|
igj, e.August 1999
/1n 5,
OMENT
Economics Report 99-1
PUBLIC PREFERENCES AND ECONOMIC
VALUES FOR RESTORATION OF THE
EVERGLADES/SOUTH FLORIDA ECOSYSTEM
*'- l ,l- -- i^ e
NAPLE S E mH [
PV PURtM*^I /L *
&-j WAVM Clf~nVTIQM JU E^' "L Cll* MS ^
UNIVERSITY OF
SFLORIDA
Food & Resource Economics Department
Florida Agricultural Experiment Station
Institute of Food and Agricultural Sciences
Gainesvilqp, FL 32611
J. Walter Milon
Alan W. Hodges
Arbindra Rimal
Clyde F. Kiker
Frank Casey
PUBLIC PREFERENCES AND ECONOMIC
VALUES FOR RESTORATION OF THE
EVERGLADES/SOUTH FLORIDA ECOSYSTEM
by
J. Walter Milon, Professor
Alan W. Hodges, Coordinator of Economic Analysis
Arbindra Rimal, Post-Doctoral Associate
Clyde F. Kiker, Professor
Frank Casey, Post-Doctoral Associate
August, 1999
Economics Report 99-1
Food & Resource Economics Department
University of Florida
Gainesville, FL 32611
PUBLIC PREFERENCES AND ECONOMIC VALUES FOR RESTORATION
OF THE EVERGLADES/SOUTH FLORIDA ECOSYSTEM
J. Walter Milon, Alan W. Hodges, Arbindra Rimal, Clyde F. Kiker, and Frank Casey
ABSTRACT
The Everglades/South Florida region is a unique, globally significant ecosystem that has been altered
by drainage and water control structures. These alterations have changed the overall quantity, quality,
and temporal distribution of freshwater flows and impacted wildlife species throughout the region.
Under state and federal legislative directives, restoration of the ecosystem is being planned yet little
research has been conducted to identify public preferences and economic values for alternative
restoration plans. This report describes the application of a multiattribute utility survey of nearly 500
South and Central Florida residents to evaluate tradeoffs between natural and social system dimensions
of the restoration problem. Both hydrological and wildlife species attributes were used to represent
alternative states of the ecosystem along with possible effects on municipal water supplies, farmland,
and annual household taxes. Statistical results show that respondents indicated strong preferences for
Everglades restoration but the responses varied depending on how the alternative states of the
ecosystem were represented. Also, these preferences were tempered by concern for the consequences
of restoration decisions on municipal water users and farmland acreage. Willingness to pay measures
derived from the sample indicate a maximum annual benefit from "full restoration" of approximately
$60 $70 per household per year over a ten-year period. Extrapolating these results to the Florida
population yields annual benefits of $342.2 $406.5 million or $3.42 $4.07 billion over a ten year
period. These benefits, however, decline rapidly and turn negative if restoration imposes high costs in
the form of water supply restrictions, losses in farmland acreage, and annual household taxes. This
survey represents an initial effort to document Floridians' preferences and economic values for
restoration of the Everglades/South Florida ecosystem. Multiattribute utility analysis provides a
flexible research tool to frame the decision problem, evaluate public preferences for alternative plans,
and develop measures of economic value. This type of social science research is an essential part of an
adaptive management approach to restoration planning and can be used to objectively evaluate how the
perceptions and economic values of Floridians and others change as new information becomes available
about the effects of restoration actions.
KEY WORDS: Multiattribute utility, decision analysis, economic valuation, ecosystem restoration
planning, survey research, Everglades/South Florida.
ACKNOWLEDGMENTS
This research was conducted under Cooperative Agreement No. 43-3AEL-6-80078 between the
University of Florida and the Economic Research Service, U.S. Department of Agriculture and the U.S.
Fish and Wildlife Service. The authors thank Dr. Mary Ahearn, Economic Research Service, U.S.
Department of Agriculture and Dr. Jon Charbonneau, U.S. Fish and Wildlife Service for their support
and insights over the course of this project. We also thank Drs. Peter Feather and Daniel Mullarkey,
Economic Research Service, for review and comment on the survey design and statistical results
reported in this paper. Dr. Andrew Laughland, U.S. Fish and Wildlife Service, also offered helpful
comments on the survey design.
A number of other individuals also made valuable contributions to this project. Stu Applebaum, U.S.
Army Corps of Engineers, and Agnes McLean and Carl Woehlcke, South Florida Water Management
District, offered helpful comments on the overall design of the survey and technical comments on the
development of hydrological attributes for the analysis. Drs. Lance Gunderson, Ronald Labisky, and
Peter Frederick, University of Florida, provided guidance and assistance on historical wildlife
populations and possible levels of these populations as a result of Everglades restoration efforts. Dr.
Ken Portier, Statistics Department, University of Florida provided guidance on the statistical design for
the field interviews. Dr. Donna Lee, Food and Resource Economics Department, University of Florida,
critiqued the multiattribute design and survey elicitation methods. Ron Thomas and Cindy Spence
from Educational Media and Services, Institute of Food and Agricultural Sciences, University of
Florida, provided excellent technical assistance and guidance during the development and production
of the video used in the field interviews. Mary Rife, Sandi Palmer and Kari Madison of Rife Market
Research, Inc., Miami, FL, provide invaluable services in conducting the focus group sessions,
administering the field interviews, and delivering high quality data sets. Dr. Elaine Lyons-Lepke of
Perceptive Market Research, Inc., Gainesville, FL, assisted in protests of the survey instruments. Drs.
Mark Harwell and John Gentile, University of Miami, and participants in the U.S. Man and the
Biosphere Human-Dominated Ecosystems Directorate meetings provided many helpful insights on
analytical concerns relating to Everglades/South Florida ecosystem restoration. Finally, Dr. Bonnie
Kranzer, Executive Director, Governor's Commission for a Sustainable South Florida, was a valuable
source of information and encouragement throughout the project. To all these individuals and many
others, too numerous to mention, who have graciously assisted us in this project over the past three
years we offer our very sincere thanks.
The views and opinions expressed are solely those of the authors and do not represent the opinions of
other individuals or sponsoring agencies. Any errors or omissions are the responsibility of the authors.
PUBLIC PREFERENCES AND ECONOMIC VALUES FOR RESTORATION
OF THE EVERGLADES/SOUTH FLORIDA ECOSYSTEM
TABLE OF CONTENTS
ABSTRACT ....... ............................................................ ii
ACKNOWLEDGMENTS ..................................................... iii
EXECUTIVE SUMMARY ....................................................... vii
SECTION 1. INTRODUCTION ................................................. 1-1
1.1 The Study Area and Problem Setting .................................. 1-1
1.2 Study Purpose and Objectives ........................................ 1-4
1.3 Overview of the Report ............................................ 1-6
SECTION 2. MULTIATTRIBUTE UTILITY THEORY AS A PLANNING TOOL ......... 2-1
2.1 Multiattribute Utility Theory ......................................... 2-1
2.2 Multiattribute Utility Preference Elicitation Methods ..................... 2-2
2.3 Scoring and Ranking Alternatives with Multiattribute Utility Theory ......... 2-4
2.4 Design of Multiattribute Utility Studies .............................. .. 2-4
SECTION 3. ATTRIBUTES AND TRADEOFFS FOR EVERGLADES RESTORATION .... 3-1
3.1 Endpoints, Attributes and Ecosystem Restoration ......................... 3-1
3.1.1 Public Perceptions and Focus Groups ............................ 3-2
3.2 Functional/Hydrological Attributes and Levels ........................... 3-3
3.3 Structural/Species Attributes and Levels ................ ............... 3-6
3.4 Annual Cost, Farmland and Water Use Restriction Attributes ............... 3-8
3.5 Attribute Combinations and Plan Alternatives ......................... 3-10
3.6 Other Attributes Not Included ....................................... 3-12
SECTION 4. SURVEY DESIGN AND RESPONDENT PROFILE ..................... 4-1
4.1 Description of the Interview Process ................................... 4-1
4.2 Respondent Selection ...................... ... .......... .... ..... 4-2
4.3 Respondent Socio-Demographic Profiles ............................... 4-2
SECTION 5. STATISTICAL RESULTS .......................................... 5-1
5.1 Statistical Modeling of Pairwise Choices ............................... 5-1
5.2 Sample Data ..................................................... 5-2
5.3 Statistical Results for the Multiattribute Models .......................... 5-4
5.3.1 Basic Hydrological Multiattribute Model Results ................... 5-5
5.3.2 Basic Species Multiattribute Model Results ...................... 5-7
5.4 Statistical Results for Models with Socioeconomic Characteristics Interactions 5-9
SECTION 6. EVALUATION OF ALTERNATIVE RESTORATION PLANS ............. 6-1
6.1 Voting, Ranking and Net Willingness to Pay Measures from
M ultiattribute M odels ................................ .... .......... 6-1
6.2 Overall Evaluation of Alternative Restoration Plans ....................... 6-2
6.2.1 Evaluation with the Hydrological Multiattribute Model .............. 6-2
6.2.2 Evaluation with the Species Multiattribute Model .................. 6-5
6.3 Effects of Socioeconomic Characteristics on Evaluations of Alternative Restoration
SECTION 7.
7.1
7.2
7.3
7.4
Plans ......................................................... 6-7
6.3.1 Evaluations of Alternative Plans with the Hydrological Multiattribute Model
by Respondents' Location and Environmental Donation Status ........ 6-7
6.3.2 Evaluations of Alternative Plans with the Species Multiattribute Model by
Respondents' Location and Environmental Donation Status ......... 6-13
SUMMARY AND EXTENSIONS ................... ............... 7-1
Study Objectives and M ethods ....................................... 7-1
Survey Results and Alternative Restoration Plan Evaluations ............... 7-2
Extending the Economic Valuation Results to the Florida Population ......... 7-4
Limitations of the Study and Suggestions for Future Research ............... 7-5
SECTION 8. REFERENCES ....................................... ......... 8-1
APPENDIX A. QUESTIONNAIRE AND INTERVIEWER GUIDE ..................... A-1
APPENDIX B. PAIRED CHOICE SETS FOR SPECIES MULTIATTRIBUTE MODEL ..... B-1
APPENDIX C. PAIRED CHOICE SETS FOR HYDROLOGIC MULTIATTRIBUTE MODEL C-l
APPENDIX D. INFORMATIONAL VIDEO SCRIPT ................................. D-1
APPENDIX E. SUMMARY OF INTERVIEWER EVALUATIONS OF RESPONDENTS .... E-1
LIST OF TABLES
Table 2-1. Example of a Multiattribute Decision Problem for Purchase of an Automobile...... 2-2
Table 3-1. Description of Attributes and Levels for the Hydrological Multiattribute Model .... 3-5
Table 3-2. Description and Results for the Species Multiattribute Model .................. 3-7
Table 3-3. Description of Attributes and Levels for Cost, Farmland and Water Use
Restrictions ... ....................................... ................ 3-9
Table 3-4. Example of Pairwise Choice for the Hydrological Multiattribute Model ......... 3-11
Table 3-5. Example of Pairwise Choice for the Wildlife Multiattribute Model ............. 3-11
Table 4-1. Socioeconomic Characteristics for the Survey Sample and Comparisons to the
Florida Population ................. .................................. 4-4
Table 4-2. Race/Ethnicity and Median Household Income for the Survey Sample, by County, and
Comparisons to the County Population ...................................... 4-5
Table 4-3. Priorities for State of Florida Expenditures by Program Area by Survey Sample
Respondents and Comparisons to the Florida Population ....................... 4-6
Table 4-4. Survey Respondent Attitudes about the Environment and Water Supply .......... 4-7
Table 5-1. Variable Definitions and Summary Statistics for the Hydrological and Species Attribute
Models............ ........ ... .............. ........ ....... ........ 5-3
Table 5-2. Coefficient Estimates for the Hydrological Multiattribute Models ............... 5-6
Table 5-3. Coefficient Estimates for the Species Multiattribute Model. .................. 5-8
Table 5-4. Coefficient Estimates for the Hydrological Multiattribute Model with Socioeconomic
Characteristics Interactions ........ .................. .................. 5-10
Table 5-5. Coefficient Estimates for the Species Multiattribute Model with Socioeconomic
Characteristics Interactions ............... .... ........ .. ............... 5-11
Table 6-1 Evaluation of Selected Restoration Plans with the Hydrological Multiattribute
Model ................................................................ 6-3
Table 6-2 Evaluation of Restoration Plans with the Species Multiattribute Model .......... 6-6
Table 6-3. Evaluation of Selected Restoration Plans with the Hydrological Multiattribute Model by
Respondent Location ................. .... ............................... 6-9
Table 6-4. Evaluation of Selected Restoration Plans with the Hydrological Multiattribute Model by
Respondent Past Donations ............ .............. ................... 6-11
Table 6-5. Evaluation of Selected Restoration Plans with the Species Multiattribute Model by
Respondent Location ................. .................................. 6-15
Table 6-6. Evaluation of Restoration Plans with the Species Multiattribute Model by Respondent
Past Donations ....................................................... 6-17
LIST OF FIGURES
Figure 1.2. The South Florida Ecosystem and Its Components ....................... 1-2
Figure 7.1. Relative Weightings of Attributes in the Multiattribute Models ............ 7-3
PUBLIC PREFERENCES AND ECONOMIC VALUES FOR RESTORATION
OF THE EVERGLADES/SOUTH FLORIDA ECOSYSTEM
EXECUTIVE SUMMARY
The Everglades/South Florida region is a unique, globally significant ecosystem encompassing over
69,000 square kilometers. For more than 100 years, the natural flow of water through the Everglades
has been altered by drainage and water control structures, in order to provide flood control and water
supplies for the rapidly growing human population and large agricultural industry in the South Florida
region. These alterations have lead to harmful changes in the overall quantity, quality, and temporal
distribution of freshwater inflows to the Everglades, saltwater intrusion into aquifers in coastal areas,
invasion by exotic plant species causing changes in plant community structure, and dramatic declines
in certain wildlife species typical of the Everglades.
Under state and federal legislative directives, restoration of the Everglades/South Florida ecosystem
is now in progress under the leadership of the US Army Corps of Engineers and the South Florida
Water Management District. Proposed plans for restoration are expected to require decades to complete
with costs in excess of $6 billion. Despite considerable efforts to base restoration decisions on sound
science, there has been relatively little formal research to inform the decision-making process about
public expectations for the restoration, or willingness to pay for restoration costs. The purpose of the
present study was to quantitatively evaluate the perceptions and preferences of the general public in
Florida regarding alternative possible outcomes for restoration of the Everglades/South Florida
ecosystem, and to use these results to estimate economic values associated with alternative restoration
plans.
Multiattribute utility theory has been widely applied to problems involving tradeoffs between multiple
conflicting elements. This study was designed as an application ofmultiattribute utility theory in order
to accommodate the inherent tradeoffs between natural and social system dimensions of the restoration
problem. After extensive consultation with agency staffs and two focus group sessions with the public,
the many different aspects of Everglades restoration were represented in terms of nine independent
attributes, as shown in the following table. A group of three hydrologic attributes represented water
levels and timing in different geographic portions of the Everglades: Lake Okeechobee, the Water
Conservation Areas, and Everglades National Park. Another set of three attributes represented different
groups of species and habitat types: wetland species, dryland species, and estuarine species. Finally,
three attributes were used to represent the social services and costs associated with restoration:
restrictions on water use by households, changes in farmland acreage in South Florida, and the annual
cost to households. Each of the attributes had three levels associated with complete restoration,
intermediate restoration, and no restoration as a baseline case. The values for each level of the attributes
corresponded to expected outcomes based on various hydrologic or biological models, agency planning
documents, or consensus views of experts consulted. The hydrologic and wildlife species attribute
levels were specified in relation to "historic" levels.
Attribute Names, Descriptions and Alternative Levels for Multiattribute Analysis of Everglades
Restoration.
Attribute
Attributes Description Attrib
Levels
Hydrologic Lake Okeechobee. Percentage of time that water levels and timing in Lake Okeechobee are 60%*
Model similar to historic, predrainage conditions 75%
Attributes 90%
E% erglades Water Conservation Areas. Percentage of area in the E\ erglades Water 50%*
Conservation Area having water levels and timing similar to historic, predrainage conditions 75%
90%
Everglades National Park and Florida Bay. Percentage of the area in E\ erglades National Park 50%*
and Florida Bay that has water levels and timing similar to historic, predrainage conditions 75%
90%
Species Wetland Dependent Species. Percentage of historic, predrainage population levels of wetland 20%*
Model dependent species such as wading birds and alligators 50%
Attributes 80%
Dry Land Dependent Species. Percentage of historic, predrainage population levels of dry land 50%*
dependent species such as deer, hawks and songbirds 60%
70%
Florida Bay Dependent Species. Percentage of historic, predrainage population levels of Florida 60%*
Bay dependent species such as pink shrimp, mullet and sea trout 75%
90%
Socio- Annual cost per household. Additional annual cost per household for water utilities $O*
Economic $25
Attributes $50
(both Restrictions on Household Water Use. Number of days per week that outdoor water use 3/10%*
Hydrologic allowed during dry years, and percentage reduction in indoor water use 2/25%
and Species 1/40%
Models) Farm land Reduction. Area (acres) of farm land that would be reduced in the Everglades 0*
Agricultural Area and western portions of Palm Beach, Broward and Dade counties 100,000
200,000
* Denotes baseline condition currently
Primary information in this study was collected through personal interview surveys of 480 randomly
selected households in five south and central Florida metropolitan areas (Miami, West Palm Beach, Fort
Meyers, Tampa and Orlando). Interviews were conducted in respondent's homes by an independent
market research firm. The survey samples were representative of their respective county populations
and the state as a whole in terms of sociodemographic characteristics such as income, race, age,
education, household size and political affiliation. The interview included presentation of an 11 minute
informational video about the Everglades/South Florida ecosystem in order to orient respondents to key
issues addressed in the survey. Survey respondents were given a pairwise choice task in which pairs
of systematically-generated combinations of six attributes were presented, and the respondent was
asked to choose the preferred attribute set. Either the hydrological or species attribute specifications
were presented to half of the respondents in each county, in order to compare preferences for alternative
restoration plans based on different representations of the ecological dimensions of the problem.
