD
3ES QD
[Hj LIE
household, «wage
u •homemaker, •
CTj
ijj
(Tj
. fTi
* •* '
IT I
(55-
(67-
(65-
(71-
(1-2
(3-4
(5-6
15.
INTERVIEWER NOTE: ASK QUESTIONS 16 THROUGH 19 IF CODE 1 OR 2 IS MARKED IN
QUESTION 15 OTHERWISE GO TO QUESTION 20.
16. How many hours do you usually work per week?
ff-.
Hours
17. How many paid vacation days wi 11 you have • Itogether in 1984, in-
cluding those you've Already taken?
Vacation Days . . ]_
-------
18. For •typical recreational outing, if you did not go, could you work
• tsome paying job instead?
No fTi
Yes |"Tj
19. If you could have worked, what hourly wage might you have been paid
specifically for the hours you worked?
a. $3.35/hour "oTi
b. $3.36 - $5.00/hour foTi
c. $5.00 - $7.50 r5T>
d. $7.50 - $10.00/hour . . . • m
». $10.00- $15.00/hour .... foTj
f. $15.00 • $20.00/hour .... ("581
g. $20.00 - $25.00/hour .... Qoj
h. Over $25.00/hour I 11 i
i. Don't knew i 98 !
j. Refuse i 97 i
20. Are there *ny (other) major wage tamers in your family?
NO Q"^0010 Q-25. )
Yes jTj
The next few questions tre tbout the other major wage tamer.
21. How many hours does he/she usually work per week?
Hours
22. How Mny paid vacation days will he/she have Itogether in 1984?
Vacation Days . . I 1
23. For the typical recreational outing, if he/she did not go could
he/she work-at some paying job initead?
No T
Ye» m
-------
24. If he/she could have worked, what hourly rate would he/she have been
paid specifically for the hours worked?
• \ $3.35/hour j 01 .
b. $3.36 - $5.00/hour roTj
c. $5.00 - $7.50 f03"l
d. $7.50 - $10.00/hour .... \W\
e. $10.00 - $15.00/hour ....
f. $15.00 - $20.00/hour ....
g. $20.00 - $25.00/hour .... \Wi
h. Over $25.00/hour 0m
i. Don't know |~98l
j. Refuse I 97 I
25. We need «n •stimate of your household's incoae for •!! of. 1984. I
will read • series of income categories. Please stop «e when I read
the category which best describes the total aaount of income •!!
members of your household will receive during 1984.
INTERVIEWER CHECKPOINT II
ENTER THE LAST DIGIT OF TIE CASE ID NUMBER HERE.
IF THE LAST DIGIT START READING THE
IS ANSUER CHOICES AT
l,3,s,7,9 a. less than $5,000
AMD ASCEND
2,4, 6,8,0 1. over $100,000
AND DESCEND
• \ less then $5,000 . . . I 01 . |.
b. $5,000 to $10,000 . . I 02 I h.
c. $10,000 to $20,000 . . rbTl i.
d. $20,000 to $30,000 . . \W\\.
». $30,000 to $40,000 . . PoTI k. '
f. $40,000 to $50,000 . . foe"! 1 .
Don't know
Refused ....
$50,000 to $60,000 . . 101
$60,000 to $70,000 . . flTl
$70,000 to$80,000 , • I 12 !
$80,000 to $90,000 . . ! 13 I
$90,000 to $100,000. . :"TTj
Over $100,000 .... ~
. . 3E
. • iTTi
-------
26. CODE SEX BASED ON NAME, PREVIOUS ANSWERS/REMARKS OR >SKi > - you
female or male?
Female ..... fTT
,
Hale I 2 ;
27. Which racial group do you identify with?
White • Olj
Black '_02 .»
Oriental • °3J
Other (SPECIFY) I"0*1
Refused " I 97 i
Don't know 1 98J
COLS. 28-73 * blank •
COLS. 74-80 = CASE02i
-------
28. •. This is the last question. We would like to send short ques-
tionnaires about the Chesapeake Bay to people through the mail.
We would include •postagepaid 0 nvelope to return the completed
questionnaire, so it would not cost anything to mail it back to
us.
Would you be willing to receive and complete such a
questionnaire?
No JTI (GO TO C.)
Yes (Tj (GO TO b.)
b. What is your mailing tddress?
(VERIFY NAME)
(P.O. Box/Street number • nd name)
CityStateZip"
ENTER CASE ID NUMBER
Thank you for taking time to • nswer our questions. Your respon-
ses will be very helpful in determining the status of twinning
• nd other •ctivities on the Chesapeake Bay.
IF YES TO 28*. ALSO SAY: When the questionnaire comes through
the mail, please complete •nd return it «s quickly »s possible.
-------
Appendix C
The User Survey and Sampling Procedures1
This section is devoted to a description of the samplingprocedures used
in a survey of Chesapeake Beach Use conducted in 1984. Unlike the data
used in analyses of boating and fishing in Chapters 5 and 6, the data used in
this chapter were collected during an earlier budget period of this cooperative
agreement. Great care was taken with the sampling frame to improve
confidence in the results. Because the survey itself dsimportant to the
project, we describe the content and procedures extensively. Copies of the
survey instruments can be found in Appendix C.
From May 26, 1984 to August 19, 1984, Research Triangle Institute (RTI)
interviewed individual on the western shore beaches in Maryland. The study
population consisted of all residents of the Baltimore and Washington, D. C.,
SMSA's, age 14 or older, that used these beaches for recreation in 1984. More
specifically, the population waa limited to recreational users of the following 12
beaches:
Strata
Geographic size
10 sandy Point north. large
2. Point Lookout south large
3. Fort Smallwood north small
4. Miami north small
5. Rocky Point north small
6. Elm's Beach south small
7. Bay Ridge south large
8. Kurtz north small
9. Breezy Point south small
10. Rod & Reel south small
11. Morgantown south ml 1
12. North Beech south Oman
Four hundred and eight individuals were interviewed at the beach to
learn of their recreational patterns and perception of water quality at these
beaches. These individual were randomly selected from sample beaches and
days. The sampling design can be described as a two-stage stratified sample
in which a probability sample ofbeaches and days was selected, and a random
systematic sample of persons was interviewed at each sample site (day-beach
combination) .
