|
An Economic Evaluation of Nitrate in Groundwater: A Contingent
Valuation Survey in Northwest Florida
A THESIS
Presented to the Faculty of the College of Engineering Sciences, Technology
and Agriculture and
The School of Graduate Studies and Research
In Partial Fulfillment of the Requirements
For the Degree of Master of Science
In Agricultural Sciences
Fall 2003
By
Carmen J. Lyttle-N'Guessan
The members of the Committee approve the thesis, entitled, An Economic Evaluation of
Nitrate in Groundwater: A Contingent Valuation Survey in Northwest Florida of Carmen Juanita
Lyttle-N'Guessan defended on November 13, 2003.
Mchal Thomas, Ph.D.
Professor Directing Thesis
Cassel Gardner, Ph.D.
Committee Me er
Zacc Olorunnipa, Ph.D.
Committee M
a s,Ph.D.
Committee Member
Approved:
Mitwe Musingo Ph.D., Coordinator
Gduate Programs, CESTA
Verian D. Thomas, Ph.D., Associate Dean
Academic Programs, CESTA
Charles Mag eOD., Interim Dean
College of Engineering Sciences, Technology
and Agriculture
Chanta Haywood, P .., Dean
School of Graduate Studies and Research
TABLE OF CONTENTS
Signature Page......................................................................................................... iii
Table of Contents .................................................................................................... iv
Acknowledgement.................................................................................................... vi
List of Tables .............................................................................................................. vii
Abstract .........................................................................................................................ix
Chapter One
Introduction ...................................................................................................... 1
I. Objectives of the Study ................................................ ........................ 3
II. Justification of the Study....................................................... ....... .......... 4
Chapter Two
Literature Review ................................................................................................ 6
I. Background.............................................................................................. 6
II. Contingent Valuation M ethod Estimation............................. ............ .............. 8
III. The CV Survey Instrument.......................................................................... 12
Chapter Three
M methodology ................................................. ............................................... 14
1. Design and Structure of the CV Survey Instrument ............................14
A. CV Guidelines.................................................................................. 14
B. The NOAA Panel Report Methodological
Guidelines for CV Instrument........................................................ 15
C. Development of the CV Survey Instrument.................................. 19
D. Pre-test and Approval of the Survey Instrument.................................. 27
E. Preparation of the Final Survey Instrument......................... ............ 27
F. Sample Selection and Data Collection Procedures................................ 28
G. Implementation of the CV Survey Instrument................................... 29
H. Data Analyses .................................................................................... 34
Chapter Four
Results and Discussion .............................................................................. 40
Part 1. Survey Response Results .............................. ...................................40
Part 2. Results and Data Analyses............................................... ............... 58
Chapter Five
Summary and Conclusion ......................................................................... 64
Future Studies.............................................................................................. 68
Limitations of the Study.................................... ......................................... 68
A ppendices ............................................................................................................. 70
R eferences..................................................................................................................... 83
ACKNOWLEDGEMENTS
I wish to express gratitude to my major advisor, Dr. Michael Thomas and my
committee members Drs. Cassel Gardner, Zacch Olorunnipa, and Nick Stratis, for their
invaluable help, facilitation and hours of dedication in completing this thesis. I would like
to make special mention of my major advisor, Dr. Michael Thomas, for his tireless
support in the pursuit of funding for the survey.
I would like also to extend my gratitude to Drs. Katherine Milla, Larry Robinson,
Odemari Mbuya, Verian Thomas and Mitwe Musingo, for providing the financial
contributions and stewardship needed in completing this research. I wish also to express
my gratitude to the Environmental Cooperative Science Center of the National Oceanic
and Atmospheric Administration, for their substantial financial contribution to the survey,
and to the College of Engineering Sciences, Technology and Agriculture (CESTA) and
School of Graduate Studies and Research, for their financial contributions to my research.
I am grateful to Mrs. Rosalyn Ross for her help in data entry during the survey
implementation process, and also for her guidance in arrangement and pagination of my
thesis. I thank also my colleagues who assisted in the survey implementation process.
Thanks to my family, relatives and friends for their immeasurable support and
encouragement throughout the long and sometimes tedious process of earning this
degree. Finally, I would like to acknowledge collectively, the moral support of the faculty
and staff of CESTA.
LIST OF TABLES AND FIGURE
Tables:
1. Comparison of procedures followed in implementation of the survey................ 30
instrument.
2. Survey response rate .................................................................................... 40
3. Pooled survey response rate ........................................................................ 41
4. Question 1: Preferences of the respondents concerning public spending on
public policy issues ...................................................................................... 43
5. Question 1: Attitudinal measures of the respondents .......................................45
6. Section 2: Response by institutional-based payment vehicle for proposed
program ............................................. .................... ....................................45
7. Section 2: Responses of survey by offer prices for proposed program ..............46
8. Section 2: Question 2; yes/no responses to WTP..............................................46
9. Question 3: Frequency distribution for observed WTP ....................................48
10. Question 4: Was the most you are willing to pay $0, because:.......................... 49
11. Question 5: Have you had your well tested for nitrates? ....................................50
12. Question 6: In your opinion, the quality of water you used at home is: ..............50
13. Question 7: For drinking, which do you use? .............................................51
14. Question 8: Number of years lived at current location .....................................52
15. Question 9: Number of persons in the household by age and sex
category ....................................................................................................... 53
16. Question 9: Number of males....................................................................... 53
17. Question 9: Number of females ................................................................. 54
18. Question 10: Highest level of formal education completed .............................. 55
19. Question 11: Respondents' employment status.............................. ............. 56
20. Question 12: Respondent's gross household income .......................................57
21. Description of variables used in this research .............................................. 58
22. Mean WTP by institutional basis ............................................................. 59
23. Summary of probit model results ............................................... ............. 60
24. Actual vs. predicted yes-no responses.......................................... .............. 61
25. Summary of cross tabulation .................................................................... 62
26. Study data vs. U. S. Census Bureau ............................................ ............... 63
Figure:
1. Map of study site, Franklin and Gulf counties, Florida...................................28
ABSTRACT
A double-bounded dichotomous choice contingent valuation survey was
administered to 2,000 residents of Franklin and Gulf Counties, Florida, with the objective
of documenting any bias as a result of the institutional basis in the hypothetical payment
vehicle. The mean willingness to pay (WTP) estimates for state and private payment
vehicles are $4.39 and $5.08 per month respectively, significantly different at the 0.15
level. This is evidence of the importance in choosing the institutional foundation for the
payment vehicle. The major determinants of WTP appear to be education, perception of
existing water quality and the attitudes of respondents towards the environment.
However, WTP responses may vary temporally and spatially with differences in
demographics and goods and services valued.
CHAPTER 1
INTRODUCTION
Groundwater is a primary source for drinking and domestic water use in the
United States of America. According to the U. S. Environmental Protection Agency
(USEPA, 1995), over 20 percent of water withdrawals come from groundwater and more
than 50 percent of the nation's population depends on groundwater for domestic use.
Florida has shown high dependence on groundwater with 90 percent of its population's
drinking water coming from groundwater. Because groundwater is important to the U. S.
water supply, there are concerns over the present levels of nitrate pollution.
Nitrates occur naturally and as a result of various human activities such as
agriculture, forest management and septic systems (Leeds and Brown, 1992; Hallberg
and Keeney, 1993). Federal nitrate concentration standards allow for a maximum
contamination level (MCL) of 10 milligrams of nitrate per liter of water. A U. S.
Geological Survey (USGS) report (USGS, 1995) showed that nitrate concentrations in the
nation's groundwater supply have increased steadily to the point where nine percent of
wells tested in 1995 exceeded federal standards compared to two point four percent in the
1990 report.
Excess nitrate in drinking water is a potential risk to human health. An U.S.
Environmental Protection Agency report stated in 1998 that nitrate is one of two
contaminants (the other is bacteria) that pose an immediate threat to human health when
concentrations exceed recommended levels, and suggested that water above 1 mg/L nitrate
concentration should not be used for feeding infants. The most commonly reported
illness from drinking nitrate-polluted water is methemoglobinemia (blue baby syndrome)
which affects infants (Pierzynski, Sims and Vance, 2000). According to Rail (1989), the
10 mg/L federal guideline for nitrate concentration is conservative and levels below it are
probably harmful. To many, the long-term health consequences of nitrate contamination
are of great concern since there is no definitive information on which to base
recommendations (Water Quality Initiative 101, 1995).
Public concerns regarding the potential nitrate pollution of groundwater, and the
subsequent health and environmental risks have created a demand for public policies to
more closely examine and monitor water quality. Best management practices are
recommended as control measures but more rigorous controls may need to be
implemented. In either case, policies to improve groundwater will benefit from a rational
approach that includes an economic assessment (Taylor and Frofberg, 1977). Assigning
costs to environmental disamenities (e.g. nitrate-polluted water) or benefits to
environmental amenities (nitrate-reduced water) becomes important in determining the
health risks of nitrate in groundwater.
One approach is to elicit people's willingness to pay (WTP) to avoid nitrate-
polluted water. Contingent valuation (CV) is one method that may be used to estimate
WTP. With a CV approach, participants are asked in a survey to estimate what they
would be willing to pay for nitrate-reduced drinking water. When aggregated with other
individual responses, their answers can be used to estimate a demand function for the
benefit of nitrate-reduced water (Mitchell and Carson, 1989).
While CV is widely acknowledged in the literature as a method for measuring the
benefits/damages for environmental amenities/disamenities (Mitchell and Carson, 1989),
critics have identified several potential limitations. For example, because the survey
relates a contingent or hypothetical scenario to the respondent, great care must be taken
in designing the survey instrument. Any information that is provided in the CV survey
instrument could potentially influence the respondent's answer (Mitchell and Carson,
1989). A potential sensitivity in CV design is the way participants may pay or receive
compensation. One concern is the sensitivity of WTP to the institutional basis of the
method of payment. For example, should one pay the state or a private firm? Previous
studies have shown that the WTP may be negatively affected if a participant doubts the
ability of the institution to act effectively, or positively if he perceives that the chances to
free-ride (avoid paying for a desired amenity) are improved. Thus responses can be
biased by a strategic answer where individuals may intentionally understate or overstate
their WTP if they perceive that by so doing they can favorably influence the policy
outcome (Ajzen and Peterson, 1986; Mitchell and Carson, 1989, Arrow et al., 1993,
Whitehead and Van Houtven, 1997).
I. Objectives of the Study
The objectives of this study are to:
(1) Document the sensitivity of CV design to the hypothetical payment
vehicle by developing and administering a CV instrument and using
the data to develop an economic model that explains choice behavior
Specifically:
(a) Develop a double-bounded dichotomous choice survey
instrument, with two different institutional-based payment
vehicles and four starting points for price.
(b) From the survey response, estimate the mean and median
willingness to pay for nitrate-reduced drinking water.
(c) Develop an economic model to explain consumer choice
behavior.
(d) Compare the difference in mean WTP between the payment
vehicles to determine if the institutional-basis of a payment
vehicle affects WTP.
H. Justification of the Study
There is concern about the detrimental health effects of excess nitrate
consumption in drinking water. This concern is particularly relevant in Florida, where
90% of the population consumes drinking water that comes from groundwater (USEPA,
1995).
Although numerous agronomic studies have been conducted to assess the level of
nitrate in groundwater, to date, no research has been conducted in Florida to determine
the cost of nitrate pollution to consumers of groundwater. While there may be a desire to
protect drinking water from nitrate pollution, policymakers need to better understand the
cost of water-borne nitrates. In so far as the process of contingent valuation is one key
method to determining this cost, it is important to document the role of the hypothetical
payment vehicle in estimating willingness to pay.
5
More specifically, this research will estimate benefits from groundwater
protection in Franklin and Gulf counties Florida, and contribute to CV methodology by
testing the importance of institutional basis in designing the hypothetical payment in a
CV survey.
CHAPTER 2
LITERATURE REVIEW
L Background
To assign value to many types of environmental damages, it is necessary to use
non-market valuation methods. These are methods that permit the estimation of value for
goods and services lacking well defined markets, such as environmental amenities.
Typically, these methods follow two different approaches: revealed preference and stated
preference. Both approaches have well developed theory and the literature is replete with
applications (Mitchell and Carson, 1989).
The revealed preference approach is based on observing consumer transactions
with goods from complementary markets (Carson, Flores and Meade, 2001). For
example, the value of a recreational park may be determined by assessing the related
costs of travel; or in the case of livestock odor, farm proximity on the price of residential
property. Within a revealed preference approach, assessing the valuation of water quality
could be determined by observing consumer behavior for related goods, such as bottled
water or water treatment systems. However, because many people may not have access to
bottled water or treatment systems, this may underestimate the importance of water
quality. (Whitehead and Van Houtven, 1997).
On the other hand, rather than relying on the observed behavior of individual
choices, the stated preference approach directly asks individuals to value incremental
levels of environmental amenities, e.g. a specified quality of drinking water (Mitchell and
Carson, 1989). The contingent valuation method is one stated preference technique that
CHAPTER 2
LITERATURE REVIEW
L Background
To assign value to many types of environmental damages, it is necessary to use
non-market valuation methods. These are methods that permit the estimation of value for
goods and services lacking well defined markets, such as environmental amenities.
Typically, these methods follow two different approaches: revealed preference and stated
preference. Both approaches have well developed theory and the literature is replete with
applications (Mitchell and Carson, 1989).
The revealed preference approach is based on observing consumer transactions
with goods from complementary markets (Carson, Flores and Meade, 2001). For
example, the value of a recreational park may be determined by assessing the related
costs of travel; or in the case of livestock odor, farm proximity on the price of residential
property. Within a revealed preference approach, assessing the valuation of water quality
could be determined by observing consumer behavior for related goods, such as bottled
water or water treatment systems. However, because many people may not have access to
bottled water or treatment systems, this may underestimate the importance of water
quality. (Whitehead and Van Houtven, 1997).