Responses from the pairwise choice tasks were used to statistically estimate multiattribute utility
functions for the hydrological and species attribute sets. Results of the statistical analysis and evaluation
are summarized in the figure below.
Relative Weightings of Attributes in the Multiattribute Models
LakeOkeechobee 5.8 Welland Speces 2.9
Water Conservation Areas 6.2 Dryland Speces .7
Everglades Natonal Park 2.4 Estuarine Spe 8.1
Cost Per Household .3 Cost Per Household -2.4
HM Water Use I -0.1 HHWater Ue I -0.4
HH Water Use II -2.3 HH Water Use I -1.7
Farn Land n EAA -1.9 Farm Land in EAA -0.9
-4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 10
Hydrological Model Species Model
* The hydrological multiattribute utility model indicated that respondents gave positive weights
to all three hydrological attributes and preferred potential restoration plans that would lead to
water levels and timing more similar to historical conditions in all three of the South Florida
hydrologic regions. However, the social service and cost attributes in the model revealed that
higher levels of annual cost, water restrictions, or reductions in farmland acreage were
negatively weighted.
* The species multiattribute utility model showed positive weightings for wetland and estuarine
species attributes indicating that restoring these species populations to be more comparable to
historical levels was preferred. However, a significant negative weighting on the dryland
species attribute indicated that restoration of these species was not preferred. Also, the
preferences for restoring wetland and estuarine species were tempered by negative weightings
for the annual cost, water restrictions, and farmland reduction attributes.
* The negative sign and magnitude of weightings assigned to the annual cost, water restriction,
and farmland reduction attributes in both the hydrological and species multiattribute models
were generally similar indicating that respondents expressed consistent preferences regardless
of the type of attribute used to represent the ecological dimensions.
* Higher levels of the wetland and estuarine species attributes were consistently preferred across
the sample indicating that the general public would more readily identify with a restoration
program that emphasizes species restoration rather than hydrological management.
Weightings for the multiattribute utility models were used to evaluate various hypothetical full and
partial restoration plans, both with and without social service and cost components.
* The results presented in the following table showed that Floridians expressed strong preferences
for plans that fully restore the hydrological conditions and wetland/estuarine species
populations of South Florida. Net willingness to pay amounts for the full restoration plans of
$59 and $70 annually for the hydrological and species multiattribute models, respectively,
provide measures of the maximum annual benefits per household.
* While the results for the hypothetical restoration plans indicated generally strong support for
restoration, respondents would not support a restoration plan that imposed high costs on
Floridians. Hypothetical restoration plans that included annual costs of $50 per household (or
$500 over 10 years) coupled with either farmland reductions of 100,000 or more acres or severe
restrictions on municipal water use during dry years received poor rankings and less than
majority support. Moreover, the net willingness to pay values for these high cost plans were
negative suggesting a potential loss in economic welfare. On the other hand, intermediate cost
levels were clearly not viewed as too overbearing, such as moderate restrictions on water use,
or annual costs of $25 per household.
* When the results were extrapolated to estimate the aggregate benefits of alternative
Everglades'South Florida restoration plans for the Florida population, the net willingness to pay
for full hydrological restoration without costs was $342.2 million annually, or $3.42 billion over
a ten-year period. Similarly, for full wetland/estuarine species restoration without costs the
aggregate willingness to pay was $406.5 million annually, or $4.07 billion over ten years. A
more realistic restoration plan which provided full hydrological restoration with annual costs
of $25 per household, a 100,000 acre reduction in farmland, and moderate water restrictions had
a net willingness to pay of $15.60 per household annually, or $907.9 million for the Florida
population over ten years.
This survey represents an initial effort to document Floridians' preferences and economic values for
restoration of the Everglades.'South Florida ecosystem. This type of social science research is an
essential part of an adaptive management approach to restoration decisions in which scientists and the
public learn about the effects of management actions. Future surveys of Floridians are necessary to
objectively evaluate how public perceptions and economic values may change over time. In light of the
national significance of the Everglades/South Florida ecosystem, additional surveys are necessary to
quantify the preferences and economic values of citizens in other regions of the U.S.
Evaluation of Selected Restoration Plans for the Hydrological and Species Multiattribute
Models.
Percent of Net
Plan Description Respondents Willingness
in Favor to Pay
Hydrological Multiattribute Model
Partial Hydrologic Restoration without Costs 71.7 $34.32
Full Hydrologic Restoration without Costs 71.7 $58.79
Partial Hydrologic Restoration with Minimized Costs 44.3 $9.32
Full Hydrologic Restoration with Minimized Costs 54.3 $15.60
Full Hydrologic Restoration with Full Costs 31.1 -$61.09
Species Multiattribute Model
Partial Wetland Wildlife Restoration without Costs 92.7 $34.93
Full Wetland Wildlife Restoration without Costs 92.7 $69.86
Partial Dryland Wildlife Restoration without Costs 17.9 -$11.95
Full Dryland Wildlife Restoration without Costs 17.9 -$23.90
Full Wetland Wildlife Restoration with Minimized Costs 67.9 $26.63
SECTION 1. INTRODUCTION
1.1 The Study Area and Problem Setting
The southern portion of Florida is a complex mosaic of hydrologically interrelated terrestrial,
freshwater, and marine ecosystems. The entire regional ecosystem is linked through surface and ground
hydrology from the Kissimmee River watershed, through Lake Okeechobee, through water conservation
areas, the Big Cypress Swamp, and the Everglades, and out through Florida Bay and the passes between
the Florida Keys to the offshore corals reefs (Figure 1). The uniqueness of the region has been
recognized through the designation of a number of state and federal parks and sanctuaries including Big
Cypress National Preserve, Biscayne Bay National Park, Everglades National Park, Pennekamp Coral
Reef State Park, and the Florida Keys National Marine Sanctuary. Also, Everglades National Park has
been designated an international Biosphere Reserve, a World Heritage ecosystem, and a Wetland of
International Importance (Maltby and Dugan 1994). The entire area covers more than 69,000 square
kilometers and is larger than nine states. More than one-third of the land area is in public ownership.
For more than 100 years, the flow of water through this network of habitats and ecosystems has been
modified from its natural conditions to satisfy the demands of a growing human population within the
region. More than half of the original Everglades system has been drained and water management
structures now channel more than four times as much freshwater to the Atlantic Ocean as natural
drainage flows (Davis and Ogden 1994). While many warned of the consequences of these
interventions (e.g. Beard 1938; Douglas 1947), a series of ecological and institutional crises over the
past two decades has focused scientific and public attention on the problems of the region and efforts
to restore the ecosystems (Light et al. 1995).
Restoration efforts for the Everglades/South Florida ecosystem have been proceeding under state and
federal legislative directives. The Everglades Forever Act of 1994 (Chapter 373.4592, Florida Statutes)
outlined a general plan to restore the ecosystem by improving the quantity and distribution of
freshwater, reducing the amount of phosphorus enriched agricultural storm water entering the system
(from agricultural production areas south of Lake Okeechobee), removing exotic plant species, and
setting time deadlines. The South Florida Water Management District is responsible for overall
coordination and administration of these efforts and has developed a planning document that describes
restoration projects over the next two decades (South Florida Water Management District 1994).
The Water Resources Development Act of 1992 (Public Law 580, 102nd Congress) authorized the U.S.
Army Corps of Engineers to initiate a study to determine whether the existing Corps' Central and
Southern Florida Project (and the water control structures built by the Corps) should be modified to
improve the quality of the environment, improve protection of the aquifer, and improve conservation
of urban water supplies. The Central and Southern Florida Project (C&SFP) dates back to the Flood
Control Act of 1948 and has been the focal point of federal efforts to provide flood control, drainage,
water control, navigation, water supply for Everglades National Park, and other purposes (U.S. Army
Corps of Engineers 1994, Appendix J). The objectives and accomplishments from the earlier years of
this Project have been cited by scientists as the source of many of the environmental problems
throughout the region (e.g. Light and Dineen 1994).
AweNT PALM
F.T C VO BEACH
FiIg urel ATSyh Foi : an om
NATIONI ".- -- --- -
A EV .W^-*ADE &
RExPANION AREA C -
^ -- r Kihometa
I-- OUTH FLOatoA, WATEB MANAflEMENT o ao so a 4
0,.TR 3CT i.nUNOAUNDARH^Y t -BI Y I
..... iELDl. NATIONAL PAK OUNNOATIY MiA
Figure 1-1 The South Florida Ecosystem and Its Components
The first phase of the Corps' study was a Reconnaissance Report that identified problems and
opportunities, described alternative plans, and recommended further studies (U.S. Army Corps of
Engineers 1994). The report presented ten basic restoration plans that ranged from changes in operating
procedures with existing physical structures to major redesign of canals, levees, and flowways
throughout the system. No specific restoration plan was recommended. At about the same time, other
restoration plans were proposed including: The Greater Everglades Ecosystem Restoration Plan by the
Everglades Coalition (1994), the Action Plan, South Florida Ecosystem Restoration by the U.S. Fish
and Wildlife Service (1993), and the Federal Objectivesfor The South Florida Restoration by the
Science Sub-Group of the South Florida Management and Coordination Working Group (1993).
The Water Resources Development Act of 1996 (Public Law 303, 104th Congress) directed the Corps
to proceed from the reconnaissance phase to the development of a comprehensive plan for the purpose
of restoring, preserving and protecting the Everglades/South Florida ecosystem. The plan also would
address other water related needs of the region such as water supplies and flood control (Section
309(1)). The Corps and the South Florida Water Management District (as local sponsor) must present
the plan to Congress by July 1999. Under the Act, cost sharing would be 50/50 between state and local
government.
As this planning effort (commonly referred to as "the Restudy") proceeds, it is useful to describe the
major planning problem as the choice of restoration goals. While a great deal has been written about
the need to restore the ecosystem, little has been offered to help the public understand what would be
achieved with restoration (Vogel 1998). In this regard, it is helpful to understand that the choice of
goals for a restoration project is equivalent to a choice of ecological endpoints. Ecological endpoints
are defined as those characteristics of the ecosystem that if changed, would constitute a change in the
health of the ecosystem (Harwell and Long 1992; Harwell et al. 1992). These endpoints represent
vectors of functions and services from the ecosystem that have both biological and social importance
(Russell 1995). The dilemma arises from the fact that choosing endpoints is in part an ecological issue
and in part a social issue. The essence of the problem is clearly identified in the Corps' Reconnaissance
Report:
The vision of the future wetlands in south Florida is influenced by different views of
how we determine restoration goals for the system. The future Kissimmee River, Lake
Okeechobee, Everglades, Big Cypress, and Florida Bay ecosystems can be, to some
extent, what we want them to be, based on our value systems, and our decisions about
what conditions and components constitute a restored ecosystem (U.S. Army Corps of
Engineers 1994, p. 109).
This dilemma might be resolved solely at an administrative level by agencies participating in the
planning process. But as Russell (1993) points out, public preferences and economic values for
alternative endpoints must be considered since restoration projects compete with other public projects
for funds. And, economic values for alternative endpoints cannot be easily measured since many of
the possible endpoints involve environmental goods that are not valued in traditional markets or directly
related to observable human behavior. Hence, any complete feasibility analysis of alternative
restoration plans must consider public preferences and values for alternative ecological endpoints.
Recognizing the need for information about public preferences and values for alternative ecological
endpoints does not, however, readily translate into a well-defined research program. Most economic
studies of environmental goods have focused on a single dimension or endpoint while assuming (most
often implicitly) that other endpoints were fixed in time or irrelevant (Russell 1993). Studies that have
considered multiple alternative endpoints (e.g. Crocker 1985; Hoehn and Loomis 1993; Loomis et al.
1990; Opaluch et al. 1993) encountered the well-known principle that individuals ability to make
consistent choices decreases as the number of options increases (Miller 1956). Professionals from both
the natural and social sciences who have considered the current state of affairs have concluded that
research should focus on identifying the ways that individuals nonscientistss) think about environmental
problems and begin to evaluate the effectiveness of alternative value elicitation procedures involving
multiple ecological endpoints (Green and Tunstall 1991; Harwell et al. 1992; Russell 1993).
While there are many ways to characterize ecological endpoints for ecosystem restoration (Bratton
1992; Cairns 1988; Harwell and Long 1992), two broad classifications are "structural" endpoints and
"functional" endpoints. Structural endpoints focus on the number and diversity of individual species,
and may include keystone species or other distinctions to differentiate the species (Paine 1980,
Westman 1985, Wilson 1992). For the Everglades/South Florida ecosystem, these species indicators
might include bird populations, fish populations, and/or populations of endangered species such as the
Snail Kite and Wood Stork. On the other hand, functional endpoints emphasize broad ecosystem
processes and dynamic properties (Holling 1987, Westman 1985, 1991). In an uplands/wetlands
ecotone such as the Everglades, these endpoints might include hydroperiod levels and timing, salinity
fluctuations, spatial extent, and fire frequency (Harwell and Long 1992, Holling et al. 1994). While the
exact linkages between these alternative endpoint classifications are not well known, in general the two
sets of endpoints represent different dimensions of an ecosystem. Ecosystem management based on
one set of endpoints may lead to different restoration decisions than for the alternate endpoints (May
1973, McNaughton 1988, Westman 1985, 1991).
1.2 Study Purpose and Objectives
The purpose of this study was to develop and implement a public survey elicitation procedure that could
be used to evaluate public perceptions of alternative ecological endpoints for what may well be the
'granddaddy' of all ecosystem restoration efforts, the Everglades/South Florida region. The elicitation
procedure would also provide estimates of individual's economic value (willingness to pay) for bundles
of environmental goods that could result from alternative restoration plans. Results from the survey
could then be used to rank alternative restoration plans and provide measures of the economic benefits
(net willingness to pay) associated with alternative ecological endpoints. This study also contributes
to the state-of-the-art in ecosystem valuation methodology by evaluating respondents' perceptions of
alternative representations of environmental functions. Specifically, this will involve a test of the
hypothesis that sets of wildlife species population levels (species endpoints) elicit the same preferences
for ecosystem restoration as functional endpoints such as the level and timing of water flows in different
areas within the region. Specific objectives for this study were:
1) To develop a public survey instrument for eliciting preferences for alternative restoration plans using
a multiattribute utility framework.
Rationale: Management of a complex ecosystem such as the Everglades/South Florida
ecosystem involves hydrologic, vegetative, and faunal processes that imply multiple objectives and
multiple alternatives for human decision making. Multiattribute utility theory (MAUT) provides a well-
known analytical framework in which the attributes of decision alternatives can be arrayed and
statistical procedures can be used to measure respondents' preferences for different alternatives (Keeney
and Raiffa 1976; von Winterfeldt and Edwards 1986). MAUT can be implemented in a discrete choice
framework (McFadden 1986) to evaluate alternatives using several different response formats such as
paired comparisons (e.g. Opaluch et al. 1993) or multiple rankings (e.g. Beggs et al. 1981). For
example, Opaluch et al. (1993) evaluated public preferences for alternative landfill locations based on
site attributes such as acreage of the landfill and adjacent land uses and location attributes such as
proximity to residential areas, access roads, and hauling costs. In addition, the MAUT approach can
be used to derive estimates of economic value for changes in attribute levels (Louviere 1988; Swallow
et al. 1994).
2) To develop two representative sets of ecological attributes based on functional and structural
endpoint concepts for use in a multiattribute utility survey instrument.