1 The discussion of the sample is composed ofselected excerpts from Devore,
McDonald, Myers and Williams, Chesapeake Bay Beach Use Survey. Research
Triangle Institute, 1984.
139
-------
A sample of at least 300 completed interviews was intended so that relative
sampling errors would be approximately 11 percent or less for estimating
proportions of .30 or larger. This assumes that a design effect of
approximately 1.5 might result because of clustering (a beach-day unit
constitutes a cluster of individuals) and unequal, weighting. A self -weighting
sample was sought but difficult to obtain because of errors in size measures
(projected number ofusers for a beach-day unit) .
The number of sample primary sampling units (PSU) waa baaed on the
requirement to obtain at least 300 completed interview, such that each PS U
would involve approximately one-half day ofinterviewing for a two-person
team. To match this design, each PSU was randomly designated a. a morning
or afternoon interview period, beginning at 1000 and 1300 hours, respectively.
Hence, an average of eight or nine completed interview, waa expected for each
of the 36 PSU's (see Table 4,1 for sample allocation).
Stratification and Sample Selection
Primary sampling unit. (PSU's) consisted of beach-day units (1,204 PSU's
= 66 days x 14 locations). The 14 locations consisted of 11 beaches plus three
beach locations at Sandy Point - partitioned into three segments because of
its relatively large size and usage. Beaches were stratified into north (Sand y
Point and beaches north) and couth (the prior listing of beaches indicates the
stratification). The beaches wore further stratified by size (expected weekly
usage). Large is defined as greater than 7,000 visitors per week in north
beaches and greater than 3,000 in the south beaches. Additionally, days were
stratified into weekdays, Saturdays, and Sundays (and holidays) . The strata,
their population and sample PSU counts are indicated in Table Cl.
Table C. l
Population Counts and PSU Allocation by Stratum8
Anticipated Number of Beach
Visitors May 26 to
August 19, 1964 (thousands)
Sat Sun Weekday Total
Number of PSU's (beach-day
units) Allocated to Bach
Stratus
Sat Sun Weekday Total
North/large
Nbrth/srBll
south/large
South/ small
Totals
77
6 7
37
20
201
110
6 4
43
24
261
106
94
44
17
259
293
243
124
61
721
3
3a
?2
22
10
5,
42
2
2
13
5
4
22
2
13
13
13
6
6
36
1 Population counts are baaed on cite interviews prior to the survey, and PSU
allocation is proportional to number of visitors with the constraint that
greater or equal to two per cell are selected.
20ne fewer resulted in these cells because of random subsampling needed in
the latter part of the survey period. (A slightly larger number of
interviews than expected were obtained early in the survey.)
140
-------
The selection of PSU's within strata involved equal probability y for days
and probability proportional to size (expected number ofvisitors) for beaches.
Within selected PSU's, individual were selected on site with equal
probabilities. Approximately eight or nine interviews were needed from each
PSU. The procedure used to satisfy this objective waa for the interviewers to
estimate the number of beach users just prior to starting time for
interviewing. By comparing this estimated count to a table prepared
specifically for this survey, the interviewers obtained a sampling interval and
a random starting point. By using the interval and starting point in a
pro-designated pattern for the entire site, a valid systematic sample of users
was obtained.
Sampling Weights
Whenever observation units (in this caae the individual users) enter the
sample with different probabilities, weights must be used with the
observation to obtain unbiased estimates for the study population. Because
of the complexities in an intercept-type survey, selection probabilities are
often not known. In the present survey, however, PSU selection probabilities
are known, and final-stage selection ofindividual can be reasonably estimated
(even though a systematic sample of visitors at the site is taken), their
chance of selection will vary with the amount of time spent at the beach
during the sample day as well as number of visits to this and other stud y
beaches during the survey period. A sampling weight for the jth individual
is calculated aa follows:
where
p = * +*j « the "selection probability" (expected number of hits)
1 >+> for PSU (i) ;
d = the number of daya during the survey period for the particular
type of day being sarpled (d = 13 for Saturdays, for example);
n = number of PSU's being selected for the stratum (12 strata);
s, = size measure for PSU (i) = the expected number of visitors for
that type of day (S+ = tSt);
p, j * sampling rate of users within sample PSU (i); and
and = factors to adjust for number of houra user spent at the beach
, .
'
on the sarple day and number of trips to the study beaches
during the sumar, respectively.
141
-------
Note that fjj and fjj are baaed in part on intentions. In fact, f» j is
particularly uncertain if the interview is being conducted early in the season.
To verify the accuracy of these intentions both to obtain selection multiplicity
and to estimate related statistics, intention-based questions were verified b y
telephone at the end of the season for interviews taken early in the survey.
Screening and Confirmation
A screening form waa designed to identify as eligible respondents only
those people living in certain counties of Virginia or Maryland, or Washington,
Do C. In addition, the screening process ruled out as ineligible any person
who was under 14 years of age; however, if the selected sample individual was
over 14 but less than 18 years of ageand in the company of someone he/she
lived with who was over 18 years of age, the interviewer deferred to the
older individual.
The User Intercept Survey Questionnaire was designed to record and
collect the following:
Frequency of visits made to beaches on the western shore of the
Chesapeake
Activities that the respondent (and his/her family) participated in
when visiting beaches
Activities not participated in and the reason why they were not
Coat related to a typical trip to each beach that the respondent had
visited since January 1, 1984
The respondent's perception of the quality of the beach and the
beach facilities 11 each beach wtth which he/she was familiar
Factors that influence a respondent's decision to visit or not visit a
beach
The respondent's willingness to continue to viait the sample site if
coots related to the use of the beach were to rise.
In addition, a series of demographic questions to enable analysts to
establish profiles of beach users were also included in the questionnaire.
A Confirmation/Intention Contact sheet (used in conjunction wtth a Record
of Calls sheet) was developed not only to confirm the number of visits to
sample beaches reported in the Intercept Survey Questionnaire, but also to
ascertain the number of visits the respondents intended to make at any
sample beach for the rest of 1984. The confirmation/intent ion contacts were
only made with those respondents who had provided adequate information to
contact them by telephone.