On the other hand, rather than relying on the observed behavior of individual
choices, the stated preference approach directly asks individuals to value incremental
levels of environmental amenities, e.g. a specified quality of drinking water (Mitchell and
Carson, 1989). The contingent valuation method is one stated preference technique that
may be used to directly determine the value people place on environmental amenities or
disamenities. With the CV approach, individuals are asked to state the value they
personally place on specific water quality in terms of either their WTP to maintain that
level of water quality or their willingness to accept (WTA) compensation for the loss of
that level of water quality. Their response, when aggregated over individuals, can be used
to estimate the market demand for a specific level of water quality. The choice between a
WTP or WTA format is a matter of property rights for the existing environmental
amenity (Mitchell and Carson, 1989).
Ciriacy-Wantrup (1947) was the first to suggest using stated preference, and
described how to use direct interviews to measure non-market values associated with
natural resources. However, Davis (1963) was the first economist to actually apply the
CV technique when he surveyed visitors to rural Maine and directly determined the
benefits of outdoor recreation in forested areas. He also set the foundation for a market
bidding method. Several important CV researchers followed Davis. Krutilla (1967)
measured benefits from fresh water quality by observing changes in market prices of
commercial activities such as commercial fishing, irrigation and water treatment while
Mitchell and Carson (1986) studied the benefits of reducing drinking water risks and
bond issues. Poe and Bishop (1992) studied groundwater protection program to prevent
contamination while Whitehead (2003), and Whitehead, Hoban, and Clifford (2000),
emphasized the benefits of safe water and effects of perceptions. Still others such as
Hanemann (1978); Binkley and Hanemann (1978), and deZoysa (1995) worked on
valuing the effect of water quality and wetland habitat protection. deZoysa (1995)
conducted research on programs to enhance water quality and wetland habitat in
northwest Ohio and considered the economic value of controlling the disamenity of
nitrates from surface water. Crutchfield and Cooper (1997) measured nitrate risk in
drinking water in four regions. Finally, Clemons, Collins and Green (1995) emphasized
wellhead protection programs to eliminate risk of nitrate contamination.
Within the field of welfare economics, CV is widely recognized as a robust
approach to dealing with compensating and equivalent measures of welfare changes.
Contingent valuation was included as one of three methods (the other two were the travel
cost and the unit day value method) recommended by the Water Resources Council
(1979) and accepted for determining project benefits. More recently, the U.S. Army
Corps of Engineers has begun to use the contingent valuation method to measure project
benefits. Contingent valuation has also been recognized as an approved method for
measuring benefits and damages under the Comprehensive, Environmental Response,
Compensation and Liability Act of 1980 (CERCLA), according to the final ruling by the
U. S. Department of the Interior (USDI, 1986).
I. Contingent Valuation Method Estimation
Because a survey is used, CV is often referred to as a direct rather than indirect
approach of determining the value of non-market goods. According to Randall, Hoehn,
and Brookshire (1983):
Contingent valuation devices involve asking individuals, in survey or
experimental setting, to reveal their personal valuations of increments (or
decrements) in unpriced goods by using contingent markets. These markets
define the good or amenity of interest, the status quo level of provision and the
offered increment or decrement therein, the institutional structure under the
good is to be provided, the method of payment, and (implicitly or explicitly) the
decision rule which determines whether to implement the offered program.
Contingent markets are highly structured to confront respondents with a well-
defined situation and to elicit circumstantial choice upon the occurrence of the
posited situation. Contingent markets elicit contingent choices.
Mitchell and Carson (1989) have shown that CV is based on the principles of
consumer preference, which assume an economic agent (individuals, households,
consumers, or firms) will select the most preferred choice from a bundle of proposed
choices and through their selection attempt to maximize overall utility. Furthermore,
consumer preference is based on the axioms of completeness, reflexivity, transitivity and
continuity (Varian, 1992).
CV attempts to determine the amount of compensation paid (WTP) or received
(WTA) that will restore an individual to his initial level of utility after experiencing an
increment or decrement in welfare. CV permits the theoretically correct measure for
changes in welfare due to losses of goods or services lacking operational markets and is
based on the concept of Hicksian demand and compensating and equivalent variations.
The compensating variation defines the amount of compensation, paid or
received, that would return an individual to her initial welfare level after the change. On
the other hand, equivalent variation defines the amount of compensation, paid or
received, that would bring an individual to her welfare level if the change did not occur.
The equivalent variation uses an alternative level of utility as a base while compensating
variation uses the initial level of utility as the point of reference. WTP and WTA allow
for both the compensating and equivalent variations, and their choices depend upon the
circumstances consumers face or the property rights assigned to the environmental
amenity.
The main objective of a CV is to obtain accurate estimates of benefits (sometimes
cost) of a change in the level of allocation of some public good. Several approaches are
used in contingent valuation to elicit values from subjects/participants. The most common
approaches are:
1. Open-ended technique: The respondents are simply asked to state the
maximum amount they are willing to pay (accept) for the non-market
good in the hypothetical market (Mitchell and Carson, 1989, 1995).
2. Iterative bidding technique: The subject is asked if he would pay
(accept) some predetermined amount, the question is repeated with
higher (or lower) amounts depending on the initial response
(Brookshire et al., 1981; Mitchell and Carson, 1989).
3. Dichotomous choice (DC) technique: Subjects are asked whether they
would pay (accept) some specific amount for the non-market good for
the hypothetical market. Different amounts are assigned to the
respondents (Rae, 1983; Mitchell and Carson, 1989). The approach
uses a large number of predetermined prices, chosen to include the
expected WTP amounts of most respondents for the amenity. Each
respondent is asked if she is willing to pay a single one of these prices
for the amenity by responding yes or no to the offered price. The
prices are randomly assigned to respondents so that each price is
administered to an equivalent sub-sample.
While the first two approaches are still applied widely, the DC approach to CV
has gained wider theoretical support in the literature. The DC approach has several
advantages over other elicitation techniques such as bidding and payment card (Mitchell
and Carson, 1989; Hanemann and Kristrom, 1995). The DC CV approach is less stressful
for the respondent, in that it simplifies the respondent's tasks to declaring his WTP. The
respondent is required to make a judgment about a given price; a type of judgment
performed frequently by consumers and by those who have experienced voting, and as
such, can be easily included in a mail survey (Mitchell and Carson, 1989). The DC
format provides the same set of guidelines for all respondents and is incentive-
compatible, so it is less vulnerable to strategic behavior than other methods (Zeckhauser,
1973; Hoehn and Randall, 1987; Weisberg, Krosnick and Bowen, 1989).
As an improvement, the double-bounded DC model (the yes/no question) is
followed-up by an open-ended question establishing an upper or lower bound for each
bid or offered price, and was later proposed and proved efficient as a welfare estimate by
Hanemann (1985) and Hanemann (1991) respectively. For econometric analysis of DC
data, statistical models such as probit and logit are frequently used through the
application of limited dependent variable software (Green, 1995).
According to Mitchell and Carson (1989), dichotomous choice was first utilized
to value an environmental good by Bishop and Heberlein (1979, 1980). Dichotomous
choice techniques have also been applied by Loehman and De (1982); Bishop, Heberlein
and Kealy (1983); Hanemann (1984, 1985); Randall and Hoehn (1989); Cameron and
Huppert 1991); Poe and Bishop (1992); Boyle (1993); Cameron and Quiggin (1994); Li
and Mattsson (1995) and deZoysa (1995).
II. The CV Survey Instrument
The CV method has been criticized for many reasons, a number of which are
particularly compelling. Examples include; the mechanism used by respondents to make
a choice, assumption of property rights (WTP vs. WTA), method of hypothetical
payment (payment vehicle), comprehensibility of the survey instrument, and the
definition of the proposed amenity (Arrow et al., 1993).
An essential part of the CV method is correct survey design and implementation.
Several sources of bias are possible in surveying human subjects. These biases can
include interviewer, starting point, payment vehicle, non-response, and elicitation
method, among others. Any information that is provided in a CV survey instrument could
potentially influence valuation and the WTP or WTA response, thus require careful
consideration (Mitchell and Carson, 1989).
The WTP and WTA responses may be significantly affected by the design and
application of the survey instrument. One specific concern is the formulation of the
hypothetical payment vehicle; the means by which the subject will pay for an
environmental amenity. These may include taxes, utility bills, higher prices and entrance
fees (Mitchell and Carson, 1989). The payment vehicle for any proposed environmental
program must be plausible and credible, and should be viewed by respondents as a
reasonable, fair and equitable method to pay for the resource. It is hoped that people will
respond to CV questions in a way that reflects their value of the resource and not cloud
their response with an emotional reaction to the method of hypothetical payment. The
more realistic the payment mechanism, the easier it will be for one to respond to the CV
question, but the most realistic may not be neutral in the eye of the subject (Mitchell and
Carson, 1988, 1989). Individuals may have some strong negative attitudes toward certain
method of payment (e.g. taxes) that may affect results from the instrument (Ajzen and
Peterson, 1986; Mitchell and Carson, 1989). Therefore, to maintain the neutrality of an
instrument, the payment mechanism should be carefully chosen.
The format of the hypothetical payment, or payment vehicle, is particularly
important. One potential consideration in designing the payment vehicle is the choice of
the "institutional basis," that is, private vs. public institution. Whether a person must pay
a government agency or private firm may significantly influence the amount of their
response (Arrow et al., 1993). For example, the WTP may be negatively affected if the
respondent doubts the ability of the organization to act effectively. Thus, the size of an
individual's WTP or WTA can be affected by strategic bias, in which individuals may
intentionally understate or overstate their response if they perceive that their actions can
favorably influence the policy outcome (Ajzen and Peterson, 1986; Mitchell and Carson,
1989; Arrow et al., 1993; Whitehead and Van Houtven, 1997).
CHAPTER 3
METHODOLOGY
I. Design and Structure of the CV Survey Instrument
A. CV Guidelines
CV was initiated in 1947 and since the 1960s has been recognized and used by
many to value environmental damage or assign benefits for environmental amenities
(Mitchell and Carson, 1989). Additionally, with the formal admission of non-market
values in damage assessments by CERCLA, and the Department of the Interior (USDI,
1986), the use of CV in litigation and policy making has increased significantly.
However, with the increased use of CV as a damage assessment tool has come growing
concern about its potential limitations. Among these are concerns that responses may be
influenced by the respondents' emotions and not reflect the good or service being valued,
that the respondents may not understand the good or service they are asked to value and
that non-binding questions may not be considered seriously (Arrow et al., 1993). In
general, critics question the validity and accuracy of the CV method by arguing that
respondents cannot truly identify their willingness to pay (Vatn and Bromley, 1995;
Brox, Kumar and Stollery, 2003). With the growing chorus of criticisms the National
Oceanic and Atmospheric Administration (NOAA) assembled a team of economists,
including Nobel Laureates, to review the CV methodology for theoretical requirements.
Their 1993 report proposed methodological guidelines for CV to adhere to economic
theory, and has now become the accepted format for CV studies.
B. The NOAA Panel Report Methodological Guidelines for CV Instruments
The NOAA Report on CV methodology addresses many key issues that must be
considered in developing the CV survey instrument. These include:
1. Careful design of the choice mechanism
The NOAA report proposes that the valuation question be posed in a dichotomous
choice (DC) format. According to the report, open-ended questions are unlikely to
provide the most reliable valuations. For example, open-ended requests for willingness to
pay or willingness to accept compensation are often erratic and invite strategic
overstatement or understatement when the scenario lacks realism.
The DC approach is preferred because it is realistic and the simple approach
presented is not uncommon in real life (such as in voting and the market place).
However, while the DC format of questioning is recommended as the most desirable
form of CV elicitation, it may be subject to some forms of bias. Therefore, yes and no
(DC) responses should be followed up by open-ended question to mitigate the limitations
of using either the open-ended or DC format exclusively. The double-bounded DC
includes a binding ingredient that allows respondents to give reasons for responses,
which avoids 'yea-saying' bias; a direct response without thorough assessment.
2. Assumption of property rights, expressed as either willingness to pay or
willingness to accept
The panel also proposes that the WTP elicitation format be used over the WTA.
The panel believes that WTP is more conservative and increases the reliability of the
estimate by eliminating extreme responses that can enlarge estimated values wildly and
implausibly, reducing the reliability and usefulness of the information.
3. Establishment of substitute commodities and budget constraints
Respondents must be reminded of substitute commodities/amenities such as other
natural resources. The respondents must be reminded that their willingness to pay for the
proposed program would reduce their expenditures for other goods. This reminder should
be introduced forcefully and directly, prior to the main valuation question to assure that
respondents have the alternatives clearly in mind. According to the panel, while
consumers may be able to sufficiently make expenditure decisions regarding familiar
goods, they may have problems doing the same for unfamiliar environmental goods often
presented in CV surveys. If respondents are not fully informed of other expenditure
possibilities they may over-spend on the proposed environmental amenity and exceed
their hypothetical budget constraint. Though respondents in CV surveys may view the
hypothetical market seriously, if they are not informed of all the issues, they may respond
without thinking carefully about how much disposable income they have available to
allocate to all issues. Thus, the CV survey must remind respondents convincingly of the
very real economic constraints where spending decisions must be made. WTP for
environmental amenities is bounded by the individual's budget constraint, thus not only
her disposable income but also expenditure on other goods or group of commodities must
be considered.
4. Clear definition of the proposed amenity
The NOAA panel recommends that accurate and adequate information be
provided to respondents about the environmental program that is being offered. The
proposed amenity must be defined in a way that is relevant to the damage assessment.
According to the panel, if CV surveys are to elicit useful information about WTP,
respondents must have a full understanding of the good or service they are asked to value,
and accept the scenario in formulating their responses. The absence of accurate and
adequate information or sketchy details about the program being valued can result in
meaningless valuation and invalid estimates. The CV survey instrument must include a
carefully worded description of the resources or change in environmental quality being
valued (Anderson and Bishop, 1986; Mitchell and Carson, 1989; Arrow et al., 1993). The
NOAA panel suggests that vague or overly technical descriptions, as well as bulky
informational materials be avoided since they may reduce a respondent's interest, making
surveying difficult and affecting results. In addition, respondents should be given enough
information to motivate informed judgements about value without biasing their answer.