Rationale: Ecosystem management has often focused on the levels of individual species within
an area as an indicator of ecosystem health (Wilson 1984, 1992; Bratton). Laws such as the Endangered
Species Act direct wildlife management efforts to species level protection and these laws focus public
perceptions on changes in specific ecosystems. In fact, declines in the number of nesting colonies for
the endangered Wood Stork and Snail Kite have been one of the primary factors driving current efforts
for South Florida Ecosystem/Florida Bay restoration (Odgen 1994). An alternative, though not
necessarily conflicting approach to ecosystem management is to focus on broad ecosystem functions
(Holling 1987). In this approach, individual species endpoints are less important than overall functional
endpoints such as the periodicity of wetland flooding and drying or the diversity of micro and meso
habitats within the overall ecosystem (Holling et al. 1994). With this approach, species levels may be
less stable (although potentially more productive) than with a structural endpoint focus. The
significance of these two approaches to endpoint indicators is that they convey different conceptual
models of ecosystems that may influence individuals' perceptions of restoration and the choice of
actions that would be undertaken as part of an actual restoration.
3) To implement the multiattribute utility theory (MAUT) survey to measure Floridians' preferences
for South Florida Ecosystem/Florida Bay restoration; to test whether structural endpoints or functional
endpoints influence individuals' preferences for restoration; and, to determine whether there are
geographical or other sociodemographic differences in preferences.
Rationale: Hypothesis tests based on multiattribute utility survey responses can be used to
understand Floridians' preferences for restoration initiatives. This information may influence the choice
of specific restoration plans. A primary hypothesis for this research is that individuals' preferences for
restoration alternatives will not be influenced by the use of structural or functional endpoints to
represent alternative restoration plans. This hypothesis can be tested using a split sample survey design
whereby different groups of respondents are presented with either structural or functional endpoint
choice sets. The results from this analysis will provide information about the structure of public
preferences for restoration and the specific attributes of restoration plans that are most important to the
public. A secondary hypothesis is that individuals' preferences are the same regardless of where they
live in Florida or their sociodemographic characteristics (e.g. age, income). This hypothesis can be
tested using a representative sample of respondents in different regions of Florida.
4) To use the results of the multiattribute utility survey of Floridians to rank alternative restoration
plans and to estimate economic values associated with these restoration plan alternatives.
Rationale: With the multiattribute utility survey results it is possible to rank alternative plans
based on the type and degree of change in endpoints provided by a specific ecosystem restoration plan.
These rankings can provide guidance to restoration planners and elected officials about the relative
merits of alternative plans. Differences in rankings can also be identified for various sociodemographic
groups. Survey results also provide measures of willingness to pay for specific restoration plans and
the marginal economic value individuals assign to attribute endpoints. These values can be used to
measure the economic benefits of alternative restoration plans to Floridians. Also, the marginal value
of endpoint attributes can pro\ ide information about the economic benefits of individual components
of a restoration plan.
1.3 Overview of the Report
This report is organized as follows. Section 2 provides an overview and description of multiattribute
utility theory and how it can be used as a tool for ecosystem restoration planning. Section 3 describes
the selection of functional attributes based on hydrological properties of the ecosystem and structural
attributes based on species populations. This section explains the basis for selecting these attributes and
how alternative levels for these attributes were specified based on a variety of information sources
including the U.S. Army Corps of Engineers, the South Florida Water Management District, and
scientists from various agencies and universities around the state of Florida. Also, this section provides
a description of the procedures used to combine the selected functional and structural attributes into
alternative "plans" to be used in personal household interviews with Florida residents. In Section 4, the
interview process and the materials used in the interviews are described. Procedures followed to
identify a stratified, random sample of residents in three South Florida counties (Dade, Lee, and Palm
Beach) and two Central Florida counties (Hillsborough and Orange) are also described. This is
followed by a summary of results from the interview surveys for sociodemographic and other
descriptive information about the respondents.
More detailed information about statistical modeling of the multiattribute choice process used in this
survey is presented in Section 5. This section also provides statistical results for various analytical
models estimated from the survey data along with an interpretation of the results of the statistical
analysis. Section 6 extends the statistical results presented in the previous chapter to develop rankings
and willingness to pay measures for alternative restoration plans based on relative scores from the
multiattribute utility models. Rankings are presented for the functional (hydrological) attribute models
and the structural (species) attribute models based on results for all survey respondents and for various
sociodemographic subgroups within the sample. Section 7 concludes the report with a summary of the
survey design and results of the analysis. This section also provides a discussion of possible uses of
this information in Everglades/South Florida restoration planning along with a discussion of the
limitations of the survey and results. Appendices to this report provide complete copies of the questions
and materials used in the interviews along with other detailed information to support results provided
in the main sections of the report.
SECTION 2. MULTIATTRIBUTE UTILITY THEORY AS A PLANNING
TOOL
2.1 Multiattribute Utility Theory
Multiattribute utility theory (MAUT) is a method among a class of procedures known as "multi-criteria
decision making" which are used for analysis of problems that have a number of disparate dimensions
that must be considered simultaneously. Basically, a multiattribute problem consists of one or more
decision alternatives that are evaluated by a decision makers) based on a set of attributes that are
deemed essential to the problem. The decision maker's choices reflect an implicit weighting assigned
to each attribute that reflects the importance of each attribute to the decision maker.
A primary strength of MAUT is in evaluating alternatives where there are tradeoffs, i.e. where one or
more alternatives are superior in one respect while inferior in another. Such tradeoffs are involved in
many, perhaps most, real world decisions. MAUT may be used to identify a single "best" option, to rank
the options in order of preference, or to identify a subset of acceptable or nondominated options for
further analysis. A virtue of the methodology is that attributes may be either quantitative or qualitative
in nature, and it is able to accommodate important but ill-defined or subjective dimensions of a
problem.
MAUT is suitable for problems in which there is uncertainty about the realization of different
alternatives. In cases where there is not any consideration of uncertainty, the methodology is known as
"multiattribute valuation" (MAV). MAUT and MAV techniques have been applied in numerous
settings, including technology selection, energy and transportation planning, strategic business
planning, product marketing and environmental management (Giocoechea, Hansen and Duckstein,
1982; Hwang and Yoon, 1981; Keeney, 1980; Keeney and Raiffa, 1976; Mollaghasemi and Pet-
Edwards; Nijkamp, Rietveld and Voogd, 1990; Saaty, 1980; Szidarovsky, Gershon and Duckstein,
1986; von Winterfeldt and Edwards, 1986).
To illustrate the basic principles of MAUT, consider the decision by a consumer to purchase an
automobile. Two possible models (A,B) may differ in terms of a number of attributes such as price,
fuel economy (miles per gallon), passenger seating capacity, performance (horsepower, maximum
speed, acceleration), safety, handling characteristics, appearance, and so forth. These attributes can be
arrayed to provide a direct comparison between attributes as in Table 2.1. A vehicle which has a lower
price may have fewer safety features while a vehicle that has superior performance may have lower fuel
economy. The decision as to which vehicle to purchase depends upon the relative weightings the
individual gives to the different attributes and the cumulative value or "score" assigned to each
alternative. If an individual attribute is strongly weighted, the choice may be determined in favor of
the vehicle that is rated most highly on this attribute. Alternatively, if there
are only weak differences in weightings among attributes, several or all attributes may influence the
decision.
Table 2-1. Example of a Multiattribute Decision Problem for Purchase of an Automobile.
Attribute Model A Model B
Price (new) $25,000 $15,000
Safety Has air bags No air bags
Performance (Horsepower) 250 Hp 200 Hp
Fuel Economy 20 miles per gallon 30 miles per gallon
Seating Capacity 6 persons 4 persons
Solution of a MAU or MAV problem typically involves the following steps (von Winterfeldt and
Edwards, 1986):
Respondents assign relative weights to the attributes.
Develop utility or value functions for individual attributes.
Respondents evaluate each alternative separately against each attribute.
Aggregate the weighted evaluations to obtain an overall evaluation of each alternative by
means of an appropriate aggregation rule.
The selection of meaningful, appropriate attributes for the decision is one of the most critical and
difficult steps for solving multiattribute problems. The attributes must be essential to the decision
problem, but each attribute should reflect independent dimensions to the degree possible to avoid
redundancy (Keeney and Raiffa 1976; Louviere 1988). The attributes must be measurable and must
also be understood by decision makers. Theoretically, any number of attributes may be used in a
multiattribute decision problem. The literature, however, indicates that in practice only seven to nine
attributes can be meaningfully evaluated by decision makers due to limited cognitive skills and memory
capacity of most individuals (Saaty 1980; Miller 1956; de Palma et al. 1994).
2.2 Multiattribute Utility Preference Elicitation Methods
Elicitation of the preferences of decision makers involves weighting the relative importance of the
attributes and an assessment of the utilities of different levels of the attributes. Because of the subtlety
and complexity of this information, it is typically gathered through extended personal interviews with
individual decision makers. The preference elicitation procedure is repeated for several choices in order
to assure consistency of the preferences expressed. There is much
experimental evidence to indicate that as the complexity of the decision increases, the reliability and
consistency of respondent judgements decreases (Borcherding, Eppel and von Winterfeldt 1991; Boxall
et al 1996; De Palma et al. 1994; Mazzotta and Opaluch, 1997). Therefore, properly designed
multiattribute surveys present the task to the decision maker in clear and simple terms and provide
opportunities to learn about the task through practice decisions before the actual decision choices are
presented. Also, the total number of times the decision task is repeated is limited to avoid respondent
fatigue.
A variety of techniques have been used to assess multiattribute utility or value functions. These
techniques differ in the degree of difficulty and the nature of the preference information expressed
(Keeney and Raiffa 1976; Giocoechea Hansen and Duckstein 1982). Perhaps the simplest technique
is the pairwise choice, i.e. to choose the preferred alternative or attribute from a set of two choices
given. A slightly more difficult task is to rank-order a set of three or more alternatives in order of
preference. These techniques provide information on the order of preferences among a set of
alternatives but do not indicate the strength or intensity of preferences.
To assess the strength of preferences among alternatives or attributes, different techniques are used. In
the simple rating method, the respondent is asked to estimate a numerical value that represents his value
or utility at various levels of the attribute based upon some arbitrarily chosen scale. These values are
then normalized so that the sum of all weights equals unity. For example, the respondent may be asked
to first choose the most important attribute which is assigned a value of 100. Then all other attributes
are rated on a scale of 1 to 100. Another class of techniques involves establishing equally preferred
conditions to which the respondent is indifferent. In a technique known as "value splitting" or
"bisection", the decision maker is first asked to determine the upper and lower bounds for the value of
an attribute which are assigned utilities of 1 and 0, respectively. Then the respondent is asked to
identify a value for the attribute which represents the utility midpoint, i.e. a value that represents a
utility exactly halfway between the upper and lower bound, and this value is assigned a utility of 0.5.
The procedure can be continued to further split intervals of the utility function, until a reasonable
approximation of the utility curve is obtained. A similar approach, known as the difference standard
sequence, involves asking the respondent to construct a series of equal marginal value intervals. These
methods may be used to assess utility or value functions that are suspected to be non-linear, since a
minimum of three points are sufficient to define a simple utility curve that may be linear, convex or
concave in shape. A convex or concave utility function represents a risk-averse or risk-seeking strategy,
respectively. In the more sophisticated cross-attribute strength method, the attribute weightings are
determined by matching the strength of preference in one attribute to a strength of preference in another.
Similarly, the cross-attribute indifference method involves systematically varying alternatives in two
attributes to generate simple equations that can be solved for the attribute weights.
For estimating individual utility functions that explicitly consider uncertainty, various types of lottery
scenarios are used. For example, in the variable probability method, respondents are asked to choose
between two alternatives, one with a certain given value, the other with specified probabilities for either
a higher or lower value. The values of the specified probabilities are adjusted until the respondent is
indifferent, i.e. the utility is equal, then a set of equations can be solved to determine the utilities at
different levels of the attribute. A related approach is the variable certainty equivalent technique in
which respondents are asked to select a certain payoff value for which they would be indifferent to a
lottery with a 50% chance of a higher payoff amount and 50% chance of zero payoff. Then, by adjusting
the value of the higher payoff amount and observing the change in the respondent's certainty
equivalent, the utility function can be derived.
Conjoint analysis is a multiattribute decision technique that combines the elicitation of preferences
among both attributes and alternatives into a single step (Louviere, 1988). A series of multiple attribute
alternatives are presented to the decision maker in which specified levels are given for all attributes.
The respondent then chooses or rank-orders the preferred alternativess. The advantage of this approach
is that the choice task is set in the more realistic context of choosing directly among alternatives with
the complete set of associated attributes. A drawback is that the most important attributes used by
decision makers to select the preferred alternative are not revealed directly. A relatively large number
of choices must be made with carefully constructed combinations of attribute levels to determine the
attribute weightings that were implicitly used by the respondent.
2.3 Scoring and Ranking Alternatives with Multiattribute Utility Theory
Once a set of individual choice decisions have been made involving a small number or a large sample
of decision makers, the individual evaluations are aggregated to give an overall value or score for each
alternative. The simplest general form of a utility aggregation function is described mathematicallN as
follows:
U(Xj) = "y I= Wi Ui(xj), forj = 1, 2, 3, ..., n,
where U(xj) is the utility of the jth alternative, Wi is the weight of the ith attribute, Ui is the utility
function for attribute I, and xii is the score given to thejth altemati e on the ith attribute. The alternative
with the highest value would be the most preferred option. The above function is additive with a linear
combination of the weighted values of each attribute. This functional form is valid when there is
mutual preferential independence among the attributes, i.e. when the preference for different values of
any pair of attributes does not depend on the level of other attributes. In cases where this does not hold.
a more complex multiplicative model may be used. However, in practice the additive form is generally
used for convenience because it reduces the number of choice repetitions required to derive acceptable
statistical results.
The output of a MAU or MAV procedure may be a cardinal (ratio scale) or ordinal ranking of the
alternatives. In the cardinal ranking, specific numerical values are given to each alternative that permits
a quantitative comparison of how much better or worse one alternative is versus another. With an
ordinal ranking, the utility model results only indicate that a particular alternative is more or less
preferred than another, but not by how much. In economic valuation studies, the marginal willingness
to pay (MWTP) for changes in a nonmonetary attribute can be estimated from the marginal utility
coefficient. The marginal utility coefficient for an attribute represents the change in utility
corresponding to a unit change in the level of the attribute. The MWTP is calculated by dividing the
marginal utility coefficient for an attribute by the marginal utility coefficient of a monetary attribute as
follows:
MWTP(I) = (QU/laI) / (aUm /lm)
where MWTP(I) is the marginal willingness to pay for changes in attribute I, dU/aI is the marginal
utility coefficient of attribute I, and aUm/lm is the marginal utility coefficient of the monetary attribute.
The net economic value of various alternatives can then be calculated by summing the MWTPs for
changes in the attributes for each alternative.
2.4 Design of Multiattribute Utility Studies
When multiattribute utility studies are conducted to characterize the preferences of a large group or
population, it is important that the set of attribute choice combinations is presented in a statistically
representative manner. This assures that the maximum amount of information is revealed by the study
and the information is unbiased. A variety of orthogonal factorial experimental design procedures are
available to assist in this task (Addelman, 1962). In a full factorial experimental design, all possible
combinations of attributes and levels are used. For example, in a study with 6 attributes (factors) with
three levels for each attribute, there are 36 or 729 possible attribute combinations. With a full factorial
design, all factor-level combinations can be tested and all possible main effects and nth-order interactive
effects can be evaluated in a statistical model. But, with a pairwise choice technique, a full factorial
design study with six attributes and three levels would require more than 360 choice decisions. Such
a large number of choices is impractical for most multiattribute studies.
Limitations in time, resources, and respondent attentiveness lead to the use of fractional factorial
designs to create a balanced sample of possible attribute combinations with some loss of information.
For example, a one-half fractional design for a study with 6 attributes and 3 levels for each attribute
would require only 33 or 27 choice decisions. This design provides a significant savings in time and
research costs yet still allows the estimation of all main effects and first-order interactive effects.
Several software packages are available for construction of optimized fractional factorial experimental
designs such as the SAS "Factex" procedure (SAS Institute).
SECTION 3. ATTRIBUTES AND TRADEOFFS FOR EVERGLADES
RESTORATION
3.1 Endpoints, Attributes and Ecosystem Restoration
Restoration of ecosystems presents one of the most difficult challenges in contemporary science and
environmental decision-making. Numerous technical questions arise over methods and procedures to
improve the ecological health of ecosystems that have been impacted by anthropogenic stresses.
Moreover, there is considerable uncertainty about what the recovery path of these systems will be in
response to existing, and potentially new, stressors (Cairns, 1988).
Equally, if not more, vexing is the problem of deciding what the objectives of restoration will be.
Several alternative points of view may exist about the services of the ecosystem that are most important
and what constitutes a 'restored' ecosystem (Bratton, 1992). The decision process to deal with these
problems requires natural science models of ecosystem processes and responses as well as social
science research to identify the importance of different ecosystem services to society and the public's
preferences for different levels of restoration (Milon et al 1997).
One way to consider the goals of ecosystem restoration is to describe ecological endpoints. The
endpoints represent characteristics of the ecosystem that, if changed, would indicate a change in the
health of the ecosystem (Harwell and Long 1992; Suter and Barnthouse 1993). These endpoints are not
fixed points but reflect multiple possible states of an ecosystem that can be observed and monitored
through one or more performance measures. In this sense, ecological endpoints are comparable to the
concept of attributes in multiattribute utility theory. One possible ecological endpoint such as
population levels of a species within an ecosystem can be conceptually related to the attributes of a
consumer product such as a car's safety or performance. While some may question the comparison of
nature with manufactured products, the decision of what and how much ecological restoration is
desirable is ultimately a social decision. And, the public's ability to understand ecosystem attributes
as well as the meaning of changes in these attributes can play a major role in determining the type and
level of restoration.