142
-------
Data Collection Procedures
The data collection schedule originally called for a total of 36 visits to the
12 sample beaches, with an expected yield of 300 or more completed interviews.
This schedule was later revised and the subsequent number of total visits wae
reduced. Table C.2 shows the original data collection schedule. Those
sampling days which were eliminated have been denoted by an asterisk. The
resulting number of visits are summarized as follows (AM indicates
interviewing beginning in the morning at 10 o'clock, and PM indicates
interviewing beginning at 1 o'clock in the afternoon):
Beach
1
2
3
4
6
7
8
9
10
11
12
Beech
Name
Sandy Point
Point Lookout
Fort Smallwood
Miami
Elms
Bay Ridge
Kurtz Pleasure Beach
Breezy Point
Rod and Reel Dock
Morgantown
North Beach
TOTALS
AM
Visi ts
5
1
1
0
0
0
0
0
1
1
1
12
PM
Vimits
6
1
0
3
0
3
1
1
0
1
0
18
Total
Visits
11
2
1
3
0
3
1
1
1
2
1
30
For the User Intercept Survey, each field interviewer waa given a
of the names, addresses, and ID numbers for each PSU (beach) in the
assignment, along with area maps with the beaches marked. In addition, each
interviewer wae given a sketch of each beach. Other materials included were
the table, to determine sampling rates and listing forma forcounting and
selecting sample individuals. The interviewers always worked as a team,
splitting up only to interview eligible respondents.
The field interviewers were asked to review their materials and
determine the moat efficient route oftravel to reach each beach. Upon arrival
at a beach, they had first to check that they had correctly identified and
located the precise boundaries of the area. Once this waa verified, the
interviewers estimated the number of sample individual on the beach ,
spending no more than 30 minutes in doing ao. When the estimate was
determined, they looked up that number in the table of sampling intervals.
Marking the estimate and sampling interval at the top of the listing form, they
next consulted their list of random numbers to determine the number of the
first person to be interviewed and marked that in the space provided on the
listing form. They circled the number of the randomly selected start person,
counted the proper interval, and circled the last interval number as the next
selected person. This activity continued until they had gone through the
entire list interviewing the selected individuals.
143
-------
Table C.2
Chesapeake Bay Beach Use Study Data Collection Schedule
Date
May 26
27
28
29
30
31
June 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
July 1
2
3
4
5
6
7
Beach Beach
Name NLnber
Pt . Lookout
Morgantown
Miami
Miami
Sandy Pt. (NofS)
Sandy Pt. (NofS)
Morgantowm
Sandy Pt . (NofS)
Bay Ridge
Sandy Pt. (NofS)
Rocky Pt.
North Beach
Sandy Pt. (E)
Rocky Pt.
Breezy Pt.
sandy Pt. (SofS)
Sandy Pt. (E)
2
11
4
4
1
1
11
1
7
1
5
12
I
5
9
1
1
Sample
Time
1300
1000
1300
1300
1000
1000
1300
1300
1300
1300
1000
1000
1000
1300
1300
1000
1300
Date
July 8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Aug 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Beach Beach Sample
Name Number Time
Kurtz
Bay Ridge
Rocky Pt.
Miami
sandy Pt. (SofS)
sandy Pt. (SofS)
Rod & Reel
North Beach
Sandy Pt. (SofS)
Pt. Lookout
Bay Ridge
Ft . Smallwood
Rocky Pt.
Kurtz
Pt . Lookout
Sandy Pt. (E)
sandy Pt. (SofS)
Sandy Pt. (NofS)
8
7
5
4
1
1
10
12
1
2
7
3
5
8
2
1
1
1
1300
1300
1300
1300
1300
1300
1000
1000*
1300
1000*
1300
1000*
1000
1300*
1000
1300*
1000
1000*
*0rijinally scheduled but subsequently eliminated.
-------
Appendix D
User Survey instrument
145
-------
I.
ASSIGNMENT INFORMATION:
CASE ID #:
SCREENING FORM
Date:QI]
Month
Date
Tiae:|
SITE #:J. I 1
INTERVIEWER NAME
Weather: s
(5-8)
PC | 2 | C m R 1 4
II, INTRODUCTION:
Hello, my name is
Researchers tt the Univer-
sity of Maryland are currently studying citizens' use of the Chesapeake Bay
and they are looking for certain types of people to aoiwer*questionnaire. I
would like to t sk you 4 questions to determine if you «re the type of person
we are looking for. First. . .
III. ELIGIBILITY: (20)
A. Do you currently live in Maryland, I 1 I (Go to B)
Virginia, (^2 I (Go to B)
Washington, D.C, or fTl (To to C)
Some other place? (4 ) (Go to F)
B. What county do you live in? (ENTER, COMPARE AND CODE)
MARYLAND
(Anne Arundel, Baltimore including city,
Carroll, Charles, Bar ford, Howard, (21[
MoatjoMry, Prince George'i) | 1 | (Go to C)
VIRGINIA
(Arlington, Fairfax, Loudoun,
Prince William) | 2 I (Go to C)
SOME COUNTY NOT LISTED fTI (Go to F)
c. How old tre you? ^2)
Less than 14 yein of age | 1 | (Go to F)
Over 14, but less than 18 years of age [ 2 | (Go to D)
18 years of age or older | 371 (Go to G)
D. Did you cone here . ..with other people including
someone you live with who is (13)
18 years of tge or older? j 1 I (GotoE)
. ..other people but none you
live with who is 18 years
of age or older? (JJ 'Go to G'
fTl (Go toG)
. ..by your self? {_3J
I would like to talk to the person you live with who as 18 years of
age or older and is here with you today. Point that person out to
me or tell me how and where to find him or her. (WHEN YOU LOCATE
THE PERSON, DETERMINE ELIGIBILITY BY VERIFYING OR REASKING ANY OR
ALL OF THE QUESTIONS IN PART III. BEFORE YOU BEGIN THE QUESTIONS
READ THE INTRODUCTION TO THE NEW PERSON.)