5. Permission to opt out of answering questions
A 'no-answer' option should be explicitly allowed in addition to the 'yes' and
'no' vote options on the main valuation question. Respondents who choose the 'no
answer' option should be given the opportunity to protest, and should be asked to explain
their choice. The no-answer option is an important component of the ability of the CV
technique to mimic an actual referendum.
6. Provision of supporting questions (cross- tabulations)
The panel proposes that the survey include a variety of other questions that help to
interpret responses to the primary valuation. The report should include summaries of
willingness to pay broken down by these categories. Among the items they suggest are
income, attitude toward the environment, education, prior knowledge of the site and
belief in the scenario. It is believed that these cross tabulations will prove useful in
interpreting and lending credibility to the responses, and may form adjustments that can
enhance reliability.
7. Careful pre-testing of the CV questionnaire
The panel states that careful pre-testing of the CV instrument is even more
important than ordinary surveys, since the CV survey contains much new and often
technical information. They recommend that careful pre-testing and pilot work be done
to determine if the respondents understand and accept the project description and
questioning reasonably well. Careful use of various pre-testing techniques to explore an
instrument's weaknesses is an effective way to enhance a study's reliability (Mitchell and
Carson, 1989).
According to the panel, although it may be evident that careful questionnaire
development was accomplished, it cannot be taken as a self-sufficient basis of validity. It
is important to anticipate the mistakes that respondents are likely to make and provide
opportunities to correct them (Mitchell and Carson, 1989). Since CV survey instruments
often present information that is new to respondents, the questionnaire should attempt to
determine the degree to which respondents understand the scenario.
8. Overall comprehensibility of the survey instrument
The panel recommends that the stated guidelines be satisfied without making the
instrument so complex that it poses tasks that are beyond the ability or interest level of
many participants. It requires that respondents of different socio-economic backgrounds
understand the wording, concepts and questions used in the survey. For example, the
survey instrument must contain clear, simple and understandable wording that clearly
directs the respondents through the questionnaire, that is, jargon and complex terms must
be avoided.
9. The method of hypothetical payment (payment vehicle)
The hypothetical payment vehicle is identified as crucial in portraying the
proposed program in believable and acceptable terms. According to the panel, the
hypothetical payment must be plausible, realistic, credible, and accepted by the
respondent. The payment vehicle should be presented fully and clearly, and the payment
scenario convincingly described, preferably in a DC format.
It is essential that respondents believe the hypothetical scenario is plausible;
otherwise responses may be biased. In addition, the type (and method) of payment
vehicle may influence a person's WTP for public goods and services (Mitchell and
Carson 1989; Arrow et al., 1993). For example, the institutional affiliation of the payment
vehicle may affect a person's WTP response if she doubts either reliability or fortitude of
the agency to act with prudence. Therefore, it is recommended that the CV practitioners
make neutral and familiar payment vehicles their choice when possible. Payment vehicles
should be accepted as reasonable and equitable methods of payment for the good or
service. In the process of documenting the sensitivity of CV design to the hypothetical
payment vehicle, it is very important to apply the NOAA recommendations to assure the
validity of CV results.
C. Development of the CV Survey Instrument
Development of this CV survey instrument in accordance with the NOAA panel
report involves a step-by-step process which is outlined in three sections. To illustrate the
extent of the survey design, the original survey instrument is dissected and discussed in
detail.
1. Section 1: The Introduction.
This section creates the atmosphere for informed decision-making and simulates a
budget constraint addressed in the NOAA fourth recommendation. To effectively and
correctly assign personal values, respondents need to weigh their alternatives subject to a
personal budget constraint (Mitchell and Carson, 1989). Respondents must be reminded
of substitute commodities and that their willingness to pay for a proposed amenity will
reduce their expenditures for other goods (NOAA recommendation #4).
Question one is included as part of the survey to remind and encourage
respondents to consider the options and issues that might be important to them, while
allocating their budget. Question one is also used to generate covariates (other
information used to explain WTP) for econometric analysis, such as attitudes toward the
environment, which may help to interpret the main valuation question. This question also
gives an overview of natural resources that are important and focuses on other public
goods such as education, health care, infrastructure, national defense/security where one
may desire to spend money. The inclusion of variables such as improved fire service and
better protection against terrorist attacks serve to remind the respondents of some issues
that are currently of major concern. The section also draws attention to some programs
(environmental and non-environmental) that the government funds. The respondents are
asked to indicate whether the government should retain current funding or spend more or
less money on the specified programs. Respondents are also given the opportunity to list
any other issues important to them. Section one of the survey instrument below addresses
NOAA recommendation number four:
1. There are many problems and Issues facing the United States today. Solving all of
them would be costly and people have different opinions about their priority. For each
of the areas of public concern listed below, please indicate whether you think it is a
good idea to: spend more money; retain current funding; or spend less public money
on these issues.
Issues Spend more Retain current Spend less
money money money
(circle one number for each issue)
Reduce global warming 1 2 3
Protect endangered species 1 2 3
Provide foreign aid to
reduce poverty in the world 1 2 3
Reduce air pollution
from automobiles 1 2 3
Improve facilities in 1 2 3
education
Improve healthcare 1 2 3
facilities
Find ways to improve
farming with less reliance 1 2 3
on chemicals and pesticides
Improve roads, bridges and
airports, etc. 1 2 3
Reduce crime to protect
public safety 1 2 3
Improve programs to
reduce terrorist actions 1 2 3
Improve roads in rural 1 2 3
areas
Improve local law
enforcement 1 2 3
Improve local fire
protection 1 2 3
Improve school bus safety 1 2 3
Preserve wetland
(e.g. marshes & swamps) 1 2 3
Protect surface water
quality (e.g. lakes and 1 2 3
rivers)
Protect groundwater
quality 1 2 3
(e.g. well water)
Please list any other Issues you consider as important.
2. Section 2: Description of Proposed Amenity
The second section of the survey instrument provides a description of the
hypothetical scenario the respondents are asked to value. To provide a valid answer, the
respondents must fully understand the good or service they are being asked to value, and
the proposed amenity must be defined in a way relevant to the damage assessment
(NOAA recommendation #3). To address the NOAA recommendation, this section starts
with describing the present level of the problem, its impact on the environment and
human health and finishes with the proposed improvement or amenity. This description is
a narrative comprised of information from available scientific sources detailing the cause
and effect of potential deterioration of groundwater quality.
Bulky, complex and unfamiliar materials should be avoided in the scenario, since
they can reduce the respondents' interest and affect responses (NOAA recommendation
#3). Simple wordings and examples are used in the following portion of section one of
the survey instrument to describe the scenario and keep the respondents interested and
focused, as well as assisting them to better understand the amenity being offered:
THE PROBLEM
Nitrates are naturally- occurring chemicals that are often detected in groundwater
(underground water used by wells). However, both the actions of nature and man
can often increase the level of this chemical in groundwater. While water provided
by public utilities is at safe nitrate levels, people who drink water from their own
wells may be exposed to measurable levels of this chemical over time. The United
States Environmental Protection Agency (EPA) states in their recent publication Is
Your Drinking Water Safe? That: "...only two substances ... pose an immediate
threat to human health whenever (safe levels) are exceeded: bacteria and nitrate."
Most often, the level of nitrate in Northwest Florida's groundwater is safe for
drinking. However, occasionally, unplanned and isolated events such as lengthy and
heavy rainfall may cause nitrates in well water to exceed safe levels.
If people drink enough nitrates along with their well water, it may affect
their health. For most people these effects may include some minor illnesses such as
diarrhea, reduced vitality and low birth weight However, scientists believe infants
less than a year old could suffer undue risk if they drink water with high nitrate
levels. The same EPA publication states that: "...in some infants, excessive levels of
nitrate have been known to react with the hemoglobin in the blood to produce a
...condition commonly known as blue baby. [Do] not give water to infants under
three months of age and do not use it to prepare a formula." Yet, blue baby
syndrome is rare, with only 200 authenticated cases caused by nitrates in the past 55
years.
Some scientists are concerned about unknown health risks from drinking
higher nitrate amounts over prolonged periods. According to a University of
Missouri publication, Understanding Your Water Test: "...as nitrate levels in water
have raised over several decades, there is growing concern about long-term health
consequences. Unfortunately, there is not much definitive information on which to
base recommendations."
The section continues with a scenario proposing a solution to the problem, how
the program would be implemented and a hypothetical method of payment. The design
and institutional affiliation of the payment vehicle may significantly affect an
individual's willingness to pay if he is not familiar with the payment scenario or doubts
the relevant agency can act efficiently (NOAA recommendation #9). For this study, a
utility bill payment vehicle is used, because of its familiarity and it avoids hot-topic items
such as taxes.
To examine the effect of institutional-based payment vehicles on WTP, some
instruments used the State of Florida (SF) while other listed a private utility (PU) as the
agency of payment collection. In accordance with NOAA recommendation #9, each
payment scenario explained the proposed implementation procedure and method of
payment portrayed in a real market. The following portion of section two describes the
proposed program for a private utility payment vehicle. The alternative instrument is
identically worded with the exception that "private utility company" is replaced by "State
of Florida:"
PROPOSED PROGRAM
In an effort to explore potentially new methods to reduce human exposure to
nitrates, researchers are reviewing different programs to control the levels of nitrate
in well water. This hypothetical program is just one of many being reviewed and
your opinion is very important to help researchers had better understand how you
value water quality.
How would the program be implemented?
A private utility company would place a nitrate-measuring device on your
wellhead to allow them remotely measure your nitrate levels. If nitrates exceed
the safe level, you will be notified and a second device will be immediately
installed on your wellhead to reduce nitrates
to safe level. This second device would remain in place until nitrates in your
well water source return to safe levels.
Participation in this program would be paid for through a monthly utility fee.
The money collected with this fee would be used exclusively to protect
Franklin and Gulf county residents against nitrate pollution in their drinking
water.
Next, the valuation question is introduced where monthly offer prices $1, $5, $10
or $20 are proposed as cost to participate in the program and are assigned randomly
across surveys. To satisfy the NOAA recommendation #1, a yes/no response about WTP
is requested in the following:
Now, with this information, please answer the following questions.
2. Scientists expect the program will cost your household a monthly payment of
$ Please circle the letter to your answer
A. YES I will be willing to pay this amount to participate in the program.
B. NO I would not be willing to pay this amount to participate in the program.
To avoid starting point bias, the bid prices are assigned randomly to the two
payment vehicles, SF and PU, for a total of eight versions of the survey instrument.
Participants are randomly assigned a single version of the survey instrument and asked a
yes/no question of her WTP. For example, a participant may receive a survey instrument
with a bid price of $1 and a proposed payment vehicle of PU.
Because of the NOAA panel recommendation #1, the yes/no DC question is
followed up by two open ended questions setting upper and lower bounds, (double-
bounded) in 3a and 3b below:
3a. Scientists cannot be completely certain about the program costs; the final cost
could be higher. What is the highest amount you would be willing to pay to
participate in the program? Please write your final answer. $_
Please go to section 3
3b. What is the highest amount you would be willing to pay to participate in this
program? Please write in your answer $
Respondents who choose the "no answer" option are given the opportunity to
protest, and asked to explain their answer (NOAA recommendation #5). Question four
provides respondents who report that they were willing to pay nothing for nitrate
reduction an opportunity to protest, that is, provide reasons for their choices. In question
four below, the study includes four DC options and an open-ended response to allow
them the opportunity to list other reasons:
4. Was the most you are willing to pay $0 because: (Please circle all that apply)
A. The program is not worth the cost to you.
B. You do not believe the program will work.
C. You are not comfortable with the idea of paying into a special
fund to provide clean water to protect human health.
D. You do not believe that there is a problem with your well water.
E. Other reason (please specify).
3. Section 3: Socio-demographic Background
Finally, in accordance with the sixth NOAA recommendation, the last section is
included to document the socio-economic background such as age, gender, education and
other information about the respondent and her household as covariates in determining
WTP values. An individual's socio-economic background may affect her choices and
values (Mitchell and Carson, 1989). Section three that follows includes other questions
that will help to interpret the responses to the primary valuation question:
For statistical analysis, we need to ask you a few questions about your household. The
information you give is strictly confidential and is governed under the Institutional Review
Board for human subjects as stated in the cover letter. However, please feel free to skip any
question you are not comfortable with.
5. Have you had your well tested for nitrate? (Please circle one letter)
A. Yes ---------> What was the level of nitrates found in your well?
Please fill in the blank if you remember the test result.
Otherwise, check the box.
Nitrate level Milligrams (mg)/liter
Don't know '-
B. No
6. In your opinion, the quality of water you use at home is... (Please
Circle one letter)
A. Very safe
B. Safe
C. Seems to be safe
D. Unsafe
E. Don't know
7. For drinking, which do you use? (Please Circle as many letters as apply)
A. Well water without additional treatment
B. Well water filtered or treated in the home
C. Purchased (e.g. bottled) water.
D. Other substitute beverages
8. How long have you lived at your current address? Indicate the number of years.
Years.
9. Please indicate the number of people in your household, including yourself, who fall
into each age and sex category.
Age Group Male Female
Under 6 months
6 months to 3 years
4 to 12 years
13 to 25 years
26 to 45 years
46 to 65 years
65 or older
All respondents: Please answer Q. 11 and Q. 12.
10. What is the highest level of formal education you have completed? (Please Circle one
letter).
A. Grade school or less
B. Some high school
C. High school graduate
D. Some college
E. College graduate
F. Graduate or professional degree
G. Other. Please specify
11. What is your present employment status? (Please Circle one letter)
A. Unemployed
B. Employed (part-time)
C. Employed (full-time)
D. Employed seasonally
E. Retired
F. Student
G. Homemaker
12. Please indicate your approximate household income before taxes in 2001. (Circle one
letter).