While there are many ways to characterize ecological endpoints for ecosystem restoration (Bratton
1992; Cairns 1988; Harwell and Long 1992), two broad classifications are "structural" endpoints and
"functional" endpoints. Structural endpoints focus on the number and diversity of individual species,
and may include keystone species or other distinctions to differentiate the species (Paine 1980,
Westman 1985, Wilson 1992). For the Everglades/South Florida ecosystem, these species indicators
might include bird populations, fish populations, and/or populations of endangered species such as the
Snail Kite and Wood Stork. On the other hand, functional endpoints emphasize broad ecosystem
processes and dynamic properties (Holling 1987, Westman 1985, 1991). In an uplands/wetlands
ecotone such as the Everglades, these endpoints might include hydroperiod levels and timing, salinity
fluctuations, spatial extent, and fire frequency (Harwell and Long 1992, Holling et al. 1994). While the
exact linkages between these alternative endpoint classifications are not well known, in general the two
sets of endpoints represent different dimensions of an ecosystem. Ecosystem management based on
one set of endpoints may lead to different restoration decisions than for the alternate endpoints (May
1973, McNaughton 1988, Westman 1985, 1991).
An additional consideration in this discussion of defining attributes to represent different types and
levels of ecosystem restoration is the distinction between endpoints and the means to achieve them.
Changing hydrological conditions may require a number of engineering projects and/or management
regulations. The exact combination of these projects and regulations is not directly related to the
consideration ofendpoints. While decisions about these projects and regulations do influence the cost
and even the technical feasibility of various endpoints, the details of these relationships are usually) not
part of the process of describing attributes in a multiattribute study.
3.1.1 Public Perceptions and Focus Groups
To help understand how the general public perceived the Everglades and the idea of ecosystem
restoration of this region, focus group sessions were conducted in Miami and West Palm Beach. Both
sessions were administered by a private marketing firm using a trained focus group facilitator (Rife
Market Research, Inc. 1998). Participants were recruited by the firm to represent a diverse cross-section
of the respective communities. As part of the sessions, participants were asked to explain what they
thought of when someone mentioned the term 'Everglades' and what personal experiences they had in
visiting places they identified with the Everglades. To facilitate this exercise, participants were asked
to draw a picture to show how they would describe the Everglades to someone who had never been to
Florida. They also were asked to describe what they thought of when someone talked about the concept
of Everglades 'restoration.'
The focus groups were helpful in establishing several points to consider about the selection of attributes
to represent the Everglades ecosystem. First, the participants' knowledge of the region varied from
those who knew little more than that the Everglades were located in South Florida (but not quite sure
where) to others who had closely followed events surrounding the Everglades over the years and had
personal recreational experiences (fishing, hunting, etc.) throughout the Everglades region. This variety
of backgrounds implied that the geographic dimensions of the problem and any attribute descriptions
used in a survey of the general public would need to be simple and clearly defined to provide a common
basis. Second, some participants had relatively strong opinions. A few thought of the Everglades as
little more than a "mosquito infested swamp" while others talked about the unique flora and fauna and
endangered species. Third, the picture drawing exercise revealed that most participants thought of the
Everglades in terms of both water and wildlife. Most drew pictures of relatively flat, watery landscapes
with different animals such as alligators, birds, and snakes. Very few included open prairies or
vegetation other than trees (mostly palms) in their drawings. Finally, some participants expressed an
understanding of the linkages between water levels in different areas of the Everglades and various
wildlife species. In general most participants were only loosely aware of this linkage. These
differences in knowledge once again indicated that some basic level of information about the
functioning of the ecosystem would be necessary to provide a common basis for the general public to
understand how the water management system was related to the ecology of the region.
Finally, the focus groups revealed that the term "restoration" was confusing and often misunderstood.
Many focus group participants expressed concerns that they thought this meant areas that had been
developed would be converted to wetlands. Others thought it would not be possible to "restore" a
system that had been altered so extensively as South Florida. Still others associated restoration with
the controversial constitutional amendment campaign to impose a tax on sugar produced in the
Everglades Agricultural Area. These participants thought that water quality was the principal problem
in the Everglades and the system would be healthy if this problem was corrected. These varied
perceptions of the term restoration indicated that this was not a useful word to describe changes in the
system and might potentially bias the results of a public survey.
3.2 Functional/Hydrological Attributes and Levels
To develop a set of functional attributes to represent different hydrological endpoints for
Everglades/South Florida restoration in this study, a number of planning documents from the Restudy
process were considered. For example, the Reconnaissance Report (U.S. Army Corps of Engineers
1994) described the use of different hydrological and ecological performance measures to evaluate the
effects of alternative restoration plans. These plans were combinations of engineering structures and
drainage system design modifications to the existing C&SFP infrastructure.
One of the most important hydrological performance measures was a comparison between the effects
of a plan as measured by the South Florida Water Management Model (SFWMM) and the Natural
Systems Model (NSM). The SFWMM is a mathematical engineering model that represents current
hydrological conditions throughout the area of South Florida from Lake Okeechobee south to Florida
Bay (see Figure 1). The model includes existing engineering structures such as canals, pump stations,
well fields, etc. and can be modified to represent new structures or the removal of existing structures
(U.S. Army Corps of Engineers 1999, Appendix B). The NSM is also a mathematical engineering
model for a comparable area that was developed to depict hydrological conditions of "predrainage"
South Florida based on rainfall patterns from 1965 to 1995. The NSM includes none of the existing
engineering structures. Estimates of overland flow rates and water levels that would have occurred in
a predrainage Everglades are based on assumptions about topography, rainfall, and evapotranspiration
(Fennema et al. 1994). Since neither accurate hydrological or rainfall data exist for earlier periods, there
is no way to know with any degree of precision how well the NSM actually represents predrainage
conditions (U.S. Army Corps of Engineers 1999, Summary pp. xv). Also, since the SFWMM and the
NSM only consider hydrological conditions, no information is provided about predrainage or current
ecological conditions in the region (Fennema et al. 1994).
Performance measures based on a comparison between existing/projected conditions and "historical"
conditions provide a convenient way to describe ecosystem restoration endpoints and are consistent
with general concepts of ecosystem restoration (Bratton 1992). Moreover, this type of comparison
provides a strong linkage between attributes that can be understood by the public and the hydrological
performance measures used in the Restudy planning process. The difficulty with the use of this type
of information in a multiattribute study is knowing what existing levels of the attribute might be and
how these might change under possible restoration scenarios. Fortunately, the South Florida Water
Management District had conducted studies to characterize hydrological conditions in key subregions
of South Florida as part of the Lower East Coast water supply planning process (South Florida Water
Management District 1997). This information can be used to characterize baseline conditions with
existing (1990) hydrological conditions and expected near-term (2010) conditions given expected water
demands and water system infrastructure. Unfortunately, since the Restudy process was developing
and evaluating alternative restoration plans up through early 1999 (U.S. Army Corps of Engineers
1999), there was no way to know with any precision how future (beyond 2010) hydrological conditions
resulting from these plans might change in relation to historical conditions (as represented by the NSM).
Therefore, to represent possible future hydrological conditions, hydrological attribute descriptions were
developed and three levels for each attribute were selected based on available information about near-
term and possible future conditions. The hydrological attributes were developed for three subregions:
Lake Okeechobee, the Water Conservation Areas, and Everglades National Park including Florida Bay
(Figure 1). These three subregions were selected because the current management system had identi field
them as distinct areas within the overall Everglades/South Florida region. And, the Restudy planning
process has identified hydrological problems and restoration alternatives that were unique to each
subregion (U.S. Army Corps of Engineers 1994). Also, information was available about hydrological
conditions in these regions based on comparisons between the SF\MINI and the NSM (South Florida
Water Management District 1997).
Other areas that are part of the Everglades/ South Florida region such as the Caloosahatchee River and
estuary, the St. Lucie River and estuary, and the Big Cypress National Preserve were not included for
two reasons. First, relatively little information was available about "predrainage" conditions in these
areas. Therefore, planning objectives have not been based on the same types of comparisons between
the SFWMM and the NSM that were used for the three subregions selected for this study. Second, the
addition of more areas into the multiattribute problem would have significantly increased the dimension
of the tradeoffs to be evaluated and required increases in the number ofpairwise choices made by each
respondent and/or increases in sample size (see Section 2.4). The relative loss of comprehensi\ eness
in this study relative to the broader Restudy planning process was necessary given the available
information and budget.
The actual hydrological attribute descriptions for the three subregions along with the levels selected
for each attribute are presented in Table 3-1. The attribute for each subregion was described in terms
of how too much or too little water impacts the primary habitats that characterize each subregion. Then,
the role that a plan to change the South Florida water management system was introduced as a
mechanism to control water levels and fluctuations in each subregion to be similar to historic,
predrainage conditions. This description of the attribute allowed the concept of "ecosystem
restoration" to be introduced through changes in hydrological conditions without actually using the term
restoration.' Respondents could then consider changes in the level of each attribute based on their own
understanding and evaluation of the merits of historic, predrainage conditions. To provide a common
reference point for all respondents about the role of water levels and timing in the Everglades
ecosystem, an informational video about the Everglades/South Florida region was presented prior to
the description of the attributes (see Section 4.1; Appendix D provides the full script for the video).
Note that no specific state or federal agency was mentioned in the attribute descriptions or in the video.
This was done to avoid any feelings, either positive or negative, individuals might have toward specific
agencies that would influence their judgment about the merits of the attribute alternatives.
As discussed in Section 4, the term 'restoration' may have caused a bias against the status
quo.
Table 3-1. Description of Attributes and Levels for the Hydrological Multiattribute Model
Attribute Descriptions Levels
Water Levels in Lake Okeechobee -- The water management system 60%, 75%, and 90% of the
controls the water levels and fluctuations in Lake Okeechobee. Too much water time, lake levels are
in the lake causes flooding of the shoreline and marsh areas. Too little water similar to historic,
causes these areas to dry out. Part of a plan to change the South Florida water predrainage conditions
management system could include ways to control the levels of the lake and
timing of fluctuations to be similar to historic, predrainage conditions. The
possible water level controls that could be included in the plan are:
Water Levels in Water Conservation Area -- The water management 50%, 75%, and 90% of the
system controls the water levels and fluctuations in Water Conservation Areas. time, water levels in the
Too much water in these Areas causes flooding of wetlands, upland areas and Water Conservation Area
tree islands. Too little water causes these areas to dry out. Part of a plan to are similar to historic,
change the South Florida water management system could include ways to predrainage conditions
control water levels and fluctuations in the Water Conservation Area to be
similar to historic, predrainage conditions. The possible water level controls that
could be included in the plan are:
Water Levels in Everglades National Park -- The water management 50%, 75%, and 90% of the
system controls the water levels and fluctuations in Everglades National Park and time, water levels in
the flow of fresh water to the Florida Bay. Too much water causes flooding of Everglades National Park
wetlands, upland areas and tree islands. Too little water causes these areas to dry are similar to historic,
out and increase the salinity in Florida Bay. Part of a plan to change the South predrainage conditions
Florida water management system could include ways to control the water levels
and fluctuations in the Park to be similar to historic, predrainage conditions. The
possible water level controls that could be included in the plan are:
The attribute levels presented in Table 3-1 denote the baseline and possible alternative levels of each
attribute. The baseline levels for each attribute are based on information presented in the Lower East
Coast Water Supply Plan (South Florida Water Management District 1997). As discussed above,
performance criteria in the Plan were based on the percent of time the current water levels and
hydroperiods (as represented by the SFWMM) would be comparable to predrainage conditions (as
represented by the NSM). Baselines represent the current system given expected service area demands
in the year 20102 and the completion of ongoing habitat restoration (e.g. Kissimmee River restoration)
and water quality improvement projects as mandated by the Everglades Forever Act (Florida Statutes
373.4592) of 1994 (South Florida Water Management District 1997, pp. III-3-III-5). Alternative levels
of the attributes were based on hypothetical changes in the water management system so that the
subregions would approximate predrainage water levels and hydroperiods 75 or 90 percent of the time.
These endpoints were selected because: a) they represent clear differences from the baseline for each
area, b) they represent 'partial' and 'full' changes in the current system relative to predrainage
conditions, and c) they avoid the notion that '100 percent' restoration of the ecosystem is possible given
the land use and engineering modifications that have occurred in the region over the past century.
These attribute levels were not intended to represent any specific restoration plan and, as discussed in
Section 3.1, the means by which the endpoints would be achieved were not considered as part of the
attribute description or in the multiattribute decision process.
2 It is important to note that this definition of the baseline for the hydrological attributes
differs from the concept of a baseline in the Restudy planning process. In the Restudy, the baseline
is service area demands projected for the year 2050.
3.3 Structural/Species Attributes and Levels
The problem of developing a set of structural attributes to represent animal and fish endpoints for
Everglades'South Florida restoration was complicated by a number of factors. First is the sheer
magnitude of the potential candidates. Due to the tropical and subtropical climate, the region includes
habitats for more than 350 bird species, 50 reptile species, 35 mammals, and more than 500 fish species
(U.S. Army Corps of Engineers 1994). Of these, more than 62 have been listed as endangered or
threatened under the Endangered Species Act. Second, information on historic and existing population
levels for many species is limited. Changes in populations of some species such as wading birds have
been well-documented (e.g. Ogden 1994) yet there is little consensus on how drainage over the past
century has impacted other populations such as deer, raptors and marine fishes. Third, one of the most
critical elements of uncertainty in the entire Everglades.'South Florida restoration process is the linkage
between hydrological changes and ecosystem response (U.S. Army Corps of Engineers 1999, pp. O-
13-0-17). This problem was recognized early in the Restudy planning process and led to efforts to
develop the Across Trophic Level Simulation System (ATLSS) to predict the effects of hydrological
changes on species and species groups (U.S. Army Corps of Engineers 1994). Unfortunately, output
from ATLSS was not available during the attribute development stage of this study.3 Therefore it was
necessary to rely on the limited published literature and discussions with various fish and wildlife
professionals from universities and agencies around Florida to develop a set of structural/species
attributes for this analysis. Table 3-2 presents the three species groupings that were selected as species
attributes and the levels associated with each attribute.
The wetland species group, which included wading birds and alligators, was selected since it
represented a broad, yet clearly defined, assemblage that most people would associate with wetlands
in South Florida.4 The dryland species group was similarly easy to associate with upland and prairie
habitats with different foraging and nesting requirements than wetland species. The third
group, Florida Bay and estuarine species, represented marine crustaceans and fishes that would most
likely be impacted by changes in hydrology in the lower Everglades region. Note that none of these
groupings explicitly identify endangered or threatened species such as the Florida Panther that many
people may associate with Everglades/South Florida fauna. This was done to avoid excessive concern
that an individual may place on a particular species in the overall consideration of each attribute.
3 As of April 1999, detailed results from ATLSS were available only for a few of the species
models; these models apparently were not a primary evaluation tool in the Restudy planning process
(U.S. Army Corps of Engineers 1999, Section 7).
4 Based in part of focus group discussions conducted as part of this study.
Table 3-2. Description and Results for the Species Multiattribute Model
Attribute Descriptions Levels
Wetland Species-- The water management system affects the food supplies and 20%, 50%, and 80% of
nesting areas available for animals and birds that live wetlands, marshes and tree historic, predrainage
islands. Part of a plan to change the South Florida water management system could population levels
include ways to influence wetland habitat quality to increase populations of birds such
as herons, spoonbills and wood storks and other animals such as alligators to levels
similar to historic, predrainage conditions. The population levels that are possible to
include in the plan are:
Dryland Species -- The water management system affects the amount of dry land 50%, 60%, and 70% of
and the food supplies available for animals and birds that live in pine forests, hardwood historic, predrainage
hammocks, and prairies. Part of a plan to change the South Florida water management population levels
system could include ways to influence dry land habitat quality to increase the
populations of species such as deer and racoons and birds such hawks, mockingbirds
and jays to levels similar to historic, predrainage conditions. The population levels that
are possible to include in the plan are:
Florida Bay Species -- The water management system affects the flow of fresh 60%, 75%, and 90% of
water into Florida Bay and fishes that are dependent on mangrove swamps and sea historic, predrainage
grasses in the Bay. Part of a plan to change the South Florida water management population levels
system could include ways to influence the habitat quality of Florida Bay to improve
the populations of pink shrimp and fishes such as mullet, sea trout and redfish to be
similar to historic, predrainage populations levels. The population levels that are
possible to include in the plan are:
Baseline and possible future levels for the wetland species attribute group in Table 3-2 were based on
Ogden's (1994) studies of wading bird nesting in the Everglades. The reductions in wading bird nesting
have been frequently cited as a primary indicator of Everglades ecosystem health (e.g. U.S. Army Corps
of Engineers 1999, Summary, pp. iii). Possible future population levels for this attribute, such as the
80 percent of historic, predrainage populations, were based on experts' opinions and the assumption
that land use and hydrological conditions in areas surrounding the remaining undeveloped parts of the
Everglades ecosystem would not significantly impair the recovery of wetland species populations. If
these changes, such as the loss of wetlands and mangroves around Biscayne Bay, do constrain wetland
species populations in the Everglades, the 80 percent level would be unrealistic.