-------
(CODE 2 or 3 IN Iv BELOW AND SAY:)
You (do not live in the area/belong to the age group) that we wish
to survey on this study. Will you give me your name •nd telephone
number in case my supervisor wants to check on ay work?
NAME TELEPHONE ( )
(CODE I IN IV BELOW AND SAY:)
You are the type of person we want to interview for this study. The
questions will take 15 to 20 minutes. Can we start now? (AFTER
MEETING OBJECTIONS SAY:) Before I ask the questions I need to keep
a record of the people I speak to. What is your name, address, and
telephone number? what time did you enter the beach? What time
will you leave?
NAME
TELEPHONE
ADDRESS
IV. RECORD OF ELIGIBILITY (Screening Status)
TIME ENTERED BEACH :| I | | |
(^-17)
TIME LEAVING BEACH:!I 1 I I
(18-21)
Eligible PHi (INTERVIEW AND/OR COMPLETE PART V)
Ineligible because of
Residence f02J (STOP. DO NOT INTERVIEW. READ
STATEMENT FROM III F, RECORD
A8e foal INFORMATION AND TERMINATE CONTACT.
Refused Screening .... I 971
Language Barrier [^5) (RECORD DETAILS
IN NOTES SECTION BELOW)
Other (Specify) [06)
(22-23)
v. RECORD OF INTERVIEW STATUS
Completed foil (EDIT CASE AND SHIP TO RTI)
Partial/Breakoff. ... QL (RECORI) DETAILS IN NOTES
Refused ("97J SECTION BELOW)
Language Barrier ....
0ther:(SpeCi£y) —' (24-25)
VI. NOTES :
COLS. 26-72 * blank
COLS. 73-80 = SITE/CASE 1
-------
CHESAPEAKE BAY BEACH USE SURVEY
Conducted by Research Triangle Institute
for the University of Maryland
NOTIOE • la/onatioa coataiatd o« tai* fern that vould pacaxt idmtificatioa of «ay iadividiul bit b»«e
eolltcttd with * fturtat** taat it vill ba bald U ttriet caafidaac* «ad vsad oaly by pcrioai
«|«iad IB or for purpetM of Uii mi*»j. All result* will b« (iBBUtixad (or ft«up* of pcopl*
•ad a* lafonutioa about iadividiuli vill »• nlt«i«d.
A. IDENTIFICATION
SITE NO. 1
CASE ID
Field Interviewer Name
Date Interview Completed
Month
B. INTRODUCTION
As I said Earlier, researchers «t the University ofMaryUadtre current-
ly studying citizen*' use of the Chesapeake Bay. I will • sk you «o«e
questions reprdinf your recreational use of the Chesapeake Bay, parti
cularly «t the beaches. I flso have to *sk IOM questions which will
tnable the researchers to establish profiles of typical users of the Bay.
There U no direct benefit froa taking pert in this study • nd you have
the right to refuse to e-nswer tny or til of the questions or discontinue
your participation tt tny time. The information that you provide will be
combined with that provided by other people who participate in the survey
to essure complete confidentiality tnd your uoe will not be released or
revealed to anyone other than Authorized project staff. It may be
necessary to recontact you later to verify the number of times you
visited beaches during the season. The results of this survey may be
helpful in • ffectively • llocating money to detains up the Bay.
cT BEACH UTILIZATION
10 How many of the following types of people ere in your party today?
Adults (tge 18 tad older)
Teenagers • ge 14 years to 18
Children under age 14
2. Are you at the beach today with your family? By family I mean people who
• re related to you by blood, marriage, or adoption, tnd who live with
you .
NO CD (7)
Yes fn
-------
3. (SAMPLE SITE NAME] is situated on the Western Shore of the Chesapeake
Bay. Since January 1. 1983, have you (and members of your family who
live with you) (READ THE FOLLOWING)
NO YES
a. Visited beaches on the Eastern
shore of the Chesapeake, for
example, shores close to
Cambridge, Salisbury, or Easton? . , . . • I LJ • • • I 2 I ( S)
b. Visited beaches on the ocean,
such tt Ocean City? [Tj 121 (9)
c. gone swimming from •boat in
the Chesapeake? • m
d. gone winning in public or
private twining pools? Ill • [_2J (22)
THIS SPACE INTENTIONALLY BLANK
-------
4 When you visit a Westers Short beach such at this 000, do you (aadMabcrs of your tastily who live with You) parti
pat* ia (READ THE rOLWwTMG)
4A. How sway adults 41. How aaay childre
NO (Nut activity YES (Aok Q.s4A.. participate ia * teenager* part
ACTIVITY »t qiwatioa) & 4B.) (ACTIVITY) cipate in (ACTIV
* b
iia( or wading?
b. hooch • ctivitioo-
tuabathing, picnicking,
ihelling?
c. beach •ctivitios-
playgrouad? . .
d. boating?
Other AetiTitits?
(Specify)
. CD
. . (T] .
[T] . Q] . (i 7)
Q] [T] . (22) .
ca
CD EH . <*•?)
i J
U5-7?)
CD
(J0-J2 )
3s-s6
-------
S If you (or any mtmbm of your family who
describe why? (READ THE FOLLOWING)
with you) don't «o into the water, which •* the following. »tJ
5A. How ewny adults SB. How uoy eh
do not jo into *t«««ugtri
No (co to aext TCS (Ask Q.a 5A. •» the water becauae |o into the
R1ASON reason) *SB.) (XCASOM)T because (WA
+ *
»\
b.
e.
d.
»
f.
Don't kao«
No bathb
at beach
Person's
precludes
Tou or th
water is
Too Deny
Other res
(SKCIFY
,n,i
ous. facilities ^ m ^^
hailrk —
it m [2] (J2)
.«, («-*:>
1 1 1
Ui-«4) (6S-66)
mm W« t f e**e> —
dirty/vollutel . . . ITI 121 . (.87)
(S
jellrfiah . FH TS1 . ( ;)
'OD ..m m.c.) r
1 I !
53S)" " " " r^-7:)
COLS, 72-SO •» SITUCASE Ul 1
~n i i i
n
-------
6. How many years have you (and members of your family who live with you)
been coming to the beaches on the Western Shore?