A. Under $10,000 E. $40,000 $49,999 I. $100,000- $124,000
B. $10,000- $19,999 F. $50,000 $59,999 J. $125,000- $149,999
C. $20,000- $29,999 G. $60,000 $74,999 K. Over $150,000
D. $30,000 $39,999 H. $75000 $99,999
D. Pre-test and Approval of Survey Instrument
According to the eighth NOAA panel recommendation, the survey instrument
must be clear and acceptable to many respondents. To help achieve clarity, a preliminary
draft of the survey instrument was presented to ten selected people for their comments
and feedback. This pre-test (NOAA recommendation #7) was used to refine the
instrument's clarity and to set bounds on the randomly assigned prices.
The draft instrument was also reviewed by the Institutional Review Board (IRB)
of Florida A & M University; a panel that reviews research projects involving human
subjects. The IRB provided feedback on the instrument's comprehensibility (NOAA
recommendation #8) and checked for potentially offending contents. After comments
and revisions, the final survey was approved by the IRB.
E. Preparation of Final Survey Instrument
Presenting an attractive and well organized survey instrument that is appealing to
the respondents is important to the response rate (Dillman, 2000). Helping to achieve this
effect, a professional graphic artist was employed to improve the quality of the
presentation. This included the addition of illustrations, photographs and other graphic
enhancements. The survey instrument was then printed and ready for distribution.
12. Please indicate your approximate household income before taxes in 2001. (Circle one
letter).
A. Under $10,000 E. $40,000 $49,999 I. $100,000- $124,000
B. $10,000- $19,999 F. $50,000 $59,999 J. $125,000- $149,999
C. $20,000- $29,999 G. $60,000 $74,999 K. Over $150,000
D. $30,000 $39,999 H. $75000 $99,999
D. Pre-test and Approval of Survey Instrument
According to the eighth NOAA panel recommendation, the survey instrument
must be clear and acceptable to many respondents. To help achieve clarity, a preliminary
draft of the survey instrument was presented to ten selected people for their comments
and feedback. This pre-test (NOAA recommendation #7) was used to refine the
instrument's clarity and to set bounds on the randomly assigned prices.
The draft instrument was also reviewed by the Institutional Review Board (IRB)
of Florida A & M University; a panel that reviews research projects involving human
subjects. The IRB provided feedback on the instrument's comprehensibility (NOAA
recommendation #8) and checked for potentially offending contents. After comments
and revisions, the final survey was approved by the IRB.
E. Preparation of Final Survey Instrument
Presenting an attractive and well organized survey instrument that is appealing to
the respondents is important to the response rate (Dillman, 2000). Helping to achieve this
effect, a professional graphic artist was employed to improve the quality of the
presentation. This included the addition of illustrations, photographs and other graphic
enhancements. The survey instrument was then printed and ready for distribution.
F. Sample Selection and Data Collection Procedures
1. Study Site Selection
Franklin and Gulf counties, Florida, were selected for study because of their
proximity to the Apalachicola National Estuarine Research Reserve (ANERR). As part of
a cooperative research effort to better document the importance of the ANERR's
environmental service to humans, the Environmental Cooperative Science Center
supported the survey effort in this area. See figure 1 for map of study site.
Figure 1: Map of study site, Franklin and Gulf counties, Florida.
Study Site Locations
OGulf County
Franklin County
N
WE
S
2. Data Collection Procedure
After selecting the study site, the sample was collected from a list of permitted
well owners maintained by the Northwest Florida Water Management District. The data
were edited for the necessary variables, such as name and address, and then merged
(Franklin and Gulf counties) resulting in approximately 1,300 names and addresses. The
data were further edited to remove duplicate names and addresses and non-specific
addresses (e.g. community well owners) and 1,000 were randomly selected to receive a
survey instrument.
A second survey used the file of licensed drivers for Franklin County, maintained
by Florida Department of Motor Vehicle (FDMV) 2. Of the 7,000 names and addresses of
licensed drivers in Franklin County, Florida, 1,000 were selected and mailed survey
instruments. Sample size was determined by the desire to achieve a 95 percent level of
confidence in a five percent level of precision. Using previous estimates of sample
variances it was estimated that a sample of approximately 200 would be needed.
Assuming a response rate of 20 percent, 1,000 residents were mailed surveys.
G. Implementation of the CV Survey Instrument
Before the surveys were sent to respondents, care was taken in thoroughly
preparing the survey with its associated documents to assure a good response. This was
accomplished by following the formal guidelines of the tailored design method (TDM),
2 Over 80 percent of questionnaires from first survey were returned as undeliverable due to an erroneous
and outdated address file. Many of the bad addresses resulted from the reassignment of addresses in Gulf
County. The high undeliverable rate prompted the need for new a effort with a more reliable data set.
(Dillman, 2000). Table 1 compares the procedures (pre-mail and post-mail) suggested by
the TDM and those followed by this study.
Table 1: Comparison of procedures followed in implementation of the survey instrument
Procedure Dillman's Survey 1 Survey 2
TDM
Pre-mailing Procedures
Appropriate cover of the
questionnaire; cover letter with
appropriate signature
Identification number on
questionnaire
Prepared return envelope
First class mail; "address
correction requested"
Post-mailing Procedures
1i follow-up (post card;
reminder/thank you)
2nd follow-up (2nd
survey/replacement survey).
Mailing address
labels affixed to
survey
Yes
Yes
Mailing address on cover
letter and key numbered
to identify participants for
thank you reminder card
Yes
Yes
Yes
Yes for request of
replacement surveys
Follow-up by certified mail Yes No No
(reminder letter and
questionnaire)
Substitution of telephone calls Yes No No
for one or more follow-ups
The survey packet sent to each respondent included items such as a detailed cover
letter, the survey instrument and a pre-stamped, pre-addressed return envelope. Studies
have shown minimal effects from the cover letter on the response rate but have indicated
a slight improvement with colored covers (Dillman and Dillman, 1995). Simple
questionnaire covers (front and back) with minimum graphics and wordings are
preferred. For example, detailed pictures, especially those selected from clip-art files
should be avoided, since they often detract or mislead the participants (Dillman, 2000).
The front cover should include a title, and the back cover end with a "thank you" and
plenty of white space with an invitation to make additional comments (Dillman, 2000).
A well-designed cover letter is key to stimulating response (Dillman, 2000). The
ideal cover letter should be limited to one page and should convey essential and
straightforward information to motivate the participant and should include the following
components (Dillman, 2000):
(1) Date of the letter
The date at the top of the letter is the first element of personalization. The absence
of or incomplete date (e.g. September 2003) on the cover letter may indicate a lack of
importance and could reduce response.
(2) Inside name and address
The use of inside name and address in the cover letter is preferred because it
expresses some measure of importance and urges response. However, where this is not
possible, an alternative personalized format such as "To resident at this address," can be
used.
(3) The salutation
The appropriateness of the salutation varies significantly but when there is no
preexisting relationship between the sender and receiver, and the gender is known,
formally addressing the last name of the participant is appropriate, or simply "Dear
participant, if otherwise.
(4) Purpose of the letter
The cover letter should explain the purpose of the letter without influencing the
response. For example, detail about the expected result should be avoided.
(5) The usefulness and importance of the request
Explaining why the action is requested of the respondent is useful, important and
an essential part of the cover letter. However, care must be taken in its formulation to
avoid bias, or mislead the participant. For example, ideas that give the impression that a
certain response is being sought should be avoided.
(6) The terms of participation
The cover letter should inform the subject when a survey is voluntary and provide
them an opportunity to opt-out. This information can be expressed in the same paragraph
as encouragement to response.
(7) Indication of enclosure
Often elements of the mail-out package are easily seen by the respondent, so a
detailed explanation is not required to inform about the enclosure. However, this can be
mentioned casually after the voluntary participation statement.
(8) Postscript addition
The postscript is a very important part of the cover letter that is often presented
prior to the remainder of the message. The postscript is an appropriate place to express
"thanks again" and the importance of the subject's participation.
(9) Contact information
The cover letter should provide the respondents with contact information and an
opportunity for questions. This helps convey the idea of accessibility, and trust that the
survey is legitimate and important.
(10) A real (appropriate) signature
Cover letters should be individually signed if possible. However, if this is not
possible, a signature stamp or machine may be used. The survey cover letter was
prepared, approved and signed bythe appropriate personnel at Florida A & M University
(FAMU). However, because of the large sample, a signature stamp was used to apply the
signature individually to each cover letter.
The 1,000 randomly selected names were then processed to produce individual
labels and letters, and the labels were affixed individually on the questionnaire. Return
labels were also prepared and placed on the return envelopes along with return postage.
Additionally, the cover letters were folded so that the letterhead was readily viewed when
opened.
For each packet, the survey instrument was located in the front with the
respondent's name in clear view, followed by the cover letter and then the return
envelope. All survey packets were sealed, boxed and mailed first class from the United
State Postal Service substation located on FAMU campus. Reminder/thank you cards
were sent to all participants one week after the first survey mailing. Follow-ups and other
procedures were omitted due to time and budget constraints.
H. Data Analysis
The data analysis will address the selected objectives of determining the mean and
median WTP for reduced levels of groundwater nitrates, testing for significant
differences in the mean WTP between public and private institution payment vehicles,
and evaluating the importance of institutional basis in CV methodology.
The mean estimate of WTP for nitrate-reduced groundwater is estimated as;
n
i=1
where,
xi = the ith WTP observation
n = # of observations
The median WTP is the middle point of rank ordered observations (i.e. number of
respondents. Both statistics provide a measure of central tendency, with the mean being
the unbiased and expected value of WTP.
The difference between private and public payment vehicles will be tested by
the following:
Ho: XPu = XSF
Ha: XPU XSF
and test statistic
(XPU-XSF)-(XPU-XSF)
t --1--- (2)
S2( -+ )
nPU nsp
where:
S2 = the pooled variance
PU = Private utility (private) payment vehicle
SF = State of Florida (public) payment vehicle
XPu & XSF = Population means for PU and SF respectively
xPu & XSF = sample means (estimation ofX ) for PU and SF respectively
n = Number of observations
To evaluate the participation choices by survey respondents, a discrete choice
model will be developed and evaluated with a probit analysis. Typically, when evaluating
behavior that involves discrete choices (yes/no), statistical models such as logit and
probit are frequently used (Bishop, Heberlein and Kealy, 1983; Green, 1993).
The choice between the logit and probit approaches is a matter of researcher's
preference and the type of data encountered. Some researchers prefer the probit model
because the logit distribution assumes more data variability, which may result in an
implausibly high estimate of the mean WTP. However, there is no inherent justification
or theoretical ground for choosing either probit or logit (Amemiya, 1981; Green, 1993;
Kennedy, 1998).
The probit model is based on the cumulative normal probability distribution
where the coefficients of the explanatory variables are measures of their influence in
terms of the probability of making a choice to participate. The probit analysis solves the
problem of how to obtain estimates of parameters as well as obtaining information about
the dependent variable. Using a linear probability model, such as ordinary least squares,
in a discrete choice data analysis leads to misspecification, that is, incorrectly assuming
the distribution is constant (Pindyck and Rubinfeld, 1991; Kennedy, 1998).
Estimation of binary choice models (probit or logit) is usually based on the
method of maximum likelihood, where each observation is treated as a single draw from
a Bernoulli distribution (binomial with one draw). Limited dependent variable software,
in particularly Limdep, is often used to estimate discrete type data in discrete choice
models such as probit and logit (Green, 1993).
The probit model from Green (1993) and Kennedy (1998) is adapted in part for
this study.
The probit model,
Prob (y=l) = Q(p'3X) (3)
is a normal probability distribution of a regression,
E(y) = 0[1 F(PX)] + 1[F(PX)] (4)
=F(P'X) (5)
where,
S= vector of estimated parameters
( = normal probability distribution, with mean 0 and variance oa
X = matrix of explanatory variable (e.g. income, age, gender, education)
y = decision variable (0 or 1)
With the DC data, the probit model is useful in establishing a relationship
between the variable of interest (e.g. WTP= y) and a set of covariates (e.g. X including
income, education).
The probit model is often used to predict whether or not people will make particular
choices, such as purchasing certain commodities, vote or seek work, with respect to the
explanatory variable (Green, 1993; Kennedy, 1998). In this study, the dependent variable
equals one when one chooses to purchases the program and zero otherwise.
Equation (4) specifies the relationship between dependent variable and the
explanatory variables and may be used for determining their relative importance and for
making predictions, that is, how well the model predicts choices (Green, 1993; Truett and
Truett, 1998). The estimated coefficients measure the proportion of total variation in the
dependent variable (y) that is explained by the variation of the independent variables (X)
(Green, 1993; Truett and Truett, 1998). The coefficient of determination (COD) is a
measure of the fit of the regression model, which may be calculated by observing
variations in the independent and dependent variables. For example, the closer the
observed data points lie to the regression line (i.e. predicted to actual), the more accurate
the regression model or better the fit (Green, 1993; Truett and Truett, 1998). There is also
interest is the size and sign of the estimated coefficients and whether they are consistent
with economic theory (Mitchell and Carson, 1989). These are indicators of the level and
type (positive or negative) of relationship existing between the independent and
dependent variable and may be used to evaluate the model based on a priori theory
(Truett and Truett, 1998; Kennedy, 1998).
Lastly, an appreciation for a model and any findings needs validating before
recommendation to a policymaker. As a probability model, the probit model allows the
comparison of actual to predicted outcomes and provides a measure of model validation
(Green, 1993; Pindyck and Rubinfeld, 1991; Anderson, Sweeney and Williams, 1996).