In contrast to wading birds and wetland species, historical data on fish and marine life populations in
Florida Bay are limited. Available time series data do suggest some overall decline, particularly for
pink shrimp populations during periods of low rainfall when salinity levels in the Bay increase
(Bohnsack et al. 1994). While it is difficult to use these fishery-dependent data to compare to historic,
predrainage conditions, they do suggest that fish and marine life populations have not declined as
dramatically as wading bird populations. This is certainly the case for spiny lobster populations (as
measured by commercial harvests) which have remained relatively stable since the 1950s (Labiskey et
al. 1980; Milon et al. 1998).5 Moreover, the loss of mangrove and seagrass habitat in and around
Florida Bay has been significantly less than the loss of wetland habitat in other areas (U.S. Army Corps
of Engineers 1999, Sections 2 and 3). Therefore, levels for the Florida Bay species attribute in Table
5 The primary nursery area for juvenile spiny lobster in the Florida Keys is Florida Bay.
While the Bay provides important habitat during the life cycle, spiny lobster larvae in Florida Bay
may recruit from a variety of locations throughout the Caribbean (Ehrhard 1994; Hunt 1994).
3-2 have a higher baseline level than wetland species (60 percent of historic, predrainage conditions
compared to 20 percent) and it is assumed that overall Florida Bay species populations could recover
to 90 percent of historic levels.
The dryland species attribute in Table 3-2, which includes deer, racoons, and a variety of songbirds, is
the most difficult to characterize in terms of historic population levels. Large areas of pine forests and
hardwood hammocks along the eastern coastal ridge of South Florida have been con\ erted to farm land
and urban development. Yet, some prairie areas are much drier now than they were prior to drainage
in the 1940s and '50s (Light and Dineen 1994). Therefore, it was assumed that the current baseline
level for dryland species was 50 percent of historic levels. This reduction roughly matches the area of
upland habitat converted to other uses. The highest possible level for this attribute was set at 70 percent
of historic levels to reflect the loss of habitat.
In evaluating the species attributes in Table 3-2, it is important to note that simultaneous increases in
the levels of all three attributes are not likely under any restoration plan. Changes in hydrology and
habitat quality that favor wetland species in the Everglades 'South Florida region are not likely to also
benefit dryland species except in isolated areas. This aspect of the species attribute combinations is
different than the hydrology attributes which could all increase to higher levels relative to the baseline.
The significance of this difference will be more obvious when the multiattribute survey results are
applied to the evaluation of hypothetical restoration plans in Section 6.
3.4 Annual Cost, Farmland and Water Use Restriction Attributes
Restoration of hydrological and/or ecological conditions in the Everglades/South Florida region cannot
be considered in isolation from competing uses for water supplies and the costs of restoration plans.
While there is a broad range of alternative ways to describe these considerations, the attribute
descriptions and levels presented in Table 3-3 were selected for this study. These attributes were
considered to be important to the planning problem regardless of the way ecological endpoints were
described. Therefore, the attributes described in Table 3-3 were used in combination with both the
hydrological and species attributes presented above in Tables 3-1 and 3-2.
The annual cost per household attribute in Table 3-3 establishes that all Floridians could have to pay
for an Everglades/South Florida restoration plan. As with the hydrological and species attributes, these
costs are presented in the context of a plan to change the South Florida water management system. A
general utility tax was used as the 'payment vehicle' (Mitchell and Carson 1989) for a plan since this
type of tax is the most broadly based tax that could be assessed under Florida's tax system.6 A broad
based tax is important in this study because it helps to avoid respondent 'free-riding' in which
respondents express preferences for public programs knowing that they will not have to pay for these
programs. Moreover, since one of the objectives of this survey was to measure Floridians' preferences
and economic values for Everglades/South Florida restoration (Section 1.2), a broad based tax
applicable to all Floridians was necessary. Also, in order to estimate economic values from
multiattribute survey results, a measure of the tradeoff between personal costs and other attributes was
necessary (see Section 2.3).
6 A commonly used payment vehicle in public surveys is an income tax (Mitchell and Carson
1989). The Florida Constitution prohibits income taxes.
Table 3-3. Description of Attributes and Levels for Cost, Farmland and Water Use Restrictions
Attribute Descriptions Levels
Annual Cost per Household --All Florida residents pay utility taxes as part of their No change in utility taxes; $25
water, electric and telephone bills. Part of a plan to change the South Florida water increase per year; $50 increase per
management system could include additional taxes and all Floridians would pay for year.
these changes over the next 10 years. Proceeds from these taxes would go into a
special trust fund that would be used only to change to South Florida water
management system. Possible tax payments that could be included in the plan are:
Farmland -- Farmland acreage can be converted to water storage to increase the No change in farm land acreage;
flexibility of the water management system, increase the extent of natural areas, and 100,000 acre reduction; 200,000
reduce irrigation demand for water. Part of a plan to change South Florida water acre reduction.
management system could include reductions in existing farmland acreage in the
Everglades Agricultural Area in western portions of Broward and Dade counties that
are adjacent to the Water Conservation Areas and Everglades National Park. The
possible levels of farm land reductions that could be included in the plan are:
Restrictions on Water Use -- Changes in the water management system can affect 3 days per week outdoor use and
the availability of water for households in South Florida. The primary effect would 10% reduction in indoor use;
occur during years with low rainfall. These dry years occur, on average, in 1 out of 2 days per week outdoor use and
every 5 years. Possible levels of restrictions on outdoor and indoor water use that 25% reduction in indoor use;
could be included in the plan are: 1 day per week outdoor use and
40% reduction in indoor use.
The levels for the annual cost per household attribute in Table 3-3 were based on a no-cost ($0), low-
cost ($25), and high-cost ($50) hierarchy. All annual costs apply over ten years, the planning horizon
used in this survey. The full attribute descriptions provided to survey respondents emphasized that the
tax payments would apply over ten years (see Appendices B and C). The $25 and $50 levels were
selected because they are readily understandable intervals and they roughly correspond to initial
estimates of Floridians' per capital costs for Everglades restoration.7
The farmland attribute description and attribute levels presented in Table 3-3 reflect the possibility that
farmland could be converted to meet the objectives of a plan to change the South Florida water
management system. The specific areas where farmland acreage could be reduced were identified as
the Everglades Agricultural Area (see Figure 1-1) and western portions of Broward and Dade counties.
The acreages identified as possible levels for the attribute (0, 100,000 acres and 200,000 acres) reflect
plan alternatives included in the Reconnaissance Report (U.S. Army Corps of Engineers 1994) and
alternatives identified in discussions with Restudy planning staff.
The last attribute described in Table 3-3 addresses the possibility that a plan to change the South
Florida water management system could influence the severity of water restrictions on outdoor and
indoor household water uses in South Florida during dry years (about 1 of every 5 years). The
possibility of water use restrictions as part of an Everglades restoration plan is necessary since domestic
water users account for a majority of existing and projected water demands (South Florida Water
7 The initial Reconnaissance Report (U.S. Army Corps of Engineers, 1994) estimated
restoration costs could vary between $3 $6 billion depending on the extent of restoration. Cost-
sharing between Florida and the federal government for Everglades restoration under the Water
Resources Development Act of 1996 would be 50/50. Using an estimate of 5.7 million households
in Florida, annual payments of $25 or $50 per household for a ten year period would amount to
approximately $1.5 $3.0 billion.
Management District 1997). The attribute descriptions and levels were based on water shortage rules
previously established for South Florida households (South Florida Water Management District 1991).
The baseline level (outdoor uses restricted to 3 days per week, 10 percent reduction in indoor use)
corresponds to Phase I (moderate) water shortages. This level of restrictions is common in South
Florida, and other regions of Florida, during dry periods. The second level (outdoor uses restricted to
2 days per week, 25 percent reduction in indoor use) corresponds to Phase II (severe) water shortages.
The third level (outdoor uses restricted to 1 day per week, 40 percent reduction in indoor use) is
comparable to Phase III (extreme) water shortages.8
3.5 Attribute Combinations and Plan Alternatives
Combining either the three hydrological or three species attributes with the annual cost, farmland, and
water use restriction attributes results in six attributes and 36 (729) unique combinations of all levels
for the six attributes. As discussed in Section 2.4, this many attribute, le. el combinations is too large
for personal interview surveys. To reduce the number of attribute combinations, an optimized factorial
design was employed which could be used to evaluate all main and first-order interactive effects. This
resulted in 27 possible attribute/level combinations for either the hydrological or species representations
of plan attributes. Pretesting indicated that more than ten pairwise choices were too burdensome for
respondents. Therefore the 27 attribute combinations were split into two groups of seven pairwise
choices (2 groups x 7 pairwise choices equals 28 alternatives with one alternative repeated in each
group). With this design, each respondent only made 7 pairwise choices from the preselected attribute
combinations.
To evaluate how respondents' selection of preferred plans varied with the description of attributes, a
split sample design was utilized. This meant that a respondent only considered the hydrological
attribute combinations or the species attribute combinations. An example of a pairwise choice between
alternative plans using the hydrological attribute combinations is presented in Table 3-4. Plans A and
B each represent one of the 27 possible combinations of the six hydrological attribute and levels
selected with the factorial design. Respondents were informed that these plans were unique and none
of the plans had been already selected by public officials. After a set of practice exercises to familiarize
respondents with the pairwise choice process, one of the 7 sets of attribute combinations was described
to respondents. Respondents were instructed to carefully consider each plan combination and then
asked to select their preferred alternative. This process was repeated for each of the 7 attribute
combinations. The full set of hydrological attribute combinations used in the interviews is included in
Appendix B. More complete details on the interview process are provided in Section 4.
A comparable example of a pairwise choice between alternative plans with the species attribute
combinations is presented in Table 3-5. The same process of describing the attribute levels for each pair
of possible plans was followed in the interviews using the species attribute combinations. The full set
of species attribute combinations used in the interviews is included in Appendix C.
8 The Phase IV (critical) water shortage plan restrictions were not included in the possible
attribute levels.
3-10
Table 3-4. Example of Pairwise Choice for the Hydrological Multiattribute Model
Plan Component A B
Lake Okeechobee, Water Levels 60% of the time, lake levels and 60% of the time, lake levels
and Timing. timing are similar to historic, and timing are similar to
predrainage conditions historic, predrainage conditions
Everglades Water Conservation 50% of areas have water levels 50% of areas have water levels
Areas, Water Levels and Timing. and timing similar to historic, and timing similar to historic,
predrainage conditions predrainage conditions
Everglades National Park and 90% of the area has water levels 50% of the area has water
Florida Bay, Water Levels and and timing similar to historic, levels and timing similar to
Timing. predrainage conditions historic, predrainage conditions
Annual Cost Per Household Increased $25 per year No change
Restrictions on Household Outdoor Outdoor use limited to 2 days Outdoor use limited to 3 days
and Indoor Water Use. per week; indoor use reduced per week; indoor use reduced
25% 10%
Farm land (acres) in the Reduce farm land acreage by No change in farm land
Everglades Agricultural Area and 100,000 acres or 15% of acreage
Western Portions of Palm Beach, farmed area
Broward and Dade Counties.
Table 3-5. Example of Pairwise Choice for the Wildlife Multiattribute Model
Plan Component A B
Wetland Dependent Species Such 20% of historic, predrainage 20% of historic predrainage
as Wading Birds and Alligators population levels population levels
Dry Land Dependent Species 70% of historic, predrainage 50% of historic, predrainage
Such as Deer, Hawks and population levels population levels
Songbirds
Florida Bay Dependent Species 60% of historic, predrainage 60% of historic, predrainage
Such as Pink Shrimp, Mullet and population levels population levels
Sea Trout
Annual cost per household Utility taxes increased $25 per No change in utility taxes
year or $250 over 10 years
Restrictions on outdoor and In dry years, outdoor uses In dry years, outdoor uses
indoor household water use restricted to 2 days per week restricted to 3 days per week
and indoor uses reduced by and indoor uses reduced by 10%
25%
Farm land in the Everglades Reduce farm land acreage by No change in farm land acreage
Agricultural Area and western 100,000 acres (15% of farmed
portions of Palm Beach, Broward area)
and Dade counties
3-11
3.6 Other Attributes Not Included
The multiattribute survey process provides a flexible tool to simultaneously consider a number of
important attributes of a decision. The design of attributes for this survey was intended to incorporate
as many of the key issues and tradeoffs in the choice combinations as was feasible given limits on
respondent's time and ability to consider complex choices. But, other key issues being considered in
the context of Everglades/South Florida restoration were excluded.
First, water quality issues were not included in the attribute set due to the variety of existing and
potential problems in different subregions. This issue was particularly troubling since a highly
contentious and well-publicized campaign to amend the Florida Constitution had occurred in 1996 to
assess a 1 cent per pound tax on sugar produced in the Everglades Agricultural Area (EAA). Proceeds
from the proposed tax, which failed to get a state ide majority, would have been used to pay for
improved water quality for water flowing from the EAA into the Everglades. Initial focus groups
conducted in 1997 indicated that many people in South Florida considered water quality as the major
environmental problem in the Everglades.
Since the attribute combinations used to represent restoration plans in this survey focused only on
changes in the water management system to enhance hydrological (water levels and timing) or structural
(species) attributes, it was necessary to deal with the water quality issue if respondents asked about how
it would be addressed under the plans described in the pairwise choice process. If this issue was raised,
interviewers were instructed to inform respondents that water quality problems were being addressed
under a different plan that had already been established by the State of Florida.9 The plans each
respondent was asked consider would not interfere with this plan. Therefore respondents were asked
to not consider water quality issues in their selection of preferred alternatives.
A second issue that was not considered in the multiatiribute survey was the potential effects of structural
modifications in the C&SFP on local areas and communities in South Florida. Given the focus on
endpoints, the external effects of achieving these endpoints were not considered. These effects, such
as the impacts of water storage reservoirs or water tables on community development patterns, may be
a strong influence on the selection of projects to achieve restoration objectives. Due to the potentially
large number of local concerns and the lack of a specific restoration plan to consider, this issue was
excluded.
9 Although not explained to respondents, the plan being referred to is the Everglades Forever
Act (EFA) of 1994 which established water quality targets and timetables for the Everglades
Protection Area. It is assumed for the purposes of this survey that these targets and timetables will
be achieved and that future modifications to the C&SFP would conform to the requirements of the
EFA.
3-12
SECTION 4. SURVEY DESIGN AND RESPONDENT PROFILE
4.1 Description of the Interview Process
This survey was designed to be conducted through household personal interviews because of the large
amount of detailed information to be presented and gathered. Interviews consisted of five parts in the
following order: 1) explanation of the nature and purpose of the survey; 2) questions about the
respondents' general attitudes toward environmental and public policy issues; 3) presentation of an
informational video about the Everglades and water management in South Florida; 4) a paired choice
task to select preferred water management plans to change ecosystem attributes; and, 5) questions about
respondent's socio-economic characteristics.
Respondents who were selected to participate in the survey (see Section 4.2 below) were told that the
purpose was to determine public opinions about proposed changes in the water management system for
South Florida. The term "restoration" was not used in the interview process because focus groups
conducted in the early stages of the study indicated that the public had many different definitions of the
term and ideas about what it meant for South Florida. Also, there was concern that use of the term
"restoration" might impart a negative image to the status quo and thereby bias the plan choice process.
Questions about public policy issues and environmental attitudes included a series of questions about
spending priorities for State of Florida programs, questions about drinking water, and a series of
statements about the environment that the respondents were asked to either agree or disagree with.
Questions about socio-demographic characteristics included political party affiliation and voting
history, educational attainment, place of birth and years residency in Florida, age, household size,
racial/ethnic background, donations to environmental groups, and household income. The paired choice
task to select conjoint sets for water management plans in South Florida consisted of 3 practice choices
and 7 test choices for either the water management or wildlife management multiattribute sets described
in the previous section. The complete set of interview questions is provided in Appendix A.
The informational video, approximately 11 minutes in length, was a key part of the survey. The video
was designed to provide a common frame of reference about the issues surrounding the Everglades and
South Florida water management, and to enable respondents to clearly understand and answer survey
questions. The video was displayed either on the respondent's video player or on a portable player
provided by the interviewer. Major sections of the video included the history of the Everglades, its
hydrology and wildlife, alteration of the natural hydrologic patterns, the current water management
system, uses of water in the region, and modem environmental problems. The narrative script for the
video is presented in Appendix D.