Number of years I I I (12-12)
7. Have you (or anyone in your family who live with you) changed your
swimming habits in the Chesapeake because of changes in the Bay's water
quality?
No ITI (GO TO Q.8)
Yes (stopped) ITl
Yes (started) f~3~l (13)
A. In what year did you (or members of your family) last change your
svinaing habits in the Chesapeake because of changes in the Bay's
water quality?
m
Year Ill ( 14-15)
The next few pueftioni deal with the frequency with which.you (and your
family) visit beachesfnd the cost related to-visits aadetofach beach.
HAND RESPONDENT CARD A.
THIS SPACE INTENTIONALLY BLANK
-------
I. look »t the beache* listed oa thia card. Pleaae give M the letter that appear* beside the MM of each I
you (*a4 your fully) riiitid ilact January 1. 1913. Pita** iaclude this vi»it to thi* beach. (CHECK (J) i
VISITED AMD THEJI ASK QOtSTIOMS SA aadTF
BEACH VISITED
Q.t. Chock if
visited
Q.IA. How many trip* did you
(aad your faaily) aiake
to (BEACH)ia 1983?
Q.SB. How Nay tript
(aad your famil
to (BEACH) in
»\ tody Point St. Park
b. Tort Seallwood ...
c. Bay Ridjt Beach
d. Kcrrisftoa Harbor
». Kurtx Pltacur* Beach
f. Caa» derrick ....
«. Brwry Point Beach
h. Chtsapeaka Beach
t. lorth Beack
m a
. . 0748)
..... ma
rn (22-231
• (mu) DTI • (27-28 )
. (M)
CD. 6
jlfl .0/^1
..... GO . («) ......... 1X1(42-43 ) ........ m. . I
0
• «M*)
- («) ......... . m , .'(28-ss)
Q] . (M) .......... lCf\
Ill .
J. lod awl Reel Dock ~. (cl) (M-«J| . | 171 I - (s
k. Poiat Lookout St. Park . . . . m . (»») I I I . $7-88 )
1. £!•'• Beach . £1 . (J) f\\ - ( 24) . .
•. Horiaatova Sooth [T] .(»).... I I I . ( 7-4 ) .
. 02-13 )
COLS. 71-80 »»»SITI/CAi
.(
•in-
a. Hlaai Sooth (Baltiaore) . . .
. CD
0 Rocky Poi»t Park m 3
p. Coarad* • R«th Villa [T] . («) . . '.
4. Porter Mew Park m . (26) .... ll-l (2748 •)....».. IN. (
. (27-18 ) .......... m | . (
r. Other (SKCX7T)
32-33 )
-------
9. Do YOU (and you* faaUly) P1" 'o "*• "T trtp* to «*T of *** h«ach«i listed ou thia card during the reMiader of
sutJBer?
° { (GOTO 0.10) (38)
Doa't kao« m t
Tea GO
A. Which of these beaches do You plaa to visit this sue»erT Pleaae give aw the letter that appears beside the
of «ach beach (CRECX(V)EACH BEACH(ffMTIOXCO AM)THEN ASK Q.9I FOR £ACH BEACH.)
Q.9B. How MBJT trips do you («nd your
Q.9A. Check if pica fMily) iatnd to take te (BEACH)
BEACH to vi«it Skis i«Mt?
»\ Sandy Poiat St. p a r k (TI . (37) I I I . (38-3$)
b. Fort Saallwood m • up ) I I I . (
»
k. Point Lookout St. Park fTI . (S?) m .W-49)
1. £!•'• Btaeh • • CD . (»> I I I . (72-72)
I COM. 73-SO = SITE/CASE 03 I
«. Heriaatow Baach . . GD • f ') • • • '• • J • ( *-J )
n. HiaaU Buck (Baltlaara) • • S . (4) I I I .( S-6)
0. Rocky Point Pork • • C2 . (7) I I I • ( 4-9)
p. Conrad's Ruth Villa • • H • («) I I I . (JJ-J2)
q. Pertat Htw Park QQ (33) I I I . (,24-1s )
r. Other (SPECIFY)
JT|.(J
-------
INTERVIEWER NOTE: ASK QUESTIONS 10A-H SEPARATELY FOR EACH BEACH CHECKED IN
QUESTIONS 8 AND 9A.
10. Now think • bout the cost of a typical (family) trip to (BEACH) in 1984.
A. What would you say the total •ntrance fee per day, including
parking, would be for you (and your family) »t fBEACH) ? (ENTER
AMOUNT IN COLUMN A BESIDE APPROPRIATE BEACHI.
B. What is the typical coat for you (and your family) to travel to
(BEACH) ? (ENTER AMOUNT IN COLUMN B BESIDE APPROPRIATE BEACH NAME) .
c. How much will it, cost you (and your family) per day in other
expenditures such «s food, hotel or camming fees, tnd entertainment
• t (w=)? (ENTER AMOUNT IN COLUMN C BESIDE APPROPRIATE BEACH
NAME) .
D. ADD AMOUNTS ENTERED IN COLUMNS A-C AND ENTER TOTAL IN COLUMN D.
E. How much travel tiM does it typically take you (and your family) to
make •round-trip fro* your home to (BEWE)? (ENTER TIME IN COLUMN
E BESIDE BEACH NAME) .
F.
How many miles if it round-trip from your borne to (BEACH)? (ENTER
NUMBER OF MILES II? COLUMN F BESIDE BEACH NAME) .
G. How much tiM do you (and your family) typically spend on the beach
per day when you visit (BEACH)? PROBE FOR BESTESTIMATE OF THE
NUMBER OF HOURS AND ENTER NUMBER IN COLUMN G BESIDE BEACE NAME.
H* What is the t verage number of days that you tnd your family spend on
• typical trip to (BEACH)? (ENTER NUMBER IN COLUMN H BESIDEBEACH
NAME)
-------
TABU 1
QUESTION 10
(Coatiauod) .
B«aCB
k. Point Lookout St.
Park
1. EU' a B«aca
a. Horgaatow Boaca
a. Hiaa* B*aca
(Bolt-se)
• Bocky Poiat Park
p. Coarad' a Butt Villa
a,. Pertar N«v Park
r. Other (SPECI1T)
A
Eatraaea
r««
( \
(2t-2t)
ffl
rn
m
( 2.1 )
ra
(*«-*•)
B
Travti
Coots
*
3-
4 j
(JO-tt]
(
j
1
J-
M.