The systematic relationship between the respondents' choice behavior, the program's
offer prices and the respondents' socio-economic backgrounds and attitudes is another
form of validation (construct validity). Also, as suggested by the NOAA panel report, a
convergent validity test will compare the demographics of the sample frame to those of
the sample to help determine if these data are representative of the population.
Conducting the following series of chi-square tests of independence will perform
the construct validity test for the probability P of choosing the program:
Ho: Pil= Pi2
Ha: Pil # Pi2
for i = 1, ..., 5 dichotomous demographic characteristics including income
(555,000, >55,000), employment (employed, unemployed), education (college degree, no
college degree), offer price (<$5, >$5) and payment vehicle (state, private).
with test statistic
x = yyN
E, (6)
where,
Nij = # of observations for demographic characteristic i, group j.
39
Ei = expected frequency for demographic characteristic i, group j.
The Pearson chi-square may be used to determine if a person's choice of a
commodity is related to their socio-economic background such as income, sex, and
education, or based on the price offered for the commodity. The chi-square test allows
one to make associations between the row variables, i and the column variables, j by
comparing their frequencies. The frequency table will be used to compare actual
frequencies with computed expected values (Eij = nij/N, where n is frequency of variable)
prior to computing chi-square, (Everitt, 1977).
CHAPTER 4
RESULTS AND DISCUSSION
Part 1. Survey Response Results
Tables 2 & 3 list the results of the two surveys that were pooled for a total of
131 useable responses and a weighted response rate of 15 percent.
Table 2: Survey response rate
Category
Total mailed
Undeliverable questionnaires'
Adjusted sample size 2
Returned questionnaires
Response rate (%)
Returned questionnaires not useable
Adjusted returns3
Adjusted response rate (%)
Survey 1
1,000
870
130
30
23
1
29
22
Survey 2
1,000
210
790
108
14
6
102
13
1 Addressees moved left no forwarding address; forwarding address expired; not deliverable as
addressed; unable to forward; insufficient address; attempt unknown; no mailing receptacle; box
closed; deceased.
2 Total mailed questionnaires minus undelivered questionnaires.
3 Returned questionnaires minus non-useable returned questionnaire.
Table 3: Pooled survey response rate
Category Response
Adjusted sample size from survey one 130
Total mailed in survey two 1,000
Pooled amount (Survey one & two) 1,130
Undeliverable questionnaires 210
Pooled adjusted sample size4 920
Pooled returned questionnaires 5 138
Pooled response rate (%) 15
Pooled returned questionnaires not useable 7
Pooled adjusted returns6 131
Pooled adjusted response rate (%) 14.3
Budget and time prevented follow-up mailings to non- respondents. While
the response rate was a relatively low 15 percent, when compared to other CV
survey efforts, the results were not unusual. The literature shows response rates for
CV mail survey as low as 8 percent (Schulze et al., 1983) with others in the 20
percent to 30 percent range (Loehman and De, 1982, Brookshire, Eubanks and
Randall, 1983).
Question one of the survey presented the respondents with a broad list of
issues as potential substitutes to the proposed water program and help create an
4 Survey one adjusted sample size plus survey two adjusted sample size.
5 Returned questionnaires from survey one plus returned questionnaires from survey two.
6 Pooled returned questionnaires minus pooled returned questionnaires not useable.
42
atmosphere for constrained decision making. It also helped provide potential
covariates for analysis. The respondents were asked whether they would prefer
government to spend more, the same, or retain current spending on the list of policy
issues. While 50 percent of the respondents were in favor of spending more money
on improving facilities in education, health care, surface water and groundwater
protection; more than 40 percent wanted more spending on improving programs to
counter (reduce) terrorist actions. Table 4 lists the outcome by issue.
43
Table 4. Question 1: Preferences of the respondents concerning public spending on public
policy Issues
Spend more Retain current Spend
Issues money spending less money
------------------------------- % --------------------------- --
Reduce global warming 23.0 42.7 23.7
Protect endangered species 27.5 46.6 18.3
Provide foreign aid to reduce poverty in the 14.5 38.9 39.7
world
Reduce air pollution from automobiles 32.1 47.3 13.0
Improve facilities in education 51.9 29.0 10.7
Improve health care facilities 55.7 32.1 5.3
Improve farming with less reliance on chemicals 36.6 42.0 14.5
and pesticides
Improve roads, bridges, airports, etc. 25.2 57.3 .9.9
Reduce crime to protect public safety 35.9 47.3 9.2
Improve programs to reduce terrorist actions 43.5 38.9 9.2
Improve roads in rural areas 32.8 48.1 11.5
Improve local law enforcement 27.5 51.9 11.5
Improve local fire protection 26.7 61.1 4.6
Improve school bus safety 22.9 60.3 9.2
Preserve wetlands (e.g. marshes & swamps) 35.9 44.3 13.0
Protect surface water quality (e.g. lakes & rivers) 53.4 35.1 4.6
Protect groundwater quality (e.g. well water) 57.3 31.3 4.6
A respondent's views, opinions and experiences may help explain their
WTP for a commodity (Mitchell and Carson, 1989). To gain an overall
understanding of respondent viewpoints, the different issues were consolidated into
eight categories; environment, education, national defense/security, infrastructure,
moral, pro-government/budget, anti-government/budget and health and listed in
Table 5. For each respondent, an index ranging from 0 to 1 was assigned for each
category based upon their mix of selected issues. For example, a person selecting
many environmental issues as important would score close to one in the
environmental category.
Overall almost all of the respondents showed some concern for the
environment. Fifty-two point seven percent and 56.5 percent of the respondents are
strongly in favor of improving education and health care facilities respectively. The
category of national defense/security also has high priority among the respondents.
The results showed that not many of the respondents are concerned about moral
issues or general government/budget policies, although some appear to have very
strong views against some aspects of the government.
Table 5. Question 1: Attitudinal measures of the respondents
Category Index range Frequency (# % Frequency
respondent)
Environment 0-0.50 45 34.4
0.51-1.00 86 65.6
National 0-0.50 56 42.8
defense/security 0.51-1.00 75 57.2
Infrastructure 0-0.50 80 61.1
0.51-1.00 31 38.9
Moral 0-0.50 110 83.9
0.51-1.00 21 16.1
Education 0-0.50 62 47.3
0.51-1.00 69 52.7
Pro-government/ 0-0.50 128 97.7
budget 0.51-1.00 3 2.3
Anti-government 0-0.50 116 88.5
/budget 0.51-1.00 15 11.5
Health care 0-0.50 57 43.5
0.51-1.00 74 56.5
Section two of the survey focused on the proposed scenario and payment
mechanism. To examine the influence of the institution basis of the payment
vehicle, the study included two institutional-based payment vehicles, State of
Florida (SF) and private utility (PU). Table 6 reveals that 53.4 percent of the
respondents received a SF payment vehicle, with the balance receiving a PU
payment vehicle.
Table 6. Section 2: Responses by institutional-based payment vehicle for proposed program
Payment vehicle Frequency % Frequency
PU 61 46.6
SF 70 53.4
Total observed 131 100.0
To avoid starting point bias, four monthly offer prices were randomly
assigned to the proposed program. The responses by offer prices were almost
evenly distributed, with responses from $1 and $5 tied for 26 percent, see Table 7.
Table 7. Section 2: Responses to survey by offer prices for proposed program
Bid$ Frequency % Frequency
1 34 26.0
5 34 26.0
10 32 24.4
20 31 23.7
Total Observed 131 100.0
Question two, section two asked the respondents if they would accept the
program for the stated price. Forty one percent answer "yes" while about 59 percent
respond "no" to the program, see Table 8.
Table 8. Section 2: Question 2; yes/no responses to WTP
Response Frequency % Frequency
Yes 53 40.5
No 78 59.5
Total observed 131 100.0
In an attempt to further refine a respondent's WTP for the program, the next
several questions establish upper or lower bounds on their initial answer to question
two, section two. These consist of two open-ended questions (3a and 3b)
establishing upper or lower bounds. Respondents who answered "yes" to the initial
price were asked in question 3a to list the highest price they would be willing to pay
to participate in the program (upper bound). Respondents who answered "no" to the
initial price were asked in question 3b to list the highest price they would be willing
to pay (lower bound). The mean upper and lower bound amounts per household are
$4.24 and $1.14 per month respectively.
The double-bounded approach is an attempt to bind the respondent's WTP
by establishing an upper or lower limit WTP from their initial acceptance or refusal
of the offer price. However, if the respondent answered absolutely "yes" or "no" to
the stated amount, his WTP is the offered price and $0 respectively; or if he
answered "no" to the offered price but was willing to pay a lower amount, or
answered "yes" and was willing to pay more, that second amount become the
respondent's WTP. The respondents' WTP were then rank ordered to determine the
median and distribution. WTP ranges from $0 $40 per month per household.
About 43 percent are willing to pay $0 that is, responded "no" to the program.
Twenty-two percent of the respondents would be willing to pay $5 to participate in
the program, see Table 9. On the average, the respondents are willing to pay $4.71
per month with a median of $1.00 and standard deviation of $7.13 per month
respectively.
Table 9. Question 3: Frequency distribution for observed WTP
WTP$ Frequency % Frequency
0 56 42.7
1 10 7.6
2 6 4.6
3 5 3.8
4 1 0.8
5 22 16.8
7 3 2.3
10 13 9.9
15 7 5.3
20 3 2.3
25 2 1.5
30 2 1.5
40 1 0.8
Total Observed 131 100.0
For respondents who reported they were willing to pay zero dollars for
nitrate reduction, question four produced an opportunity to explain their answer.
The respondents were provided four options from which to choose and an open-
ended question for other reasons. A total of 84 (64.2%) protests are made against
the program for various reasons. Many of the respondents either believe there is
either no problem with their water (13%) or list another reason (35.9%). Only 13
percent of the respondents opt out of the program because they do not believe the
program is worth the cost (7.6%) or is unworkable (4.6%). Table 10 summarizes the
reasons for $0 willingness to pay.
Table 10. Question 4: Was the most you are willing to pay $0 because:
Protest Frequency % Frequency
Program not worth the cost 10 7.6
Do not believe program will work 6 4.6
Not comfortable with paying to provide clean water to 4 3.1
protect human health
Do not believe there is problem with their water 17 13.0
Other reasons' 47 35.9
Total protests 84 64.2
'See appendix B for other specific reasons
Question five measured the attitudes and knowledge of the respondents by
asking them if they had their well tested and to provide the test result if they
answered yes. Only nine percent of the respondents tested their well for nitrates and
three percent did not respond to the question, see Table 11. Since only three
respondents provided actual test results, the variable was excluded from further
analysis. Many respondents are concerned about testing of their well water for
contaminants but feel unable because of their renter status.
Table 11. Question 5: Have you had your well tested for nitrates?
Response Frequency % Frequency
Yes 12 9.2
No 115 87.8
No response 4 3.0
Total Observed 131 100.0
Question six gathered information on the respondents' attitudes, perceptions
and knowledge of the environment. The respondents were asked about the quality
of their home water supply. About 40 percent of the respondents believe that their
water is either very safe (12.2%) or safe (27.5%) while about 60 percent have some
doubts.
Table 12. Question 6: In your opinion, the quality of water you use at home is:
Category Frequency % Frequency
Very safe 16 12.2
Safe 36 27.5
Seems to be safe 33 25.2
Unsafe 25 19.1
Don't know 14 10.7
No response 7 5.3
Total Observed 131 100.0
Question seven measured the respondent's choice of drinking water source.
About 83 percent of all respondents either use a home filter (35.9%), purchase
bottled water (33.6%) or other substitute beverages (9.2%). Only about 21 percent
of the respondents use well water with no additional treatment or did not respond to
the question, see Table 13. This result is consistent with the results of question
twelve since 60 percent express doubt about the quality of their water supply.
Table 13. Question 7: For drinking, which do you use?
Category Frequency % Frequency
Well water no additional treatment 19 14
Well water treated 47 35.9
Purchased bottled water 44 33.6
Other substitute beverages 12 9.2
No response 9 6.9
Total Observed 131 100.0
Question eight documented a participant's length of residency, and was
included as a possible covariate in the analysis of WTP. Respondents were asked to
indicate the number of years they had lived at their current location. The number of
years range from 1 to 65, with a mean of 11.93 years, a median of eight years and
standard deviation of 11.98 years, see Table 14.
Table 14. Question 8: Number of years lived at current location
Year Frequency % Frequency
0-10 70 53.6
10-20 27 20.7
20-30 15 11.5
30-40 8 6.2
40-50 3 2.4
50-60 1 0.8
Over 60 1 0.8
No response 6 4.8
Total 131 100.0
Question nine dealt with demographic information. Respondents were asked
to list their household size including the gender and age of each household member.
Most persons in the survey fall within the 46 to over 65 age categories. Only two
persons (females) in the survey are three years or younger in age, see Table 15.
Previous studies have showed that age is an important determinant in the effect of
nitrate contamination (Pierzynski, Sims and Vance, 2000; USEPA, 1998; Water
Quality Initiative 101, 1995). While high nitrate levels may affect other age groups,
children under one year are most affected, which may influence a person's concern
about nitrate reduction.
A total of 142 males and 141 female were accounted for in the survey. Most
households consist of single occupants; 76 percent of males and 67 percent of
females live in single occupant households. The highest number of males and
females per household is six and four respectively, see Tables 16 and 17. The
average number of persons per household is 2.16, which is comparable to that
reported by the U.S. Census Bureau (2000) for Franklin county (2.28) and Gulf
county (2.46).