Interviews were conducted by an independent market research firm (Rife Market Research, Inc. Miami,
FL) subcontracted by the University of Florida to implement the survey. Bilingual interviewers were
used in sampling areas of South Florida with a high percentage of Hispanic residents. An incentive of
$10 was paid to respondents as compensation for their time and cooperation to encourage participation
in the survey. Survey respondents were screened to be at least 18 years of age and a legal Florida
resident for at least 6 months. All respondents signed a consent form describing the conditions of the
interview and their rights in the interview process prior to the start of the formal interview.
The interviews averaged 57 minutes in duration and ranged from 20 minutes to nearly 2 hours. All
interview materials were organized in a set of notebooks that contained graphic representations of
response possibilities and were color-coded for the two multiattribute choice protocols (Appendices B
and C). Interviewers asked the respondent questions and recorded his or her answers on a separate
coding sheet. Interviewers also rated the respondents in terms of their seriousness about the survey,
level of distraction or attentiveness to the task, and understanding of the material, as well as recording
any general comments that were made about the survey.
4.2 Respondent Selection
A total of 480 randomly selected households were interviewed for the survey. Respondents were drawn
from selected Census tracts in 5 metropolitan areas: Miami (Dade County), West Palm Beach (Palm
Beach County), Ft. Meyers (Lee County), Tampa (Hillsborough Co.) and Orlando (Orange Co.). The
first three counties were chosen to represent the opinions of citizens most directly affected by the
restoration project since they were located within the South Florida Water Management District. The
other areas represented the opinions of urban residents outside South Florida. The margin of error was
4.5 percent for estimation of a binomial variable within the overall sample with a 95 percent level of
confidence and 10 percent within each county.
In each of the target sample areas, 96 households were sampled from 6 or 7 selected U.S. Census tracts.
The Census tracts in the study area were stratified by median household income and racial ethnic
composition. The three strata for median annual household income were: less than $25,000, $25,000
to $39,999, and $40,000 or greater. The two strata for racial ethnic composition were based on the
population of non-white and Hispanic minorities as a percentage of the county population: less than 25
percent and greater than or equal to 25 percent. These stratifications gave a total of 6 (3x2) strata
combinations. Each stratum was represented in the sample for each county by random selection from
the universe of Census tracts for that county.
The final selection of households for the survey was made by interviewers in the field following a
prescribed procedure to assure randomness. In each targeted Census tract, a typical street was selected
from two different Census blocks. A listing of all addresses on the street was made and numbered
consecutively. Then a random number table was consulted to randomly select 8 of the addresses as
primary target households for interviews and 4 as alternate households. At least two attempts were made
to contact the residents of the selected primary households at different times of day before accepting
an alternate address to avoid systematic bias in the selection based on daily work-activity patterns.
Additional sets of alternate households were generated as needed to complete the sample of 8 from each
block group. Approximately 10 percent of households contacted refused to participate.
4.3 Respondent Socio-Demographic Profiles
Socio-demographic characteristics of the respondents interviewed for the survey are summarized in
Tables 4.1 and 4.2. Comparable figures are also provided for the state population based on official
Census statistics and the Florida Annual Policy Survey (FAPS) conducted by Florida State University.
While this survey sample was not designed to be representative of the Florida population, the target
areas represent a major share of the total resident population in Florida.
In general, the characteristics of the survey sample closely matched the Florida population. Table 1
shows that respondent gender was 48.5 percent male and 51.5 percent female. Respondent ages were:
18 to 24 years 6.6 percent, 25 to 44 years 43.0 percent, 45 to 64 years 25.0 percent, and 65 years
or older 25.4 percent. The highest level of education attained by survey respondents was: primary
school (no high school diploma) 12.2 percent, high school 30.3 percent, some college 27.8 percent,
college graduates 19.6 percent, and post-graduate work 9.9 percent. This represented a slightly
higher share of sample respondents with advanced education, and a correspondingly lower share with
only primary education than the state population (Table 4.1). The number of respondents who indicated
they were not born in Florida (79 percent) was slightly greater than the state population (71 percent).
Also, respondents who had lived in Florida less than 5 years (31 percent) were somewhat more frequent
than the state population (17.5 percent), while those who had been residents over 21 years (30 percent)
were less frequent (37 percent).
The racial/ethnic composition of sample respondents was: white 82.9 percent, black 11.9 percent,
Hispanic (included in either white or black) 22.9 percent, and other 4.3 percent. Racial composition
of the sample in each county surveyed closely matched that of the county as a whole (Table 4.2).
Political party affiliation of respondents was: Republican 24.6 percent, Democrat 37.4 percent,
Independent or other party 16.7 percent, and no preference 21.3 percent. The proportion of
respondents who indicated that they voted in the last 3 years (62.6 percent) was slightly less than the
population (72.7 percent). In terms of donations made annually to environmental groups, 53.1 percent
of respondents did not make any, 33.3 percent donated less than $100, 9.0 percent donated $100 to
$500, and 1.5 percent reported donating more than $500. Annual household income was less than
$20,000 for 23.3 percent of respondents, $20,000 to $50,000 for 42.9 percent, over $50,000 for 23.1
percent; 10.6 percent refused to disclose household income levels (Table 4.1). The estimated median
household income for the survey sample in each county ranged from $25,000 in Hillsborough County
to $46,250 in Palm Beach County, but were similar to the county population in toto (Table 4.2). The
mean household size reported by respondents was 2.7 persons as compared to 2.46 for the state
population (Table 4.1).
As part of the interview, respondents were asked to indicate their support for state expenditures in 12
major program areas. These questions were identical to those contained in the FAPS in 1996 and
provide a useful benchmark for comparison. The results are summarized in Table 4.3.
Table 4-1. Socioeconomic Characteristics for the Survey Sample and Comparisons to the
Florida Population.
Socioeconomic Characteristic
Gender
Age
Educational Attainment
Born in Florida
Years Residency in Florida
Race/Ethnicity (may sum to greater
than 100 percent)
Male
Female
18-24
25-44
45-64
65 plus
No High School Diploma
High School Graduate
Some College
College Graduate
Post-Graduate Work
Yes
No
Less than 5
6 to 10
11 to 20
Over 21
White (Hispanic, nonhispanic)
Black (Hispanic, nonhispanic)
Hispanic
Other (Asian, American Indian
Pacific Islander)
Survey
Sample
(Percent)
48.5
51.5
6.6
43.0
25.0
25.4
12.2
30.3
27.8
19.6
9.9
21.0
79.0
31
16
23
30
82.9
11.9
22.9
4.3
Florida
Population*
(Percent)
47.8
52.2
10.6
38.3
26.9
24.2
25.6
31.0
26.0
12.0
6.3
28.8
71.2
17.5
16.9
26.9
37.1
85.9
12.5
12.8
1.8
Political Party Affiliation Republican 24.6 42.9
Democrat 37.4 48.4
Independent/Other 16.7 8.7
No Preference 21.3
Voted in Last 3 Years Yes 62.6 72.7
No 36.6 27.3
Amount of Donations None 53.1 45.8
Annually to Environmental Less than $100 33.3 32.6
Groups $100 to $500 9.0 12.4
More than $500 1.5 1.8
NA 3.1 7.5
Annual Household Income Less than $20,000 23.3 22.9
$20,000-$50,000 42.9 48.8
Over $50,000 23.1 28.4
Refused 10.6
Household Size (mean) 2.70 2.46
* Source for Florida Population values: Florida Annual Statistical Abstract 1997 (Floyd et al. 1998).
I
Table 4-2. Race/Ethnicity and Median Household Income for the Survey Sample, by County,
and Comparisons to the County Population.
Race/Ethnicity
(Percentage of Survey Sample or Population) Median Household
White, non-Hispanic Minority, other Income
Sample Population* Sample Population* Sample Population*
Dade 35.4 30.4 64.6 69.6 $40,625 $33,117
Lee 87.5 88.4 12.5 11.6 $32,273 $29,734
Palm Beach 77.1 79.4 22.9 20.6 $46,250 $36,013
Hillsborough 81.3 72.9 18.8 27.1 $25,000 $27,466
Orange 72.9 73.4 27.1 26.6 $34,444 $31,277
All Counties 70.8 58.5 29.2 41.5 $33,942 $31,628
* County population characteristics from 1990 US Census.
Public schools were rated as the top priority by 31.7 percent of respondents, followed by crime (18.5
percent), health care (11.9 percent), the environment (11.0 percent), the elderly (7.9 percent), low
income families (5.2 percent), endangered species (3.5 percent), highways (3.1 percent),
colleges/universities (2.5 percent), industrial development (2.3 percent), prisons (1.5 percent) and
tourism (0.6 percent). These rankings were very similar to those of the Florida population as measured
by the 1996 FAPS. For public schools, 79.4 percent of respondents indicated that public funding should
be increased in this area, 16.5 percent said it should remain the same and 3.8 percent wanted it
decreased. Again these results closely matched the Florida population. For programs for the
environment, a somewhat higher percentage of survey respondents (66.7 percent) desired an increase
in funding as compared to the state population (57.6 percent). A lower percentage wished it to remain
the same or to decrease. A similar pattern of results was found concerning priorities for funding
programs for endangered species. The percentage of survey respondents who supported increased
funding was also greater than the population for health care, highways, low income families, the elderly,
and tourism, but were lower for prisons and colleges/universities.
Respondents were also read statements about the environment and water supply issues and asked
whether they (strongly or somewhat) agreed or disagreed with each statement. Table 4.4 shows that
generally there was a very high level of concern and support expressed for statements about the
environment. Over 90 percent of respondents either "strongly agreed" or "somewhat agreed" with the
statements "the environment is easily upset" and "interference with the environment often causes
disaster." These results are highly consistent with other statewide surveys of Floridians attitudes about
the environment (Milon et al. 1998). Respondents also expressed strong support for water conservation
as in the statement "water conservation practices are a good idea." Conversely, nearly two thirds of
respondents disagreed with the statement "humans have the right to change the environment to suit their
needs." There was a somewhat lower level of agreement with statements regarding water supply: "water
will always be available for my community," "water will always be available for anybody who wishes
to move to Florida," or "water supply needs are a top priority." A majority of respondents disagreed
with the statement "protection from flooding is adequate in my community."
Table 4-3. Priorities for State of Florida Expenditures by Program Area by Survey Sample
Respondents and Comparisons to the Florida Population.
Funding Change Indicated
Program Area Sample or Top Rated (Percent of Respondents)
Priority Remain No
Population* Po increase Remain Decrease No
Popular (Percent) Increase Same De e Answer
Crime Sample 18.5 71.3 23.3 4.4 1.0
Population 16.4 70.5 21.8 4.5 3.2
Public schools
The Environment
Industrial development
Endangered Species
Health care
Colleges/universities
Highways
Low income families
Elderly
Prisons
Tourism
Sample
Population
Sample
Population
Sample
Population
Sample
Population
Sample
Population
Sample
Population
Sample
Population
Sample
Population
Sample
Population
Sample
Population
Sample
Population
31.7
35.4
11.0
8.6
2.3
4.3
3.5
na
11.9
7.6
2.5
3.9
3.1
3.7
5.2
6.0
79.4
78.5
66.7
57.6
26.3
24.4
57.7
51.1
65.8
55.0
50.0
55.3
43.3
35.5
48.5
36.5
64.2
49.7
31.5
33.7
23.3
20.8
16.5
15.0
29.2
33.0
45.2
43.0
32.3
36.3
28.5
32.1
42.3
34.3
45.6
52.9
40.0
40.9
31.5
38.8
42.3
39.5
47.1
53.6
3.8
3.0
3.1
6.3
26.5
27.4
8.3
9.3
5.4
8.6
7.1
5.8
10.2
9.4
10.8
16.2
4.0
6.6
24.8
21.5
28.5
23.5
* Population estimates are from the 1996 Florida Annual Policy Survey (Florida State University).
------- -~1-~ -- ~~-
-.
Table 4-4. Survey Respondent Attitudes about the Environment and Water Supply.
Percent of Respondents Choosing
Statement Strongly Agree Somewhat Somewhat Strongly No Answer
Agree Disagree Disagree
Water will always be available
for anyone who comes to my 38.3 32.5 18.8 6.3 4.2
community.
Water will always be available
for anyone who comes to 40.4 32.1 15.8 7.7 4.0
Florida.
Water supply needs are a top 38.5 27.3 20.4 12.9 0.8
prionty.
The environment is delicate and 63.1 28.3 5.6 1.5 1.5
easily upset.
Interference with the
environment often causes 60.8 32.3 4.2 1.9 0.8
disaster.
Protection from flooding in my 9.6 25.8 19.6 38.3 6.7
community is adequate.
Humans have the right to 6.0 21.0 24.0 46.9 2.1
change environment to suit
their needs.
Water conservation practices are 66.3 27.3 3.3 2.7 0.4
a good idea.
SECTION 5. STATISTICAL RESULTS
5.1 Statistical Modeling of Pairwise Choices
The pairwise choice (PC) approach to multiattribute utility analysis involves the elicitation of responses
to a pair of alternative choices. In an environmental choice setting such as described in Section 3, this
method uses attributes and levels to represent possible alternative states endpointss) of the environment.
Pairwise choices can be described in a utility maximizing framework based on random utility theory
(Louviere 1988; McFadden 1986). For each individual, the utility level of each alternative, I, in the pair
of choices may be represented by:
(1) U, = x, + Ei
where xi is a set of attributes for alternative choice I and E i s a random error term. If an individual
chooses alternative I consisting of a set of attributes over alternative j, the decision indicates that utility
to that individual from alternative I (U,) is greater than utility from j (U,).
In a pairwise multiattribute setting, a respondent compares the attributes of alternatives A and B and
selects the alternative that provides a higher level of utility or, in symbolic terms,
(2) U(X4)>U(XB)
where U() represents a respondent's utility function and XA, XB represent sets of attributes for
alternatives A and B. Utility can be decomposed into a systematic component, v() determined by the
attributes, and a random component, e such that
(3) U( = v(X) + e
The probability that a respondent chooses A rather than B depends on the probability that the difference
between the systematic component of A and B is greater than the difference between the random
components, such that the probability of choosing A is
(4) Pr(A) = Pr (D, > 6)
where D, = v(XA) v(XB) and 6 = c c
With data from a representative sample of individuals, the difference in the systematic component in
the utility functions (Dy) can be estimated using statistical techniques such as the conditional logit
model (Ben-Akiva and Lerman 1985; Greene 1997). The conditional logit model for data consisting
of choice-specific attributes can be expressed as the probability of observing response Y,:
exp(x, P,)
(5) Prob(Y, = j) =
J=j exp(x, p)
where the Ps are estimated coefficients (weights) for the attributes and exp represents the exponential
function.
Considering the hydrological attributes choices described in Section 3 (Tables 3-1 and 3-3), the
conditional logit model can be specified as:
(6) Do = Pf (+E,=,6 p3,( -X )
where Pi are the estimated coefficients and X, are the attributes of the alternatives in the choice sets with
X, the Lake Okeechobee attribute, X the Water Conservation Areas attribute, X, the Everglades
National Park attribute, X4 the Annual Cost attribute, X, the Water Use Restrictions attribute, and X6
the Farmland attribute.
Similarly, the species attributes choices (Tables 3-2 and 3-3) can be specified as:
(7) Dw= Po+~,,=1,6 P(XA -XB)
where X, is the Wetland Species attribute, X2 is the Dryland Species attribute, X, is the Estuarine
Species attribute, X4is the Annual Cost Attribute, X, is the Water Use Restrictions attribute, and X6is
the Farmland attribute.
In addition to identifying respondents' weighting of different attributes, it may also be desirable to
evaluate whether socioeconomic characteristics such as age, income, or geographic location influence
the attribute weights. The conditional logit model can be used to evaluate socioeconomic interactions
with the attributes variables (Swallow et al. 1994). A comparable specification for either the
hydrological attributes or species attributes models that includes socioeconomic variables interacting
with the attributes can be expressed as:
(8) DUr=Ei1,6 P,,(X AXB)+EI1,6 m =1,8 a-mS(X, _XB)
where X, denotes the ith attribute, fi are estimated coefficients for the main effects of each attribute,
ai, are estimated coefficients for interactive variables with attributes and eight socioeconomic
characteristics, S,,, that include political party affiliation, region of the state, donations to environmental
causes, gender, years living in Florida, income, and two ethnic groupings.
5.2 Sample Data
Using the survey design and split sample process described in Sections 3 and 4 resulted in 480
completed interviews which were divided evenly between the hydrological attributes and species
attributes formats. The 480 completed interviews provided 1,680 choices for each attribute set (7
choices times 240 respondents). Based upon the interviewer's appraisal that some respondents lacked
attentiveness and seriousness during the interview, 21 respondents were removed from the hydrological
attributes respondent sample and 6 respondents were removed from the species attributes sample. This
resulted in a total of 1,533 observations (choices) for the hydrological attributes model and 1,638
observations for the species attributes model. See Appendix E for complete data on interviewer
evaluations of respondents.
Table 5-1 reports descriptive statistics for the attributes and socioeconomic variables for the
hydrological and species sample groups used in the analysis. As discussed in detail in Section 3, the
attribute variables are expressed in terms of the percentage of historic levels. The mean, minimum, and
maximum values for the attributes reflect the statistical design.