*.
4 j
~~~)
1
(30.33)
( s-
*
1 1
U0-JJ)
Ei
I
Otter
VM
1 1
\
7-11)
(34-31)
| |
(
1 |
|
|
1
|
(3
\
r-ii)
4-1
1
1)
1 1
_r
(J4-U)
(
(j
1
7-11)
1-3
|
u
Total
Co
] 1
(11-17)
J 1
(39-44)
r
<
\
~\ 1
1 J '
|
|
1
(Jt-17)
} 1
(39-44)
1
(12-17)
\ 1
(J9-44)
| |
E
B-T
Tiaw
1
(l»-l>
( '
P3
ui-17,
0«-;» )
Z«-3*)
r
B-T
Hilts
\
(20-21)
«7-W)
COLS.
COU.
G
Tiaw
00
Batch
m
(i4-2$:
\
(Si-Si)
H
0*
I
Tr
n
Q
15-73 sblaak
73-80 = SITE/CASE I
1
(20-23)
(47-U)
COU.
COLS.
| |
(u
(
-23)
(51-52)
5
y
M-72 • aak
73-sO • SITE/CASE
• 1
(20-23)
1 1
(47-50)
COU.
COU.
m
(24-25)
n
T5F
s
39-72 • blaak
73-80 • SITE/CASE
(*0-2 3)
(47-50).
|
(24.
•25)
(S1-S2)
a
n
53:
-------
HAND RESPONDENT CARD B
11. Please read the water quality characteristics listed on this card and
rank them • ccording to how important they 9 re to you (ind your family)
when deciding whether ornot to visit • beach. Please rank them on a
scale of 1 to 5, with 1 being least important and 5 being most important.
RANK
a. Presence of cloudy water '. fSS)
b. Presence of floating debris or oil _____ (£S)
c. Presence of odors (s7)
d. Presence of jellyfish • . (58)
». Presence of seaweed §nd other
• quaticplants • (59)
HAND RESPONDENT CARD C
12. Each sussttr jellyfish, tlso called sea nettles, • ppearin the waters of
the Chesapeake Bay. lookt-t the statements on this card tnd tell me
which one describes your (and your faaily'i) behavior after jellyfish
tppear.
». Stop going to the beich when ^_^
jellyfish appear I 1 I
b. Still go to the beach, but less often . . . __G (60)
c. Continue to 80 to the beach, but
don't fo into the w«ter l_3 I
d. Don't consider presence of
jellyfish atoll
COLS. 61-72 = blank
COLS. 73-80 = SITE/CASE 12
-------
THIS PAGE INTENTIONALLY BLANK
-------
13* A. HAND RESPONDENT CARD A. We would like to find out bow beach users
perceive the water quality, the quality of the beach facilities, the
beach quality, *nd crowding ft the beaches on the Chesapeake Bay.
Please give me the letter that appears beside the name of tach beach
on this card that you (and your family)* re familiar with. By
"familiar," I mean that you know something tbout the beach t ither by
having been there or you have heard •bout it through some other
source. Please include this beach in the ones that you mention.
(CHECK COLUMN A FOR EACH BEACH MENTIONED.)
ASK QUESTIONS 13B-ESEPARATELY FOR EACH BEACH MENTIONED.
B. Do you consider the water quality tt (BEACE) Acceptable or unaccep-
table for twinning • rid/or other water •ctivities? (CHECK APPRO-
PRIATE CODE IN COLUMN B BESIDE BEACH NAME) .
C. Wftafabout the beach facilities, such fsspace, bathhouses, tables,
swinging guards, etc. ft (BEACH). Would yourste it «s • cceptable
or unacceptable? (CHECK APPROPRIATE CODE IN COLIMT C BESIDE BEACH
NAME) .
D. Think*bout the beach itself ft (BEACH). Do you feel that the
quality of the beach is 0 cceptable or unacceptable? (CHECK
APPROPRIATE CODE IN COLUMH D BESIDE BEACH NAME) .
E. Do you feel that the size of the crowd tt (BEACH) is 0 cceptable or
unacceptable? (CHECK APPROPRIATE CODE IK COLUMN E BESIDE BEACH
NAME) .
-------
TABU 2
QUESTIONS 13A-I
Beach
Sandy feint
st. Pttk . .
Fort Saallvoo
Bay Kidie
Beach . . .
. Harrington
Harbor . . .
Kurtz Pleaaui
Beach
. Caap derrick
. Breexy Point
Beach . . .
u Chesapeake
Death . . .
. North Beach
Rod and Keel
"tick ....
;. Point Lookout
St. Park . .
.. Ela'a Beach
t. Morgaatown
Beach
t. Miami Beach
(Baltimore)
i. Rocky Poiat
Park ....
p. Coarad't But
Villa . . .
q. Porter leu
Park ....
r. Other
(SPECIFY)
r— A — 1
Zheck if
••'liar
ffr
55
TT )"
Oj.
UJ)
S»
i
a-
en
CM)"
rri
rn
9-
•ffi.
' '™3
•$,
•a-
•3v
•ffr
•53
(IT)"
• CD.
• (28)
i 1 : r
Water
:eeptab 1 1
.m..
. CD • •
. m- •
.m-.
.03-
.m-
.23-
. m .
.CD-
. m .
. m .
.23-
-23-
. m .
. m .
.23-
.23.
. [73 •
lualitT
oaccep table
• O3 ci)
. 23 CM
. 23(">
.23(17)
. m (221
. 23
.23 (w)
• El (S4)
.23 (22)
• ffl (44)
Cti ini
Acceptable I UnacceptabLi
.CD.
. CD-
.23-
.23-
.m-
.23-
.23-
.23-
.03-
.O] (48) • 23 •
.Q3 (s4)
. 03 (««)
. S3 (64)
. 23 <«)
1 COU
. m ( 4)
. m c *)
. 1 2 1 (I*)
. m (»)
CD-
,03.
.03.
,23-
. 71-80 >
.03.