Table 15. Question 9: Number of persons in the household by age and sex
category
Age category MALE FEMALE Total Total
percent
< 6 month 0 0 0 0
6mth 3yr. 0 2 2 0.71
4 12 yr. 9 4 13 4.59
13 25 yr. 10 19 29 10.25
26-45 yr. 33 30 63 22.26
46-65 yr. 48 58 106 37.46
S65 42 28 70 24.73
Total 142 141 283 100.0
Table 16. Question 9: Number of males
Male per Frequency % Frequency Males in
household Household
1 100 76.3 100
2 10 7.6 20
3 4 3.1 12
4 1 0.8 4
6 1 0.8 6
Missing 15 11.5
Total 131 100.0 142
Table 17. Question 9: Number of females
Female per Frequency % Frequency Female in
Household Household
1 88 67.2 88
2 20 15.3 40
3 3 2.3 9
4 1 0.8 4
Missing 19 14.5
Total 131 100.0 141
In question ten respondents were asked to list their highest level of formal
education. According to the results, 22.9 percent have a formal college education
and another 23.7 percent have a technical or professional degree. This is not
substantially higher than the 12.4 percent college graduates reported for the area in
the U. S. Census Bureau in 2000.
Table 18. Question 10: Highest level of formal education completed
Education level Frequency % Frequency
Grade school or less 0 0
Some high school 8 6.1
High school graduate 16 12.2
Some college 42 32.1
College graduate 30 22.9
College/ professional 31 23.7
degree
Other 2 1.5
No response 2 1.5
Total 131 100.0
In question eleven (section three) respondents were asked about their
present employment status. Approximately 50 percent of the respondents indicate
they are employed and six percent are unemployed, see Table 19 for detail.
Table 19. Question 11: Respondents' employment status
Employment Frequency %
Frequency
Unemployed 8 6.1
Employed (part-time) 9 6.9
Employed (full-time) 56 42.7
Employed (seasonally) 1 0.8
Retired 47 35.9
Student 4 3.1
Homemaker 1 0.8
No response 5 3.7
Total 131 100.0
Finally, question twelve asked respondents to reveal their before-tax
household income. Figures range from $5,000 to $200, 000 per year, see Table 20,
with median and mean $55,000 per year and $61,296 per year respectively. The
standard deviation is $4,189.54 per year. This result is not comparable to the
median income ($26,756) reported by U. S. Census Bureau (2000) for Franklin
county.
Table 20. Question 12: Respondents' gross household income
Income' $ Frequency %
Frequency
5,000 8 6.1
15,000 11 8.4
25,000 19 14.5
35,000 12 9.2
45,000 8 6.1
55,000 10 7.6
65,000 17 13.0
85,000 8 6.1
110,000 3 2.3
135,000 6 4.6
200,000 9 6.9
No response 20 15.3
Total 131 100.0
'Median of income ranges shown in survey instrument.
Part 2: Results Data Analysis
Table 21 lists and describes the variables considered in this research. Only a
subset of these was actually used in the hypothesis test and probit model.
Table 21: Description of variables used in this research
Description
Variable
PU
SF
QIENV
Q1DEF
Q1INF
Q1EDU
Q1PGOV
Q1AGOV
Q1HEA
Q1MOR
MALE
FEMALE
YOUTH
OLDER
OLD
YOUNGER
INCOME
COLLEGE
COLLEGE
PRICE
EMPLOY
NEMPLOY
YEARS
RESPYES
RESPNO
UPPERA
LOWER
BID$
TESTYES
TESTNO
MF
PROTEST
Q6
Q7
Organizational basis, private utility payment vehicle
Organizational basis, State of Florida payment vehicle
Respondent's attitude toward environmental issues
Respondent's attitude toward national defense/security
Respondent's attitude toward infrastructure
Respondent's attitude toward education
Respondent's attitude for government/budget
Respondent's attitude against government/budget
Respondent's attitude toward health care
Respondent's attitude toward moral issues
# of males in households
# of females in households
# of males and females ages > month to 25 years
# of males and females ages 26 to 65 years.
# of males and females ages 65 years and over.
# of males and females < 12
Respondent's before taxes household income
Level of formal education of respondent such as college or professional degree
No college education. High school graduates or less.
WTP values from respondent.
If respondent employed part-time, full-time or seasonally.
If respondent unemployed, retired, student or homemaker.
# of years lived at current location.
If respondent responded yes to valuation question.
If respondent responded no to valuation question.
The highest WTP amount by respondent who responded yes to original bid.
The highest WTP amount by respondent who responded no to original bid.
The bid amount proposed the offered price ($1, $5, $10, $20).
If respondent responded yes to testing nitrate in well water.
If respondent responded no to testing nitrate in well water.
Total household component (Male + female).
If respondent objects to valuation question, reasons) for response:
1= program is not worth the cost; 2 = do not believe the program will work; 3 =
not comfortable with idea of paying in a special fund to provide clean water for
protection of human health; 4 = do not believe there is a problem with the water
supply (well); 5 = other important issues.
Respondent's opinion of the quality of water used in his/her home: question 6
1 = very safe; 2= safe; 3= seems safe; 4= unsafe; 5= don't know
Type of water/beverage respondent used for drinking.
1= well water without additional treatment; 2=well water filtered or treated in
home; 3= purchased (bottled) water; 4= other substitute beverages.
The mean WTP estimates differ significantly between payment vehicles at
the 0.15 level, see Table 22. While this is less than the 0.05 or 0.10 levels often
reported in literature, it is still noteworthy.
Table 22: Mean WTP by institutional basis
Payment vehicle Mean WTP Standard error
State of Florida 4.3857 0.772
Private utility 5.0820 1.0077
Note: the WTP values are significantly different at the 0.156 level: t = 1.489, df. = 129
The probit model helps document the relationship between a respondent's
WTP and several independent variables, and is useful in validating the model's
effectiveness by comparing its predictions to the actual observations. Six variables
are found to be significant at the 0.10 level or better. These included MALE, SF,
Q7, PROTEST, UPPERA, and PRICE, see Table 23. As one might expect, a
person's likelihood to accept the offered price and participate in the program is
negatively affected by the offered price itself (own price) or her protest against the
program, and positively affected by her water drinking patterns. Respondents with
more education, longer years of residency, willingness to support more government
spending on health care and infrastructure are more likely to accept the offered
price. Similarly, the respondents who do not exhibit the above attitudes and
demographic status are more likely to reject the offered price. Respondents who
believe the quality of water used in the home is safe are less likely to participate in
the program. However, respondents with lower income are more willing to pay for
the program, which implies that well water is an inferior good.
Table 23. Summary of probit model results
Variable Parameter p-value Variable
estimate mean
CONSTANT -3.035 0.232
EMPLOY -0.013 0.988 1.06
MALE -0.583 0.104* 1.08
OLDER 0.159 0.677 1.82
INCOME -0.0001 0.866 61,300
COLLEGE 0.327 0.626 0.793
Q1HEA 0.521 0.560 0.767
Q1ENV -0.840 0.468 0.632
Q1EDU -0.016 0.985 0.721
SF 1.041 0.037** 0.534
Q6 -0.150 0.529 2.72
Q7 0.575 0.066* 2.23
PROTEST -0.464 0.000*** 2.28
Q1INF 1.561 0.217 0.595
"Q1DEF -0.280 0.859 0.616
UPPERA 0.682 0.000*** 4.23
PRICE -0.251 0.028** 4.71
YEARS 0.031 0.109 11.3
NB: *, **, ***, indicate levels of significance at .10, .05 and .01 respectively.
See Table 21 for description of variables.
The model fits the data well, (chi-square test significant at the 0.10 level or
better) and is consistent in its ability to correctly predict a respondent's choice.
Table 24 compares the predicted and actual response to the choice of participating
in the program. The model correctly predicts 120 of 131, or 91.6 percent of the
observations.
Table 24: Actual vs. predicted yes-no responses
Predicted Total
0 1
0 73 5 78
Actual
1 6 47 53
Total 79 52 131
Note: 0= decision no, l=decision yes.
The construct test is performed to determine if an association exists between
the choice to participate in the program and the characteristics of income,
employment, payment vehicle, education, and offer price. Table 25 lists the cross
tabulation between participation and the independent variables that represent these
characteristics. Three characteristics are found to be associated with participation at
the 0.10 level or better; education, offer price and payment vehicle. There appears
to be little association between participation and either income or employment
status.
Table 25: Summary of cross tabulation
Variable Calculated X2
INCOME 1.16
(<$55,000, >$55,000)
EMPLOYMENT 0.51
(employed, not employed)
EDUCATION 12.10***
(college, no college)
OFFER PRICE 13.97***
(accept, reject)
PAYMENT VEHICLE 4.12**
(PU, SF)
Note: **, ***, indicate rejection of the null hypothesis at the .05, .005 levels respectively.
The convergent validity test measures the extent to which two measures of
the same theoretical construct correlate. For example, two different samples of the
same study site are compared to determine if they correlate at different points in
time (Mitchell and Carson, 1989; Carson, Flores and Meade, 2001). This test is
often conducted to determine statistical reliability by comparing demographics
between the sample and study site (Carson et al., 2001. For example, if a sample is
characterized by demographics different from the study site, it would imply the
study is not representative of the population. The variables selected are based on
commonly reported variables in the U.S. census. Table 26 shows a comparison
between data in this research and that of the 2000 U.S. Census Bureau, Franklin
County, Florida. The confidence interval and standard errors are used to determine
where significant differences arise.
Overall, none of the compared U. S. Census variables falls within their
respective confidence intervals, implying that there are significant demographic
differences between study sample and the population.
Table 26: Study data vs. U.S. Census Bureau
Standard U.S. Confidence
Category/household Study error Census' Interval (CI)
Bureau
Female person (%) 49.8 1.073 43.3** 50.87,48.73
Person 2 65 years old (%) 24.4 0.534 15.7** 24.93, 23.87
Median household income ($) 55,000 366.03 26,756** 55,366.03, 54,633.03
Persons per household 2.16 0.097 2.28** 2.26, 2.06
'2000 U.S. Census Bureau Report
NB: CI was tested @ .05 confidence level
** indicates that the Census data were significantly different from study estimates @ the .05 level.
CHAPTER 5
SUMMARY AND CONCLUSION
To address the objectives of this study, a double-bounded dichotomous
choice CV survey was developed and mailed to 2,000 residents of Franklin and
Gulf counties in Florida to estimate WTP, explain consumer choice behavior and
determine if the institutional basis of the payment vehicle affects WTP. Based on
131 respondents, the mean and median WTP are estimated to be $4.71 and $1.00
per month respectively. Furthermore, there is a significant difference at the 15 %
level in WTP between the institutional basis of the payment vehicle, state vs.
private.
This degree of difference between the institutional-based payment vehicles
is important to know for CV practitioners and supports the practice of including
more than one types of payment vehicle in a survey. This practice should be benign.
If a difference fails to materialize the data may simply be pooled, but if a significant
difference materializes the influence will be documented and permit a more
accurate WTP estimate.
A probit analysis was conducted to model program participation. The
significant explanatory variables include MALE, SF, Q7, PROTEST and the
PRICE and are all significant at the 0.10 level or better. The offered price and
protests against the program negatively affect WTP for the proposed program while
water consumption habits have a positive influence. On the other hand, income is
negative and insignificant, implying that well water is an inferior good. The probit
analysis correctly predicts 91.6 percent of the observations implying the model is
efficient in forecasting behavior.
The study also modeled a person's attitude and socio-economic background
in relationship to his WTP response. For example, does the respondent's attitude
and socio-economic characteristics affect his willingness to participate in the
program? The results show that the respondent's level of education, the price
offered to participate in the program and the institutional basis of the hypothetical
payment all affects choice behavior.
While the estimated mean WTP in this study differs significantly from
several studies such as deZoysa (1995), Poe and Bishop (1992), Crutchfield and
Cooper (1997), Whitehead, Hoban and Clifford (2000), and Whithehead (2003)
with WTP of $9.69, $24, $54.50, $77.23 and $75.95 per month respectively, it
compares favorably with Clemons, Collins, and Green (1995), with estimates of
$1.80 and $1.20 per month. While these studies valued water quality, differences in
the population size and demographic composition, payment vehicle, offer amount,
and number of resources valued likely explain the wide variations in the WTP
estimates across studies.
One explanation for higher mean WTP estimates in comparable studies is
the size of initial offer price. For example, Crutchfield and Cooper (1997) had a
relatively high range of offer prices ($4.50 $100 per month) and documented a
high WTP estimate. They estimated WTP for drinking water in four regions and
concluded that it varies across regions through differences in demographics and
other socio-economic factors. Whitehead, Hoban, and Clifford (2000) and
Whitehead (2003) who studied different populations and water resources also
documented relatively high mean WTP estimates associated with relatively high
offered amounts ($10 $200 per month). They concluded that the WTP response
depends on the respondent's perceptions and socio-economic background. On the
other hand, the studies that exhibited relatively low WTP estimates (e.g. deZoysa,
1995; Clemons, Collins, and Green 1995) also had low offer prices ranging from $0
to $17 per month and $10.50 to $16.50 per month respectively.
While deZoysa's (1995) range of offer prices were comparable to this study,
she researched three resources and seven programs in twelve counties with different
demographics. She measured the total value of combined programs and also tested
if diverse programs were preferred to single programs. The study concluded that
diverse programs were preferred and showed higher WTP estimates, which varied
with location. This may explain her slightly higher mean WTP. The Clemons,
Collins, and Green (1995) study, measured WTP for a program to eliminate the risk
of exposure to nitrate and volatile organic compounds (VOC) in well water in a
single location. They found that income and perceived seriousness of the
contaminants affected estimated WTP.
Previous studies have shown that an individual's attitude, background and
perception of the environment or resource could influence her WTP. For example,
if the respondent perceives that the quality of their water is good or there is no
problem with their water supply, then it is unlikely that they would be willing to
pay for a program of improvement. Also, if respondents have income problems they
may be less likely to pay for the program. Many respondents expressed concerns
about their income, household budget and taxation, which may in turn limit their
WTP.