Table 5-1. Variable Definitions and Summary Statistics for the Hydrological and Species Attribute Models.
Hydrological Attribute Models Species Attribute Models
Attribute/Variable Definition
Mean S. D. Min Max Mean S.D Min Max
Lake Okeechobee
Water Conservation Area
Everglades National Park
Wetland Species
Dryland Species
Estuarine Species
Annual Cost
Water Restriction 1
Water Restriction 2
Farmland
Political Party
Region
Donations
Gender
Years in Florida
Income
Ethnic 1
Ethnic 2
Percentage of historic level
Percentage of historic level
Percentage of historic level
Percentage of historic level
Percentage of historic level
Percentage of historic level
Annual increase in utilities tax for 10 years
Outdoor uses restricted to 2 days per week and
25% reduction in indoor consumption
Outdoor uses restricted to 1 day per week and
40% reduction in indoor consumption
Farm land acreage in EAA and South Florida
('000 acres)
1 if Republican, 0 all others
1 if Central Florida, 0 South Florida
1 if donated to environmental groups, 0
otherwise
1 if male, 0 female
Number of years as resident
1 to 9 by $10,000 increments (i.e. l=less than
$10,000, 2=$10,000 to $19,999, etc.)
1 if White-Hispanic or Black-Hispanic, 0
otherwise
1 if White-Non-Hispanic, 0 otherwise
0.74 0.12 0.6 0.9
0.72 0.17 0.5 0.9
0.71 0.17 0.5 0.9
25 20.02 0
0.36 0.48 0
0.32 0.47 0
99.84 80.19 0
0.25 0.44 0
0.41 0.49 0
0.42 0.49 0
0.48 0.50 0
20.24 14.09 1
4.07 2.06 1
0.14 0.35 0
0.74 0.44 0
0.49
0.60
0.76
25
0.36
0.25 0.2
0.08 0.5
0.12 0.6
20.03 0
0.48 0
1 0.32 0.46 0
99.88 80.19 0 200
0.27
0.39
0.49
0.48
18.71
4.26
0.44 0
0.49 0
0.50 0
0.50 0
12.69 1
2.08 1
0.12 0.33 0
0.73 0.45 0
Since the choice combinations were randomly distributed across the sample groups, the means for each
attribute are approximately the midpoint of the upper and lower levels.
For the purpose of this statistical analysis, the three levels for the water use restrictions attribute (see
Table 3-3) were converted into two dummy variables. The variable Water Restriction 1" was defined
as the second level of the attribute that would restrict outdoor uses to two days per week with a 25
percent reduction in indoor use during dry periods. The variable "Water Restriction 2" was defined as
the third level of the attribute that would restrict outdoor uses to one day per week with a 40 percent
reduction in indoor consumption. The first level of the attribute was defined as the base and included
in the estimated intercepts of the statistical models. As indicated in Table 5.1, Water Restriction 1 was
included in approximately one third of the choices in each sample group and Water Restriction 2 was
also included in about one third of the choices. These proportions are as expected given the sample
design.
Table 5-1 also presents the socioeconomic variables that were included in the statistical models.
Several of the socioeconomic variables (originally described in Table 4-1) were redefined as 0,1
variables as presented in Table 5-1. A comparison of the means for the socioeconomic variables in the
two sample groups shows that the two groups were very similar. This indicates that the alternative
attribute formats (hydrological and species) were randomly distributed across the total sample.
5.3 Statistical Results for the Multiattribute Models
The first multiattribute models estimated with the survey data were basic models that included only the
choice attributes as defined previously in this section in Equations (6) and (7). The basic model was
specified with and without an intercept term to test whether there was an order bias in respondents'
choices between alternative plans. The null hypothesis that the intercept (Po ) equals zero indicates
whether respondents based their choices on whether a plan was on the left- or right-hand side of the
page (see Table 3-5) when they were asked to choose between alternative plans. If the order of
presentation did influence respondents' decisions, then any information about the relative weighting
given to different attributes would be questionable. This is because multiattribute analysis assumes that
responses as based solely on the levels of the attributes.
The estimated coefficients and related statistics for the hydrological and species multiattribute models
are presented in Tables 5-2 and 5-3, respectively. The tables also include estimates of the marginal
willingness to pay (WTP) for changes in each attributes. The WTP for each attribute was calculated
with the formula:
(Pi (A,i) P (A,2)) /P,
where 1i is the estimated coefficient for attribute I, A,. is the base level of attribute I, Ai2 is the next level
of the attribute, and 3P is the estimated coefficient for the annual cost attribute. By dividing the
marginal change for each attribute by the cost attribute, the estimated \VTPs provide a measure of the
relative weighting assigned to each attribute. This weighting is a monetary measure of the utility
derived from each attribute. The WTP measure is appropriate for benefit-cost analysis of alternative
plans (Johnson and Desvousges 1997; Roe et al. 1996). Because the basic utility function specification
is additive, the marginal WTPs are constant (linear) for each attribute with the exception of the water
restrictions attribute that was decomposed into two separate effects. Given the description of the annual
cost attribute in the analysis (see Table 3-3), the annual WTPs would apply over a ten year period. A
comparable approach can be used to estimate WTP values in the multiattribute model with
socioeconomic interactions (Equation (8). The basic difference is that marginal values must be
calculated for each individual in the sample because the coefficients (13,) for each attribute vary across
socioeconomic characteristics.
5.3.1 Basic Hydrological Multiattribute Model Results
Table 5-2 presents statistical results for the hydrological multiattribute model. The coefficient for the
model estimated with an intercept term was statistically insignificant indicating that this subsample of
respondents' choices were not influenced by the order of presentation of the plan alternatives. Several
other coefficients, however, were statistically significant in the model estimated without an intercept.
The Lake Okeechobee and Water Conservation Areas attributes were both positive and statistically
significant indicating that respondents preferred plans that had higher levels of these attributes. The
Everglades National Park attribute was also positive but not statistically significant.
The Annual Cost and Farmland attributes were also statistically significant but had a negative sign
indicating that respondents preferred plans that had lower levels of these attributes. The water supply
restriction attributes provided interesting results. The Water Restriction 1 attribute was negative but
not statistically significant. The Water Restriction 2 attribute, however, was highly significant and the
estimated coefficient indicates that respondents placed a large negative weight on this attribute.
These statistical results indicate that higher levels of the hydrological attributes increased the likelihood
that respondents would select a plan that included higher levels of those attributes. On the other hand,
higher levels of the Annual Cost and Farmland attributes and the Water Restriction 2 attribute decreased
the likelihood of selecting a plan. The full effects of respondents' preferences for each attribute across
the sample are reflected in the marginal WTP values. The Water Conservation Areas attribute had the
highest positive WTP ($17.63) indicating that respondents derived the most satisfaction (utility) from
an increase (e.g. 50 percent to 75 percent) in this attribute. The Lake Okeechobee and Everglades
National Park attribute WTPs were also positive but relatively smaller.
The relatively large negative marginal WTP (-$37.10) associated with the Water Restriction 2 attribute
indicates the strong aversion respondents had to this component of a plan. On the other hand, the Water
Restriction 1 attribute had a relatively small negative WTP indicating that respondents would be willing
to experience some restrictions on water use for desirable changes in hydrological conditions in the
Everglades areas. The relatively large negative WTP associated with the Farmland attribute (-$16.39),
however, indicates that respondents would not prefer that these desirable changes in hydrological
conditions were accompanied by large reductions (100,000 or more acres) in farmland acreage in South
Florida.
Table 5-2. Coefficient Estimates for the Hydrological Multiattribute Models.
Model with Intercept
Attribute/Variable Coefficient Standard
Error
Model without Intercept
Coefficient Standard Marginal
Error Willingness to
Pay (WTP)
Intercept -0.1066 0.0756
Lake Okeechobee 0.3639 0.4724 0.7876* 0.3675 $9.68
Water Conservation Areas 0.9797* 0.2808 0.8606* 0.2666 $17.63
Everglades National Park 0.4445 0.3383 0.3419 0.3310 $7.01
Annual Cost -0.0136* 0.0031 -0.0122* 0.0029
Water Restriction 1 0.0354 0.1143 -0.0220 0.1063 -$1.80
Water Restriction 2 -0.4048* 0.1030 -0.4539* 0.0973 -$37.10
Farmland -0.0001* 0.0001 0.0001* 0.0001 -$16.39
Log Likelihood Function for the Model -999.36 -1000.36
Unrestricted Log Likelihood Function -1162.59 -1162.59
Chi-squared (7 degrees of freedom) 326.32* 324.46*
* Indicates significance at the 0.05 level
5.3.2 Basic Species Multiattribute Model Results
Table 5-3 presents statistical results for the species multiattribute model estimated with and without an
intercept term. The intercept in this model was also statistically insignificant indicating that this
subsample of respondents was not influenced by the order of presentation of the choice alternatives.
Most of the other coefficients, however, were statistically significant. The Wetland and Estuarine
Species attributes were positive and significant while the Dryland Species attribute was negative. The
marginal WTPs associated with these attributes indicates that respondents placed the most value on
higher levels of Estuarine Species ($27.34) and the least value on higher levels of Dryland Species (-
$29.87). Clearly respondents were expressing strong preferences for plans that would improve habitat
for wading birds and various marine fishes in the Everglades and Florida Bay.
As in the hydrological multiattribute models, higher levels of the Annual Cost and Farmland attributes
and the Water Restriction 2 attribute decreased the likelihood of selecting a plan. The coefficients for
all three of these attributes were negative and statistically significant. Moreover, it is also interesting
to note the similarity in the estimated WTPs for the Water Restriction 2 attribute in the hydrological
attributes model (Table 5-2) and the species attributes model. The relatively small difference (-$37.10
vs -$34.96) indicates that both subsamples responded in approximately the same way to the potential
of severe water restrictions despite the presence of other quite different attributes in the choice set.
Table 5-3. Coefficient Estimates for the Species Multiattribute Model.
Model with Intercept
Attribute/Variable Coefficient Standard
Error
Intercept
Wetland Species
Dryland Species
Estuarine Species
Annual Cost
Water Restriction 1
Water Restriction 2
Farmland
Log Likelihood Function for the Model
Unrestricted Log Likelihood Function
Chi-squared (7 degrees of freedom)
* Indicates significance at the 0.05 level
0.0930
0.7736*
-1.3915*
0.9589*
-0.0084*
-0.1289
-0.3815*
-0.0001*
0.0715
0.2239
0.6420
0.3594
0.0029
0.1075
0.0995
0.0001
-1078.12
-1135.38
114.39*
Model without Intercept
Coefficient Standard Marginal
Error Willingness to
Pay (WTP)
0.5991* 0.1780 $9.26
-1.1593* 0.6154 -$29.87
1.0606* 0.3518 $27.34
-0.0097*
-0.0869
-0.3391*
-0.0001*
0.0027
0.1027
0.0934
0.0001
-1078.96
-1135.37
112.82*
-$8.95
-$34.96
-$9.27
--------------------
5.4 Statistical Results for Models with Socioeconomic Characteristics Interactions
The hydrological multiattribute and species multiattribute models were also estimated using interactive
socioeconomic variables as described in Section 5.1 (Equation (8)). Tables 5.4 and 5.5 present results
for the hydrological and species multiattribute models, respectively. These coefficient estimates are
more difficult to interpret because of the large number of interactions. Moreover, WTP measures for
individual attribute interactions with socioeconomic characteristics can be misleading if not considered
along with other related interactions. Therefore, WTP values for the attributes are not presented in these
tables.
The statistical results indicate several significant interaction effects but the effects differ in each model.
In the hydrological multiattribute model, gender, region, income and ethnic variables had statistically
significant effects. On the other hand, only income and ethnic status had significant effects in the
species multiattribute model. While there were a number of insignificant interactions in both models,
the overall statistical results (i.e. log likelihood functions) indicate that the multiattribute models with
socioeconomic interactions provide important information to assess differences in respondent
preferences.
The primary importance of these results is that they provide the basis for disaggregating overall scoring
and ranking results for possible plans that can be constructed with alternative combinations of the
attribute. This application of the results will be discussed in greater detail in Section 6.
Table 5-4. Coefficient Estimates for the Hydrological Multiattribute Model with Socioeconomic
Characteristics Interactions.
Variable Coefficient Standard Error
Lake Okeechobee -0.0222 1.5309
Water Conservation Areas 1.0656 1.0944
Everglades National Park -3.6539* 1.1832
Annual Cost -0.0130* 0.0030
Water Restriction 1 -0.3111 0.4345
Water Restriction 2 -1.2725* 0.3992
Farmland -0.0001 0.0001
Political Party/Lake Okeechobee 0.1995 0.9244
Political Party/Water Conservation Areas 0.3207 0.6757
Political Party/Everglades National Park 0.5636 0.7257
Political Party/Water Restriction 1 -0.1225 0.2773
Political Party/Water Restriction 2 0.1598 0.2305
Political Party/Farmland 0.0001 0.0001
Region/Lake Okeechobee 0.4523 0.7936
Region/Water Conservation Areas -0.3920 0.5723
Region/Everglades National Park -1.9221 0.6081
Region/Water Restriction 1 0.1474 0.2304
Region/Water Restriction 2 -0.0674 0.1997
Region/Farmland -0.0001 0.0001
Donations/Lake Okeechobee 1.4508 0.7962
Donations/Water Conservation Areas -0.1894 0.5829
Donations/Everglades National Park -0.2004 0.6128
Donations/Water Restriction 1 0.2569 0.2298
Donations/Water Restriction 2 0.2534 0.2006
Donations/Farmland 0.0001 0.0001
Gender/Lake Okeechobee -1.6513* 0.7763
Gender/Water Conservation Areas -0.1156 0.5600
Gender/Everglades National Park 0.8863 0.6018
Gender/Water Restriction 1 0.4791* 0.2283
Gender/Water Restriction 2 0.5932* 0.1975
Gender/Farmland 0.0001 0.0001
Years in Florida/Lake Okeechobee -0.0103 0.0285
Years in Florida/Water Conservation Areas -0.0281 0.0201
Years in Florida/Everglades National Park -0.0013 0.0218
Years in Florida/Water Restriction 1 -0.0144 0.0083
Years in Florida/Water Restriction 2 -0.0115 0.0073
Years in Florida/Farmland 0.0001 0.0001
Income/Lake Okeechobee 0.2408 0.1982
Income/Water Conservation Areas 0.1756 0.1440
Income/Everglades National Park 0.5424* 0.1540
Income/Water Restriction 1 0.0703 0.0578
Income/Water Restriction 2 0.0741 0.0502
Income/Farmland -0.0001 0.0001
Ethnicl/Lake Okeechobee 0.7024 1.5441
Ethnic l/Water Conservation Areas 0.2611 1.1123
Ethnic 1/Everglades National Park 0.9575 1.1794
Ethnic /Water Restriction 1 -0.2855 0.4435
Ethnic 1/Water Restriction 2 0.4555 0.3942
Ethnic 1/Farmland 0.0001 0.0001
Ethnic2/Lake Okeechobee 0.0234 1.2438
Ethnic2/Water Conservation Areas -0.2963 0.8771
Ethnic2/Everglades National Park 2.6056* 0.9424
Ethnic2/Water Restriction 1 -0.1129 0.3604
Ethnic2/Water Restriction 2 0.2937 0.3170
Eihnic2 'Fannland 0.0001 0.0001
------------- ------------------------0 ---------------
Log Lt'klhood Funciion for the Moddel -951.91
Unrestricted Log Likelihood Function -1162.59
Chi-squared (55 degrees of freedom) 421.36*
Indicates significance at the 0.05 level
5-10
Table 5-5. Coefficient Estimates for the Species Multiattribute
Characteristics Interactions.