.23-
.23-
.OD.
p ( s )
.CE t:s
. GDCi
. m (20
.23 (2s
.CD (30
. m CM
.23(40
.23 (4s
. 1 2 1 (-50
.CD(sS
. me,
. rn (5^
..mo,
» SITE/ CASI
. .mc^
. . 23 t:
. . 23 di
. . n~' ca
-------
INTERVIEWER CHECKPOINT 1
CHECK THE RESPONSES TO QUESTION SB FOR THE NUMBER OF TRIPS THE RESPONDENT
HAS MADE TO TEE SAMPLE SITE IN 1984 AND QUESTION 9B FOR THE NUMBER OF'
TRIPS THE RESPONDENT PLANS TO TAKE TO THE SAMPLE SITE DURING THE REMINDER
OF THIS SUMMER. ADD THE NUMBER ENTERED FOR BOTH QUESTIONS AND RECORD IT
HERE. •»
TOTAL NUMBER OF 1984 TRIPS =
CHECK COLUMN D IN TABLE 1 FOR THE TOTAL COST OF A TYPICAL 1984 TRIP TO
SAMPLE SITE . ENTER THE AMOUNT
THE
TOTAL COST OF TYPICAL 1984 TRIPS
INSERT THE TOTAL NUMBER OF 1984 TRIPS ANDTHE TYPICAL 1984 COST PER TRIP
IN THE APPROPRIATE SPACES WHEN QUESTION 14 IS ASKED.
ENTER THE LAST DIGIT OF THE CASE ID NUMBER »
If last digit it
0
1
2
3.
4
5.
6
7.
9..
Ust $ amount below in
Question 14
$50.00
$45.00
.$40.00
.$3s.OO
$30.00
$25.00
$20.00
$15.00
$10.00
$s.OO
14. According to your responses to previous questions, you (and your family)
take •bout (TOTAL NUMBER OF 1984 TRIPS) to this site per year *t an
approximate cost of (TOTAL COST OF TYPICAL 1984 TRIP) per day. If your
costs.per day were to riit by (USE INITIAL AMOUNT FROM TABLE ABOVE),
would you still visit this site? Keep in mind that the costs of visiting
other sites on the Chesapeake or participating in other •ctivities would
rauia the
NO.
YES
1
2
(22)
IF NO, DECREASE THE DOLLAR
AMOUNT IN $5.00 Increments
UNTIL A "YES" ANSUER IS GIVEN.
WHEN A YES ANSUER IS GIVEN,
RECORD DOLLAR AMOUNT BELOW.
14.
IF YES, INCREASE THE DOLLAR
AMOUNT IN $5.00 INCREMENTS
UNTIL A "NO" ANSUER IS GIVEN.
WHEN A "NO" ANSUER IS GIVEN,
RECORD DOLLAR AMOUNT OF LAST
"YES" RESPONSE.
DOLLARS
(22-25)
-------
ENTER THE LAST
If last digit is
1 ....
2. ...
3. ...
4 ....
5
6
7 ....
8 ....
9
0. ...
INTERVIEWER CHECKPOINT 2
DIGIT OF THE CASE ID NUMBER =
Use $ amoun
Questions
$ 5
$10
$ 1
$20
. $25
.$30
$35
$40
. $45
$50
t below in
L5 and 16
.00
.00
5.00
00
.00
00
.00
00
.00
.00
15. Jellyfish «re frequently identified «s •nuisance to tvimm. Would you
(and your fully) be willing topty (AMOUNT FROM CP 2) per yeas io • xtra
state or federal taxes if jellyfish could be t laminated «s •nuisance
without tny tdverse ecological tffects?
NO roil
Yet 0 CH3 (26-2?)
Don't Know .... ^ m
INTERVIEWER CHECKPOINT 3
REFER TO COLUMN B IN TABLE 2 (QUESTIONS 13A-D)
WAS THE WATER QUALITY OF THIS SITE RATED AS UNACCEPTABLE?
No fTl •» (GO TO Q. 17)
Yes 0m" (28>
16. You indicated that the water quality «t this site is unacceptable for
•wining. Would you be willing to pay (AMOUNT FROM CP 2) in extra state
or federal taxes if the water quality were improved so that- you found
it Acceptable to swim here?
(29-30)
-------
17. The next few questions • re •bout you •nd your household. How many of
each of the following types of people live in your household? (READ EACH
OF THE FOLLOWING AND ENTER THE NUMBER OF EACH TYPE).
a.
Adults (age 18 ind older)
b. Children between the ages of 14 and H
c.
Children under fge 14
m
(32-32)
(33-$4)
(ZS-ZS)
18.
HAND CARD D. Which of the relationships listed on this card best de-
scribes your status in your household?
I. Grandparent
b. . Husband
c. Wife
d. Child
» Other Relative
f. I liveflone or with
unrelated individuals . '. . .
m
(37)
m
EJ
19. How many.years have You (and your faaily) lived in either Maryland,
Virginia or Washington; D.C.?
Number of years J.J.J. (38-39)
20. Do you or tny other lenber of your household own (READ THE FOLLOWING)
NO YES DK RF
• \ tboat?
b. •boattrailer? . . .
c. fishing tackle (rod,
reel, tackle box,
etc.)?
d. •recreational
vehicle (RV)? . . . .
m * * m *
cm=mzl*
TT
TT
m
* *
m
e.
f.
• swming pool? I 01 1 . .
Other recreational
iteu (SPECIFY)
roTi M
rsn . .
I 02
FoT
f02
. C2D -. dS • • EEI • • QD
9i"i. . nrn
(40-4:
(42-43
(44-45
(45-4?
. r?n . .
. nn . . irn
(so-s:
-------
21. Are you one of the principal wage earners in your household, «wage
earner but not the principal wage • arner, or «re you •homemaker, a
student orretired?
a. One of the principal wife earners
in the family .„»,,,,,,,»,, I 1 I
b. Awage»arner but not the principal
wage earner 6 0 IT]
c. Homemaker I 3 I (S2)
d. Retired 0 CD
e. Student CE3
f. Other (SPECIFY)
INTERVIEWER NOTE: ASK QUESTIONS 22 THROUGH 24 IF CODE 1 OR 2 IS MARKED IN
QUESTION 21. OTHERWISE GO TO QUESTION 26.