Results from this study show that the respondents with lower incomes are
more likely to participate in and pay for a program to protect well water. Also,
previous studies have shown that lower income earners are likely less educated and
have higher" non-response rates to surveys than their counterparts (Kanuk and
Berenson, 1975; Mitchell and Carson, 1989). This may have important implications
for policymakers who seek effective measures to controlling nitrate pollution. For
example, the Florida Department of Environmental Protection (DEP) reported that
their efforts to reduce contamination levels in well water in some parts of Florida
have failed in large part from a lack of full cooperation from well owners and
operators (DEP, 1999). Because these owners and operators are likely lower income
and less co-operative, officials may wish to consider targeting their efforts to reach
this specific demography.
It is noteworthy to mention that the Franklin and Gulf county area has a
history of disagreement between many of the residents and the government of
Florida. This is particularly true with many residents deriving their primary income
from the fishing industry. The 1995 statewide ban on commercial net fishing and
routine closures of the oyster beds for phyto-sanitary reasons have caused a strain
on communications between these resource-dependent residents and the state
regulatory agencies. This degree of suspicion may influence the WTP response for
some, particularly those provided the state payment vehicle, and may influence the
outcome.
FUTURE STUDIES
Future studies should examine how varying demographics, goods and
services might affect this result. For example, one could conduct payment vehicle
research in counties with relatively high government employment, a younger or
older pollution and by using a different water supply contaminant. Future studies
could also repeat the payment vehicle research (in Franklin County) with other
goods or services such as well water bacteria and compare the results with this
study.
LIMITATIONS OF THE STUDY
The low response rate is a primary limitation. The initial sample frame, with
its many erroneous addresses, when combined with the limited budget and time
constraints contributed to lower the response rate. However, even with the small
sample, the payment vehicles still differ at the 0.15 level.
A second limitation also involves the sample selection. As noted earlier, the
demographic composition of the sample is not reflective of the counties studied.
This restricts the application of the WTP figures and it is not appropriate to use
these estimates to make inferences about the area residents' valuation of nitrate-
reduced water. However, this limitation has no impact on the study's stated
objectives. Also, the study included people without well water. Even though they
were instructed to complete the survey as if they used well water, many respondents
stated they found the process difficult to follow. Lastly, the absence of
some, particularly those provided the state payment vehicle, and may influence the
outcome.
FUTURE STUDIES
Future studies should examine how varying demographics, goods and
services might affect this result. For example, one could conduct payment vehicle
research in counties with relatively high government employment, a younger or
older pollution and by using a different water supply contaminant. Future studies
could also repeat the payment vehicle research (in Franklin County) with other
goods or services such as well water bacteria and compare the results with this
study.
LIMITATIONS OF THE STUDY
The low response rate is a primary limitation. The initial sample frame, with
its many erroneous addresses, when combined with the limited budget and time
constraints contributed to lower the response rate. However, even with the small
sample, the payment vehicles still differ at the 0.15 level.
A second limitation also involves the sample selection. As noted earlier, the
demographic composition of the sample is not reflective of the counties studied.
This restricts the application of the WTP figures and it is not appropriate to use
these estimates to make inferences about the area residents' valuation of nitrate-
reduced water. However, this limitation has no impact on the study's stated
objectives. Also, the study included people without well water. Even though they
were instructed to complete the survey as if they used well water, many respondents
stated they found the process difficult to follow. Lastly, the absence of
69
a control in the survey design likely limited the study, making it unclear whether
respondents' WTP were impacted by the institutional basis of the payment vehicle
or if it was their actual WTP for nitrate-reduced water.
APPENDICES
71
LIST OF APPENDICES
1. Appendix A: The cover letter ........................................................72
2. Appendix B: Comments to questions 1 & 4 of the survey instrument ...............73
3. Appendix C: Institutional Review Board letters of approval of the survey .......81
NOTE: Appendices are arranged in the order as they appeared in the list.
APPENDIX A: The cover letter
November 24, 2002
Name xxx
Street address xxx
City xxx, State xxx Zip code xxx
Dear Participant
As you know, the quality of drinking water is important to your health. Like food, water is
essential for life. While Florida has large quantities of quality groundwater, unfortunately, the
strains of a growing population has resulted in some instances of temporary and localized water
pollution. One potential source of ground water pollution is the presence of nitrates, naturally
occurring and man-caused chemicals that can often be found in untreated groundwater. While water
provided by public utilities is safe, some wells may contain levels of nitrates high enough to cause
minor health problems for humans.
As part of an on-going research effort by the Environmental Cooperative Science
Center (ECSC), we would like to ask for your help in evaluating the cost effectiveness of a
prospective system to remove nitrates from well water. The ECSC was established to conduct
research on science and policy issues pertaining to the Apalachicola Bay National Estuary Research
Reserve in partnership with scientists and managers from Florida A&M University, National
Oceanic and Atmospheric Administration and the Florida Department of Environmental Protection.
Your participation is extremely important and your input is crucial. Whether you purchase
bottled water, drink water from your own well or from a public utility, we still value your
opinion. Please help us by taking the time to complete this survey and return it to us in the
enclosed postage- paid envelope.
Again, your help is very important. While the nitrate-removal system you are asked to
evaluate is only hypothetical, your response will assist us in understanding the cost of nitrates
in water. If you have any questions about this survey, or comment you wish to share, please contact
either Ms. Carmen Lyttle-N'Guessan or Dr. Michael Thomas at (850) 599-3729.
Thank you for your help in this important matter.
Sincerely,
Larry Robinson, Ph.D.
Director
Environmental
Cooperative Science Center
7 Led by Environmental Sciences Institute, the College of Engineering Sciences, Technology and
Agriculture, and the College of Arts and Sciences.
APPENDIX B: Comments to questions one and four
COMMENTS ON SURVEY RESPONSE BY NUMBERS
Question Number 1
(1) "All of these issues are important, but is spending more money the answer,
no, science education is".
(2) "Remove George Bush from office. Also Jeb Bush".
(3) "Morality issues, especially in our young people (under age 50)!"
(4) "Reduce crime should be a local problem and solved by better law
enforcement."
(5) "Spend less money on silly survey."
(6) "Spend more or whatever it takes to stop abortion altogether. That is why
the judgments of God are coming more and more on the United States.
Individuals, read your bible, it's in there.
(7) "Quit raising taxes and use the lottery money for improving schools; which
was all a lie. But that's what the politicians promised, right." They could
also use some of to repair roads, stop crime, our pollution problem, help
endangered species and the list goes on."
(8) "I cannot offer valid responses. While I endorsed all proposals listed, money
is not the answer. Case in point: Education. Washington D.C. has the
highest per student expenditure in the country, yet it languishes at the
bottom of education performance. Spend more? No!"
(9) "All issues raise above have validity. But without knowing how well the
current expenditure is being handled, I have no way of answering these
statements intelligently."
(10) "Reduce development and Franklin County. We must save some pristine,
(and ours is the last of it! Areas!) both for our children and grandchildren, and
save the last opportunity for them to see the Bald eagle, Ospreys, Dolphin pods,
fishing and marine, and estuarine integrity of Apalach Bay areas, so they can
see them- in the wild! (instead of pictures in a book!). We don't have the right
to rob them of this- it's our greatest resource! .... And theirs!"
(10) "Pray in schools and government buildings."
(11) "Consolidation of many government agencies."
(12) "Alternative fuel options for automobiles"
(13) 'Think of all the money and resources it would free up, if they legalized
marijuana, not to mention the tax $$ it could generate."
(14) "Saving forest land. Stop offshore drilling for oil."
(15) "Entertainment or after school programs for teen. Too much teen drugs
use."
(16) "Unfair income tax law. Service to veterans stink!! Water quality in
Franklin County is non-existent."
(16) "Enforce present laws instead of spending to mitigate crimes' results."
(17) "More money should be spent on teachers pay to encourage teaching as a
profession."
(18) "Spend on us instead of outside country."
(19) "Clean up forests where people have dumped their garbage."
(20) "Get rid of St. Joe Land and Timber would make one hell of an
improvement in well water. If you really give a 'shit', look at what has been
dumped behind Highland View- near Port St. Joe, FL (Tim Stein, 850-648-
4522)."
(21) "Reducing taxes. Reduce socialistic trends in government."
(22) "Live within present budget."
(23) 'To accomplish many of the goals outlined, a mass transit system is
critical to reduce reliance on fossil fuel. At some point, we must abandon
the single person automobile on the way to and from the workplace!"
(24) "Control our borders."
(25) "Build more nuclear power plant. Cost per k.w.h considerably less and no
pollution."
(26) "See Attachment."
(27) "I am not convinced that spending more money is the answer to improving
some of these situations."
(28) "Help local rural hospitals stay open"
(29) "Improve EPA DEP. Eliminate 'cost' of 'engineer.'"
(30) "Affordable good health cares for everyone. Controlled growth."
(31) "Stay out of Iran."
(32) 'Too many high rising condos and buildings that block the beauty and
fresh air of the Gulf of Mexico. Government controls all natural resources
for all and not just the wealthy who can financially afford these luxuries.
Question Number 4 (e)
(1) "Why do we have to burden the owner with monetary costs of the
program? If this a problem, then it affects more than just well owners.
Why not channel some of the foreign aid for a change and help me the
citizens of the country." [Pu]
(2) "Clean Water Act should cover this." [Pu]
(3) "My drinking water comes from the utility I pay for good water." [SF]
(4) "We already pay taxes. The money should come from tax money." [SF]
(5) "Well water use for irrigation only." [SF]
(6) "We have a shallow (16 ft.) well that we use only for irrigation, car
wash, boat wash, etc. We are on city water for all household uses."
(7) 'There appears to be little/no evidence that a problem exists." [Pu]
(8) "People do not use money wisely in the State!"
(9) "I do not have a well. I am a utilities customer in the town in which I
live in (Carrabelle). The information sought and its remedy is immaterial
to me and does not apply. Call upon well owners, and have them pay the
fee."
(10) "I am on city water. At what our water rates are, I feel that the East
Point Water or Sewage District should be the one to pay."
(11) "The water company get most of my money know, and I can't drink it
half the time. I pay a $100 water bill." [Pu]
(12) "Scientists are not always reliable." [Pu]
(13) "I am on city H20 and the city should check H20 level
contaminations." [SF]
(14) "We anticipate having city water available in the near future." [Pu]
(15) "I don't believe many people in these counties drink well water." [Pu]
(16) "Financial situation." [Pu]
(17) "I pay for H20 that is bad already. They need to fix that first. I spent 3
days in hospital for 'H-Pylori' that I got from drinking H20." [Pu]
(18) "Can't afford it." [SF]
(19) "We are already taxed nearly 70% of our income. We barely have
enough to get by now; use the taxes already taken, and quit trying to
take the way I make a living away from me; get the money from
N.O.A.A." [Pu]
(20) "There has never being a well on this property." [SF]
(21) "State and government funds can be obtained through the adjustment
of other non-effective programs still receiving money." [SF]
(22) "Slanted and'biased survey. Only those in favor reply." [SF]
(23) "I do not have a well." [SF]
(24) "I do not have well water." [SF]
(25) "It would be cheaper to buy a charcoal/carbon filter for my faucet."
[SF]
(26) "City water bill is very high." [Pu]
(27) "I cannot see them making it work smoothly, and the monies would be
used to line someone's pockets." [Pu]
(28) "We do not have private well from which we get our drinking water."
[Pu]
(29) "Need more information." [Pu]
(30) 'The Govt. needs to stay out of my life." [SF]
(31) "I do not like people telling me you need to do this when my well is
better the city water that has bleach smell in it!!" [Pu]
(32) "Insufficient information on what level of nitrates is harmful." [SF]
(33) "Only use this well 30-40 days per year and don't drink the water."
[Pu]
(34) "Use current tax revenues." [Pu]
(35) "No specific number of months to be billed the $10.00. Perhaps
residents of Gulf and Franklin should bear the cost of the analysis." [Pu]
(36) "I live on social security fixed income. I could not pay. I could never
pay out anymore because I do not have enough to get by on now." [SFJ
(37) "Pouring money into the program is not the answer." [SF]
(38) "I am on the city water system." [SF]
(39) 'Too much government interference in private sector." [Pu]
(40) "No money." [SF]
(41) "Selling property" [Pu]
(42) "My well is at a vacation cottage used only a few weekends per year,
thus minimum nitrate intake." [SF]
(43) "I believe it's a scam." [SF]1
(44) "We are purchasing water conditioning system for our home that cost
over $2000.00." [Pu]2
(45) "I already have a filtration system." [SF]
(46) If this is a pilot- volunteer program then no one should be expected to
pay anything, unless it becomes clear that there is a large scale problem.
Other Comments
(1) 6 (E) "But it smells!"
(2) 7(E) is added by a number of respondents as 'City Water' and 'City water
treated in home'.
(3) 6 (D) "This city (Apalachicola) is in process of replacing all parts of the
water system. The water has been terrible for years."
(4) "Please find something worthwhile to do. I am a graduate of Florida A &
M." NB. Same respondent commented: "Spend less money on silly
surveys."
(5) 6 (D) "City does not do as they should. Fire hydrants are not flushed,
reservoir not clean."
(6) 6 (D) "Very. Water leaves brown settlement in glass."
(7) "Might have know you'd take on what could be a' thankless task',
eventually the? will figure it out, especially when they take a ride along the
Coast and can't even see the water because of shoulder to shoulder multi-
1 State of Florida
story condos (largely unoccupied for most of the year!). Don't need to
remind you of the Florida Keys and too, many sections of South and Central
Florida!
(8) "Respectfully, I do not use and have never used well water to my
knowledge. Therefore, I do not think I am qualified to respond to your
survey. However, I do pay taxes in the highest 2% category and become a
little sensitive when someone is seeking 'just one more $ per month'. It is
always just one more $ per month."