22. How many hours do you usually work per week?
Hours I I I (53-54)
23. How many paid vacation days will you have W lto@er in 1984, including
those you've Already Uken?
Vacation Dayi 1 I I (5s-56)
24. For your typical trip tothebeach, if you hadn't gone to the beach,
could you have worked •tsoat job instead?
No 0m
(57)
-------
25. if .You could have worked, what hourly wage Bifht you have been paid
specifically for the hours 'you Worked?
a. $3.35/hour I 01 I
b. $3.36-$5.00/hour
c. $5.00-$7.50/hour
d. $7.50-$10.00/hour
». $10.00-$15.00/hour % |JKI
f. $15.00-$20.00/hour • QJE
g. $20.00-$25.00/hour 0 OH
h. Over $25.00/hour 0 QD
i. Don'tKnow I 98 J
j . Refuse Q I 97 I
26. Are there • ny (other) major wage»arners in your family with you today?
No • El (00 TO Q.31)
^ i (60)
Ye« 0| (ASK Qf . 27-30) ' '
The next few questions • re about the other major waget-arner who is with you
today .
27. How many hours does he/she usually work per week?
Hours IT! (51-?72)
28. How many paid vacation days will he/she have 0 1 together in 1984?
Vacation Days . . . .
(83-64)
29. For the typical trip to the beach, if he/she hadn't gone to the beach,
could he/she have worked • t some paying job instead?
1)0 .........
Yes
-------
30. If he/ibt could have worked, what hourly rate would he/she have been paid
specifically for the hours worked?
• \ $3.35/hour foTl
b. $3.36-$5.00/hour ' 02 I
c. $5.00-$7.50/hour 0 "m
d. $7.50-$10.00/hour • m
». $10.00-$15.00/hour f"bT"l (66-$7)
f. $15.00-$20.00/hour • m
g. $20.00-$25.00/hour 9m
h. Over $2 5. 00/hour CITI
i. Don'tKnow 9m
j. Refuse
31. HAND CARD E. Which one of the categories on this card best describes
your (family's) incoae during 1984? Pleate give me the letter that
• ppears beside the category?
• • ££/ s . • . . rrb~i
b ron h rrn
c V m i
H am i rrn (68-69)
d 9 m J UU
e rsn * nn
f ron i rm
DK !~98~1
RF r?n
32. CODE SEX BY OBSERVATION.
Female I 1 !
Male FTI
-------
33. CODE RACE BY OBSERVATION.
White
Black m
Oriental FT1
Other (SPECIFY)
. . . . m
I COLS. 72-80 = » SITE/CASE 141
Thank you for participating in our survey of btach users. Your responses will
be helpful to u»0nd hopefully to the Sutetnd Federal government* in deter-
•ining the itatui of sviaaiBf and other •ctivities on Chesapeake Bay.
-------
TELEPHONE CONFIRMATION QUESTIONS
FOR EACH BEACH LISTED ASK 34. A. AND B. BEFORE GOING TO KEX1 BEACH .
34. A.
B.
During 1984 how many times have you (or members of your family
live with you) visited (ENTERED NAME OF BEACH)?
RECORD ANSWER IN COLUMN A AND ASK:
During the remainder of 1984 how many times will you (or members
your family who live with YQU) visit (ENTER NAME OF BEACH)?
RECORD ANSWER IN COLUMN BAND GOTO NEXT BEACH.
34.B
PLANNED
VISITS
who
of
BEACH
•\ Sandy Point St. Park ....
b. Fort Smallwood
c. Bay Ridge Beach
d. Her ring ton Harbor
» Kurtz Pleasure Beach ....
f. Canp Merrick .
g. Breezy Point Beach
h. Chesapeake Beach | |_
i. North Beach I j_
j. Rod tnd Reel Dock I I
k. Point Lookout St. Park . . .
1. Elm'sBeach
m. Horgantown Beach I I i
n* MiMi Beach (Baltimore) . . I I I
o. Rocky Point Park I I I
P« Conrad's Ruth Villa . . . .
q. Porter New Park I I I
r. Other (SPECIFY) I I I
(1-2)
(3-6)
(13-14) •
(17-18) -
(21-22) »
(25-28) •
(29-30) •
(33-34)" '
(37-38)"
(41-42) •
(45-46) •
(49-50)"
(53-54) •
(57-58) •
(61-62)-
(65-66)*
(69-70)
OZJ
en
(3-4)
(7-8)
(11-12)
(25-16)
(19- 20)
(23-24)
(27-28)
(31-32)
(35-36)
(39-40)
(43-44)
(47-48)
(51-52)
(55-56)
(59-60)
(65-64)
(67-68)
(71-72)
I COLS. 73-80 SITE/CASE IS I
AFTER PART rSAY: Thank you for taking part in the survey tnd tilkinf to
•e today. 'The information you-have provided will be very helpful in determin-
ing use of the Chesapeake.
-------
CHESBEEBKE BAY BEACH USE SURVEY
TELEPHONE CONFIRMATION
SCRIPT AND INTRODUCTION
I am calling to speak with ( INSERT NAHE ). IF NOT THERE,
FIND OUT WHEN HE/SHE WILL BE AVAILABLE. WHEN AVAILABLE SAY:
My name is (INSERT NAME ) »nd I'm calling for the University of Maryland.
On (INSERT DAY AND DATE you were interviewed »t (INSERT THE NSME OF THE BEACH
WHERE INTERVIEW TOOK PLACE) by • lady working for the University on • survey
about the Chesapeake Bay. Do you remember that interview? (IF YES PROCEED.
IF NO, TRY TO REFRESH THE RESPONDENT'S MEMORY BY TELLING HIM OR HER THE TYPE
OF QUESTIONS ASKED. IF STILL NO, TERMINATE THE CALL).
I'm calling today to confirm some of the information you reported the day
you were interviewed. I want to know which beaches you have visited or plan
to visit on the Western Shore of the Chesapeake Bay. I will read through*
list of beaches •nd • sk •bout each one individually. First,
GO TO QUESTION 34.A.
------- |