SPrivate utility
ftw[Iriba Mgriultural an1 Molpnil pniuitr-rsity
___ .__ TALLAHASSEE. FLORIDA 32307-4100
MA
STELEPHON.: (850) 599-8816
FAX: (850) 561-2794
xrlenacewtICaring 'APPENDIX C: Institutional Review Board letters of approval
of the survey
INSTITUTIONAL REVIEW BOARD
May 24, 2002
Mrs. Carmen Lyttle-N'Guessan
302 South Perry Paige Building
Florida A&M University
Tallahassee, Florida 32307
Dear Mrs. Lyttle-N'Guessan:
The Institutional review Board has reviewed your proposal entitled, "The Economic
Evaluation of Nitrogen on Groundwater Quality, using a Contingent Valuation Survey,"
reference number 002-029.
This project was approved, pending revisions. The Institutional Review Board Members
made the following recommendations:
1. The survey should be simplified; the language should be less complex, and
understandable to all participants.
2. It might be necessary to debrief the participants, if the current language about "Blue
Baby Syndrome" remains. The recommendation was made to remove the reference
to "Blue Baby Syndrome," which could cause unnecessary alarm in participants.
Perhaps, you could state that "high levels of nitrate in drinking water could be
harmful." This statement should be sufficient.
Please revise your proposal, and submit a copy of the survey at your earliest convenience.
Sincerely,
Verian D. Thomas, Ph.D.,
Chair
Institutional Review Board
FAMU IS AN EQUAL OPPORTUNITY/EQUAL ACCESS UNIVERSITY
porrita gri'ulturals, an a 32307-4100i tet
TALLAHASSEE, FLORIDA 32307-4100
INSTITUTIONAL REVIEW BOARD
MEMORANDUM
TO:
FROM:
DATE:
SUBJECT:
Ms. Carmen Lyttle-N'Guessan
Verian Thomas, Chairperson '
Institutional Review Board
April 18, 2003
"Economic Evaluation of the Effect of Nitrates in Groundwater: A
Contingent Valuation Survey in Northwest Florida" (003-13)
The Florida A&M University Institutional Review Board (IRB) has approved the
above-named project.
Please be reminded that you are now required to submit a written report, for review
by the Institutional Review Board, describing project activities completed, any
proposed changes in activities, and any problems encountered in the use of human
subjects for this project, by April 18, 2004.
FAMU IS AN EQUAL OPPORTUNITY/EQUAL ACCESS UNIVERSITY
Excellence With Caring
TELPHONE: (850) 599-8816
FAX: (850) 561-2794
REFERENCES
Ajzen, Icek and George L. Peterson. "Contingent Value Measurement:
The Price of Everything and the Value of Nothing?" Paper
presented at the National Workshop on Integrating
Economic and Psychological Knowledge in Valuation of
Public Amenity Resource. Fort Collins, Colorado, May
1986.
Amemiya, Takeshi. "Qualitative Response Model: A Survey." Journal
of Economic Literature. Vol. X1X, (1981) pp. 1483-1536.
Anderson, G. D., and R.C. Bishop. "The Valuation Problem." National
resource Economics: policy, Problems and Contemporary
Analysis. D. Bromley, ed., pp. 89-135, Washington, DC.
Urban Institute Press, 1986.
Anderson, David R., Dennis J. Sweeney, and Thomas Williams.
Statistics for Business and Economics, 6th. Ed., pp. 247-
260 West Publishing Company. 1996.
Arrow, K. Robert Solow, Paul R. Portney, Edward E. Leamer, Roy
Radner, and Howard Schuman. "Report of the NOAA
Panel on Contingent Valuation." Resource for the Future,
Washington D.C. 1993.
Binkley, Clarke S., and W. Michael Hanemann. 'The Recreation
Benefits of Water Quality Improvements: Analysis of Day
Trips in an Urban Setting." Report to the U. S.
Environmental Protection Agency (Washington, DC. 1978.
Bishop, Richard C., and Thomas A. Heberlein. "Measuring Values of
Extra-Market Goods: Are Indirect Measures Biased?"
American Journal of Agricultural Economics, Vol. 61,
(1979), no. 5, pp. 926-930.
Bishop, Richard C., and Thomas A. Heberlein. "Simulated Markets,
Hypothetical Markets, and Travel Cost Analysis:
Alternative Methods of Estimating Outdoor Recreation
Demand," Staff Paper Series no. 217, Department of
Agricultural Economics, University of Wisconsin. 1980.
Bishop, R. C., T. A. Heberlein, and M. J. Kealey. "Contingent
Valuation of Environmental Assets: Comparisons with a
Simulated Market." Natural Resource Journal, Vol. 23,
(1983), pp. 619-633.
Boyle, Kevin J. "A Review of Contingent Valuation Studies of
Groundwater Protection." Cooperative Agreement. CR 818
760-01-0. 1993.
Brookshire, David S., Ralph C. d'Arge, William D. Schulze, and Mark
A. Thayer. "Experiments in Valuing Public Goods," in V.
Kerry Smith, ed., Advances in Applied Microeconomics.
Greenwich, Conn., JAI Press. 1981.
Brookshire, David S., Larry S. Eubanks, and Allan Randall.
"Estimating Option Prices and Existence Values for
Wildlife Resources." Land Economics, Vol. 59, (1983), pp.
1-15.
Brox, James A., Ramesh C. Kumar, and Kenneth R. Stollery."
Estimating Willingness to pay for Improved Water Quality
in the Presence of Item Non-response Bias." American
Journal of Agriculture, Vol. 85, no. 2, (2003), pp. 414-428.
Cameron, T. A., and D. D. Huppert. "Referendum Contingent
Valuation Estimates: Sensitivity to Assignment of Offered
Values." Journal of American Statistic Association, 86,
(1991). pp. 910-918.
Cameron, Trudy A., and John Quiggin. "Estimation Using Contingent
Valuation Data from a Dichotomous Choice with follow-up
Questionnaire." Journal of Environmental Economics and
Management, Vol. 27, (1994), no. 3, pp. 218-234.
Carson, Richard T. Nicholas E. Flores and Norman F. Meade.
"Contingent Valuation: Controversies and Evidence,"
Environmental and Resource Economics, Vol. 19, (2001),
pp. 173- 210
"Contingent Valuation: Controversies and Evidence,"
Environmental and Resource Economics, Vol. 19, (2001),
pp. 173- 210
Clemons, Roy, Alan R. Collins, and Ken Green. "Contingent
Valuation of Protecting Groundwater Quality by Wellhead
Protection Program." Division of Resource Management,
West Virginia University. 1995.
Ciriacy-Wantrup, S. V. "Capital Returns from Soil-Conservation
Practices," Journal of Farm Economics, Vol. 29, (1947),
pp. 1181-1196.
Crutchfield, Stephen R. and Joseph Cooper. "Valuing Risk Reduction:
The Example of Nitrate in Drinking Water." FoodReview,,
pp. 38 41. January April 1997
Davis, Robert K. "Recreation Planning as an Economic Problem."
Natural Resources Journal, Vol., 3, (1963), no. 2, pp. 239-
249.
DEP (Florida Department of Environmental Protection). Environmental
Problem Solving ." (1999). Available online at
http://www.dep.state.fl.us/ospp/eps/tncfail.htm.
deZoysa, A. Damitha N. "A Benefit Evaluation of Programs to
Enhance Groundwater Quality, Surface Water Quality and
Wetland Habitat in northwest Ohio". Ph.D. Dissertation,
Department of Agricultural Economics and Rural
Sociology, Ohio State University. 1995.
Dillman, J. J., and D. A. Dillman. "The Influence of Questionnaire
Cover Design on Response to Mail Surveys." Paper
presented at the International Conference on Measurement
and Process Quality, Bristol, England. April, 1995.
Dillman, Don A. "Mail and Internet Surveys: The Tailored Design
Method." Second Edition, John Wiley & Son, Inc. 605
Third Ave, New York, NY 10158-0012. 2000.
Everitt, B. S. 'The Analysis of Contingency Table: Monographs on
Applied Probability and Statistics. pp. 1-37. Chapman and
Hall Ltd. 1977.
Green, William H. Econometric Analysis. 2nd. New York Macmillan.
1993.
Green, William H. Limdep User's Manual and Reference Guide.
Version 7.0, Econometric Software, Inc. 1995.
Hallberg, G. R., and D. R. Keeney. "Nitrate Alley." In Williams A.
ed., Regional Groundwater Quality, Van Nostrand
Rainhold, New Yolk, pp. 277-322. 1993.
Hanemann, W. Michael. "A Methodological and Empirical Study of
the Recreation Benefits from Water Quality Improvement."
Ph.D. dissertation, Harvard University. 1978.
Hanemann, W. M. "Statistical Issues in the Discrete-Response
Contingent Valuation Studies." Northeastern Journal of
Agricultural Resource Economics, Vol. 14, (1984), pp. 5-
12.
Hanemann, W. M. "Some Issues in Continuous- and Discrete
Response Contingent Valuation Studies." Northeastern
Journal of Agricultural Economics, (1985) pp. 5-13.
Hanemann, W. M. "Willingness to Pay and Willingness to Accept:
How Much can they Differ?" American Economic
Reviews, Vol. 18, (1991), pp. 635-647.
Hanemann, W. M., and B. Kristrom. "Preference Uncertainty, Optimal
Designs and Spikes". Chapter 4 in Current Issues in
Environmental Economics, ed., P.O. Johansson, B.
Kristrom and K. G. Maler. Manchester UK: Manchester
University Press. 1995.
Hoehn, John P., and Alan Randall. "A Satisfactory Benefit Cost
Indicator for Contingent Valuation." Journal of
Environmental and Management, Vol. 14, (1987), pp. 226-
247.
Kanuk, Leslie, and Conrad Berenson. "Mail Surveys and Response
Rates: A Literature Review." Journal of Marketing
Research, Vol. 12, (1975), pp. 440-453.
Kennedy, Peter. A Guide to Econometrics. 4th. Ed., MIT Press.
Cambridge, Massachusetts. 1998.
Krutilla, John V. "Conservation Reconsidered." American Economic
Reviews, Vol. 57, (1967), pp.787-796.
Leeds, R. and Larry C. Brown. "Non-point Source Pollution: Water
Primer." AEX-465 Agricultural Engineering Department,
Ohio State University, Ohio. 1992.
Li, C. Z., and L. Mattsson. "Discrete Choice under Preference
Uncertainty: An Improved Structural Model for Contingent
Valuation." Journal of Environmental Economics and
Management. Vol. 28, (1995), pp. 256-269.
Loehman, Edna T., and Vo Hu De. "Application of Stochastic Choice
Modeling to Policy Analysis of Public Goods: A Case
Study of Air Quality Improvements." Review of
Economics and Statistics. Vol. 64, no. 3, (1982), pp. 474 -
480.
Mitchell, R. C., and R. T. Carson. "Valuing Drinking Water Risk
Reductions Using the Contingent Valuation Method: A
Methodological Study of Risks from THM and Giardia,"
draft report to the U. S. Environmental Protection Agency,
Washington D. C. 1986.
Mitchell, R. C., and R. T. Carson. "Evaluating the Validity of
Contingent Valuation Studies." In G. L. Peterson, B. L.
Driver, and R. Gregory, ed., Amenity Resource Valuation:
Integrating Economics with other Discipline. pp. 187-200.
State College, PA: Venture Publishing, Inc. 1988.
Mitchell, R. C. and Richard T. Carson. "Using Surveys to Value
Public Goods: The Contingent Valuation Method."
Resource for the Future. Washington, D.C. 1989.
Mitchell, R. C., and R. T. Carson. "Current Issues in Design,
Administration, and Analysis of Contingent Valuation
Surveys." in P. Johansson, B. Kristrom and K. Maler, eds.,
Current Issues in Environmental Economics, Manchester
University Press. 1995.
Pierzynski, G. M., J. T. Sims, and G. F. Vance. Soil and
Environmental Quality. Second edition. CRC Press Boca
Raton, London, New York, Washington D.C. 2000.
Pindyck, Robert S., and Daniel L. Rubinfeld. Econometric Models and
Economic Forecasts. 3rd. ed., McGraw Hill. Inc. 1991.
Poe, Gregory L., and Richard C. Bishop. "Measuring the Benefits of
Groundwater Protection from Agricultural Contamination:
Results from a Two-Stage Contingent Valuation Study."
University of Wisconsin, Madison and Hatch Grant. 142 -
297. 1992.
Rae, Douglas A. "The Value to Visitors of Improving Visibility at
Mesa Verda and Great Smoked National Parks," in Robert
D. Rowe and Lauraine G. Chestnut. Eds., Managing Air
Quality and Scenic Resources at National Parks and
Wilderness Areas, (Boulder, Colorado, Westview Press).
1983
Rail, C.D. "Groundwater Contamination: Source, Control, and
Preventive Measures." Technomic, p. 139. Lancaster, PA.
1989.
Randall, A., John P. Hoehn, and David S. Brookshire. "Contingent
Valuation Surveys for Evaluating Environmental Assets."
Journal of Natural Resources, Vol. 23, (1983), pp. 635-648.
Randall, A., and J. P. Hoehn.. "Benefit Estimation for Complex
Policies." In Folmer, H., and E. Vanlerland (ed.), Methods
and Policy Making. Environmental Economics, pp. 219-
230. Amsterdam: Elsevier. (1989)
Schulze, William D., Ronald G. Cummings, David S. Brookshire,
Mark A. Thayer, and R. Whitworth, and M. Rahmalian.
"Methods Development in Measuring Benefits of
Environmental Improvements: Experimental Approaches
for Valuing Environmental Commodities, Vol. 2, draft
manuscript of a report to the Office of Policy Analysis and
Resource Management, U.S. Environmental Protection
Agency, Washington, D.C. 1983
Taylor, C. Robert, and Klaus K. Frohberg. 'The Welfare Effects of
Erosion Controls, Banning Pesticides, and Limiting
Fertilizer Application in the Corn Belt." American Journal
of Agricultural economics, Vol. 59, (1977), pp. 39-83.
Truett, Lila J., and Dale B. Truett.. Managerial Economics: Analysis,
Problems and Cases, pp. 471-472, second Edition South
Western College Publishing, Cincinnati, Ohio. 1998.
|
RSS
HIDE