Model with Socioeconomic
Variable
Wetland Species
Dryland Species
Estuarine Species
Annual Cost
Water Restriction 1
Water Restriction 2
Farmland
Political Party/Wetland Species
Political Party/Dryland Species
Political Party/Estuarine Species
Political Party/Water Restriction 1
Political Party/Water Restriction 2
Political Party/Farmland
Region/Wetland Species
Region/Dryland Species
Region/Estuarine Species
Region/Water Restriction 1
Region/Water Restriction 2
Region/Farmland
Donations/Wetland Species
Donations/Dryland Species
Donations/Estuarine Species
Donations/Water Restriction 1
Donations/Water Restriction 2
Donations/Farmland
Gender/Wetland Species
Gender/Dryland Species
Gender/Estuarine Species
Gender/Water Restriction 1
Gender/Water Restriction 2
Gender/Farmland
Years in Florida/Wetland Species
Years in Florida/Dryland Species
Years in Florida/Estuarine Species
Years in Florida/Water Restriction I
Years in Florida/Water Restriction 2
Years in Florida/Farmland
Income/Wetland Species
Income/Dryland Species
Income/Estuarine Species
Income/Water Restriction 1
Income/Water Restriction 2
Income/Farmland
Ethnic 1/Wetland Species
Ethnic /Dryland Species
Ethnic 1/Estuarine Species
Ethnic I/Water Restriction 1
Ethnic 1/Water Restriction 2
Ethnic 1/Farmland
Ethnic2/Wetland Species
Ethnic2/Dryland Species
Ethnic2/Estuarine Species
Ethnic2/Water Restriction 1
Ethnic2/Water Restriction 2
Ethnic2/Farmland
Log Likelihood Function for the Model
Unrestricted Log Likelihood Function
Chi-squared (55 degrees of freedom)
* Indicates significance at the 0.05 level
------------------ --____
-1051.02
-1135.37
168.70*
5-11
Coefficient
-0.9084
-2.4959
0.2779
-0.0100*
0.0614
-0.6717*
0.0000
0.0813
-0.4097
-0.5412
0.1221
-0.2261
-0.0000
0.5020
1.2192
0.0371
0.1276
-0.0551
0.0000
0.1908
0.3639
-0.5563
0.1499
0.0693
0.0000
0.3046
-0.0713
-0.7121
-0.2746
-0.2066
0.0000
0.0075
0.0360
-0.0123
-0.0045
-0.0005
-0.0000
0.2460*
0.2445
0.3519*
-0.0124
0.0469
-0.0000
0.7429
-4.8581*
-0.3573
-0.2751
-0.4575
-0.0000
-0.2601
-0.5092
0.3932
-0.0062
0.4470
-0.0000
Standard Error
0.6952
2.0541
1.3650
0.0028
0.3881
0.3484
0.0000
0.4346
1.2692
0.8297
0.2490
0.2135
0.0000
0.3825
1.1279
0.7384
0.2199
0.1872
0.0000
0.3693
1.0930
0.7156
0.2120
0.1825
0.0000
0.3671
1.0824
0.7079
0.2106
0.1805
0.0000
0.0145
0.0429
0.0275
0.0084
0.0070
0.0000
0.0908
0.2700
0.1767
0.0523
0.0447
0.0000
0.8104
2.2277
1.4914
0.4444
0.4043
0.0000
0.5288
1.5440
1.0188
0.3003
0.2608
0.0000
SECTION 6. EVALUATION OF ALTERNATIVE RESTORATION PLANS
6.1 Voting, Ranking and Net Willingness to Pay Measures from Multiattribute Models
The paired comparison, multiattribute utility approach is one method to derive quantitative measures
of respondents' preferences for alternative levels of choice attributes based on statistically reliable
procedures. The results reported in Section 5 provide multiattribute utility functions for both
representative and socioeconomic-specific preferences using hydrological and species attributes of the
Everglades/South Florida ecosystem. These utility functions can be used to evaluate how respondents
would respond to hypothetical alternative combinations of the attributes based on the overall score
assigned to an alternative (see Section 2.3). The scoring information can be summarized in three ways:
1) as rankings of alternatives based on overall scores, 2) as the percent in favor of an alternative relative
to the status quo, and 3) as the net willingness to pay for an alternative. Because each summary measure
aggregates the scoring results in a different way, the ordering of alternatives may vary with these
measures.
A ranking evaluation of alternatives can be derived directly from the multiattribute utility function.
Using the basic model described in Equations (5) and (6) in Section 5, the net score (S) for the jth
alternative would be computed as:
sj = (6PxJ + Prx2j + r3X3j + PXr4j + P ,J + + P7X71 )
where the 3, s are the estimated coefficients and X1 is the level of the ith attribute in alternative j. This
net score can be calculated for as many hypothetical alternatives as are feasible with the attribute
combinations or are applicable to the overall problem setting. Once the net scores have been calculated
for all alternatives, the alternatives can then be arrayed in order of the net score to provide a unique,
ordinal ranking. The top ranked alternative would have the highest overall utility to respondents. This
same computational approach could be used with the socioeconomic interactive variable models
described in Equation (8) in Section 5 to provide rankings of alternatives according to respondents'
location (e.g. South Florida vs Central Florida) or other socioeconomic characteristics.
The net scores for each alternative could also be used as measures of whether respondents would vote
in favor of an alternative if it were presented as a referendum. If the status quo (i.e. a plan with no
change from baseline attribute levels) has a net score defined as So, then a voter would vote in favor
of the jth alternative plan if S ) So or against the alternative if S, ( So. By adding up the votes of all
respondents, the percent in favor of the alternative can be calculated. Note that this calculation only
requires a comparison of two net scores for each individual. Therefore two alternatives with similar
net scores may have the same percent in favor relative to the status quo.
The third way that the scoring information can be used is to calculate net willingness to pay (WTP) for
each alternative. The net WTP is a measure of the economic benefits (value) each respondent would
derive from all attributes included in the alternative. Following the basic approach to calculate the
marginal WTP for each attribute described in Section 5.3, the net WTP can be defined as:
Net WTPj = [((PX AX1,)/ (()/P)+ A pX20)/ 4) + ((p3 p/,3)/ ) + ((35XJ
PXo0)//4) + ((I6X6j J6X60)/l4) + ((y7X7j 7T70 )//14)) -X4j ]
where the f, s are the estimated coefficients, X, is the level of the ith attribute in alternative j, Xi is the
level of the ith attribute in the status quo (baseline) alternative, and 4, is the coefficient for the annual
cost attribute. Dividing the net score for the jth alternative (relative to the baseline) by the annual cost
coefficient converts the net score (measured in units of utility) into a money metric of individual
benefits (Johnson and Desvousges 1997; Roe et al. 1996). In contrast to the other two evaluation
measures, the net WTP measure provides some indication of the intensity of preferences. WTP
measures can be used in benefit-cost analysis of alternative plans by aggregating across the sample and
extrapolating the sample results to the population.
6.2 Overall Evaluation of Alternative Restoration Plans
6.2.1 Evaluation with the Hydrological Multiattribute Model
A set of alternative Everglades/South Florida ecosystem restoration plans were constructed with
different levels of the attributes in the hydrological multiattribute model.'0 These plans were designed
to represent a variety of possible scenarios: full and partial restoration with no direct costs to Floridians,
full and partial restoration with varying levels of costs to Floridians, and the status quo. Each plan was
then evaluated for the average survey respondent using the three measures described in the previous
section. Not all of these plans are intended to be technically or administratively feasible. For example,
any restoration plan conducted under the Water Resources Development Act of 1996 (see Section 1)
would require 50/50 cost-sharing between federal and state government. While a restoration plan
scenario involving no direct costs to Floridians is not likely, the no cost scenarios are useful as
indicators of Floridians' preferences for Everglades/South Florida ecosystem restoration. The WTPs
for the no cost scenarios can be interpreted as measures of the maximum benefits the average
respondent would receive from restoration.
Descriptions of 13 alternative restoration plans, the attribute levels included in each plan, and results
for the plan with the three evaluation measures are presented in Table 6-1. The baseline plan was
represented as the status quo levels of the three hydrological attributes. Since no changes in water
management attributes occur under the status quo, the annual cost, farmland and water restriction
attributes would not change from their baseline levels." The baseline ranked 8th overall. Since the
'percent in favor' and net WTP measures are calculated relative to the baseline, there is no result for
the baseline with these measures.
10 The reader should note that none of the plans were intended to fully represent the
components being considered as part of the Everglades/South Florida Restudy. Also, the definition
of the status quo (baseline) differs from that used in the Restudy. For a further discussion of the
differences between this evaluation and the Restudy process, see Section 3.
The annual cost attribute would be $0 since no additional money would be paid by
Floridians to finance restoration plans; the farmland attribute would be 0 since no farmland would
be converted for water management purposes; and water use would be restricted to 3 days per week
w ith 10 percent reductions in indoor use during dry years. These are essentially the 2010 conditions
as described in the Lower East Coast 11'ater Supply Plan (South Florida Water Management District
1997).
Table 6-1 Evaluation of Selected Restoration Plans with the Hydrological Multiattribute Model.
Net
Plan Description Percent in Ranking Willingness
Favor to Pay
HI-Baseline (no change) Hydrology.
Lake Okeechobee: 60% Costs: 0 NA 8 NA
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 50% Water Restriction: 3 days/ 0%
H2-Partial Everglades Park Restoration.
Lake Okeechobee: 60% Costs: 0 54.3 6 $7.01
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 75% Water Restriction: 3 days/10%
H3-Full Everglades Park Restoration.
Lake Okeechobee: 60% Costs: 0 54.3 4 $11.21
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 90% Water Restriction: 3 days/l0%
H4-Partial Hydrologic Restoration without Costs.
Lake Okeechobee: 75% Costs: 0 71.7 2 $34.32
Water Conservation: 75% Farmland Reduction: 0
Everglades National Park: 75% Water Restriction: 3 days/I 0%
H5-Full Hydrologic Restoration without Costs.
Lake Okeechobee: 90% Costs: 0 71.7 1 $58.79
Water Conservation: 90% Farmland Reduction: 0
Everglades National Park: 90% Water Restriction: 3 days/1 0%
H6-Partial Everglades Park Restoration and Water
Conservation Area with no costs, minimum restrictions.
Lake Okeechobee: 60% Costs: 0 54.3 7 $6.45
Water Conservation: 75% Farmland Reduction: 100,000 acres
Everglades National Park: 75% Water Restriction: 2 days/25%
H7-Partial Hydrologic Restoration with Minimized Taxes.
Lake Okeechobee: 75% Costs: $25 54.3 5 -$8.87
Water Conservation: 75% Farmland Reduction: 100,000 acres
Everglades National Park: 75% Water Restriction: 2 days/25%
H8-Partial Hydrologic Restoration, No Water Supply
Restrictions.
Lake Okeechobee: 75% Costs: $25 44.7 9 -$7.07
Water Conservation: 75% Farmland Reduction: 100,000 acres
Everglades National Park: 75% Water Restriction: 3 days/10%
H9-Partial Hydrologic Restoration with MinimizedCosts
Lake Okeechobee: 75% Costs: $25 44.3 10 $9.32
Water Conservation: 75% Farmland Reduction: 100,000 acres
Everglades National Park: 75% Water Restriction: 3 days/10%
H10-Full Hydrologic Restoration with Minimized Costs.
Lake Okeechobee: 90% Costs: $25 54.3 3 $15.60
Water Conservation: 90% Farmland Reduction: 100,000 acres
Everglades National Park: 90% Water Restriction: 2 days/25%
Table 6-1 Continued.
Net
Plan Description Percent in Ranking Willingness
Favor to Pay
H11-Full Everglades Park Restoration and Water
Conservation Area
Lake Okeechobee: 60% Costs: $25 29.7 12 -$55.45
Water Conservation: 90% Farmland Reduction: 200,000 acres
Everglades National Park: 90% Water Restriction: 1 day/40%
H12-Full Hydrologic Restoration with Costs, No Water
Supply Restrictions
Lake Okeechobee: 90% Costs: $50 41.6 11 -$23.99
Water Conservation: 90% Farmland Reduction: 200,000 acres
Everglades National Park: 90% Water Restriction: 3 days/10%
H13-Full Hydrologic Restoration with Full Costs
Lake Okeechobee: 90% Costs: $50 31.1 13 -$61.09
Water Conservation: 90% Farmland Reduction: 200,000 acres
E erglades National Park: 90% Water Restriction: 1 day/40%
NA=Not Applicable
Table 6-1 shows that the plan which ranked first overall was H1 'Full Hydrological Restoration
Without Costs' with the highest level of restoration for each hydrological area (Lake Okeechobee,
Water Conservation Areas, and Everglades National Park) and baseline levels for the other attributes.
Nearly 3/4ths of the respondents would vote in favor of this plan and it had the highest WTP at $58.79
per respondent household. Because the annual cost attribute description encompasses a ten-year time
period, the cumulative WTP per household for this plan would be $587.90.
By comparison, Table 6-1 shows that the lowest ranked alternative was H13 'Full Hydrologic
Restoration with Full Costs' which would include the highest levels of restoration for the hydrological
attributes and the highest levels of the annual cost, farmland, and water restriction attributes. Less than
1/3 of respondents would vote in favor of this plan and the net WTP for the plan is $61.09. Because
this plan is ranked lower than the baseline and the net WTP is negative, it would result in most
respondents being worse off than the status quo.
The sensitivity of the average respondent to the cost components of a plan is clearly evident in the
difference in the evaluation results between H10 and H12. Both plans include full hydrological
restoration of the h1 drological areas, but H12 has annual costs of $50, a reduction in farmland of
200,000 acres, and baseline water restrictions. H10 has annual costs of $25, a reduction in farmland
of 100,000 acres, and water restrictions of 2 days outdoor use/25 percent indoor use reductions. Yet
H10 is the 3rd ranked alternative with a majority in favor and a net WTP of $15.60 per household while
H12 is the 11th ranked alternative with less than a majority and a net WTP of -$23.99. This result
suggests that respondents would be willing to trade-offsome reductions in water availability in lieu of
higher direct annual costs and farmland reductions if full hydrological restoration were achieved.
The importance of the cost components of a plan is also demonstrated in other results in Table 6-1. For
example, all of the full or partial restoration plans without costs (H2 H5) are ranked higher than the
baseline, have majorities in favor, and positive net WTPs. When direct costs are minimized such as in
H6 and H7, even partial hydrological restoration plans are favored by a majority of respondents and
have positive net WTPs.
In summary, these evaluations of alternative hydrological restoration plans with the multiattribute
model results indicate that the average respondent has a strong desire for restoration of the
Everglades/South Florida ecosystem. But, these preferences to restore the ecosystem are tempered by
other preferences to minimize the direct costs of restoration to households and to avoid severe water
use restrictions.
6.2.2 Evaluation with the Species Multiattribute Model
A comparable set of alternative Everglades/South Florida restoration plan scenarios were constructed
for the attributes in the species multiattribute model. As with the hydrological model scenarios, these
plans were designed to represent a variety of possible attribute combinations and were not intended to
represent actual restoration plans or to be technically or administratively feasible. The variety of plan
scenarios helps to illustrate the preferences of the average survey respondent who was interviewed using
the species attribute pairwise choice process.
Descriptions of 11 alternative restoration plans, the attribute levels included in each plan, and results
for the plan with the three evaluation measures are presented in Table 6-2. The baseline plan was again
defined as the status quo levels of the three species attributes with all other attributes at their baseline
levels. The baseline ranked 7th of the 11 plans considered indicating some plans were preferred to the
baseline while others would leave respondents worse off.
The top ranked alternative in Table 6-2 was S3 'Full Wetland Wildlife Restoration Without Costs."
This plan had a very large majority in favor and a net WTP of $69.86 indicating significant potential
economic benefits would accrue from this restoration scenario. The similarity in this net WTP value
with the net WTP value for H1 'Full Hydrological Restoration Without Costs' suggests that
respondents were expressing a relatively consistent set of preferences regardless of the survey
instrument used in the interview.
The effects of the large negative weighting given to dryland species restoration in the species
multiattribute model (Table 5-3) are evident in the plan rankings in Table 6-2. Both partial and full
dryland species restoration plans without costs (S4 and S5) were ranked below the baseline scenario
and have negative net WTPs. Clearly the average respondent did not view restoring dryland species
populations as a desirable outcome.
As in the hydrological restoration scenario evaluations, however, the cost components of a plan had a
major impact on the overall evaluation. The full wetland wildlife restoration scenario S11 'Full
Wetland Wildlife Restoration With Full Costs' was the lowest ranked alternative and had a net WTP
of -$33.64. Other full and partial wetland wildlife restoration plans illustrate the sensitivity of the
average respondent to the cost components. But, the species model evaluations did not change as
dramatically in response to costs as in the hydrological model evaluations. Plans S6 and S9 in Table
6-2 would yield either full or partial wetland wildlife species restoration with annual costs at $25,
farmland reductions of 100,000 or less, and no severe water restrictions. These plans were ranked 3rd
and 4th, respectively, had majorities in favor, and had positive net WTPs. Even plan S10 with full
wetland wildlife species restoration and annual costs of $50 and farmland reductions of 200,000 acres
had a majority in favor and a positive, yet small, positive net WTP. Only when partial wetland species
restoration was coupled with annual costs of $25, farmland reductions of 100,000 acres, and water use
restrictions of 2 days/25 percent as in plan S8 did the percent in favor fall below a majority and the net
WTP become negative.
Table 6-2 Evaluation of Restoration Plans with the Species Multiattribute Model.
Percent Net
Plan Description in Favor Ranking Willingness
to Pay
S-1 Baseline (no change) Wildlife.
Wetland Species: 20% Costs: 0 NA 7 NA
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/10%
S-2 Partial Wetland Wildlife Restoration without Costs.
Wetland Species: 50% Costs: 0 92.7 2 $34.93
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 75% Water Restriction: 3 days/ 0%
S-3 Full Wetland Wildlife Restoration without Costs.
Wetland Species: 80% Costs: 0 92.7 1 $69.86
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 90% Water Restriction: 3 days/10%
S-4 Partial Dryland Wildlife Restoration without Costs.
Wetland Species: 20% Costs: 0 17.9 9 -11.95
Dryland Species: 60% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/10%
S-5 Full Dryland Wildlife Restoration without Costs.
Wetland Species: 20% Costs: 0 17.9 10 -23.90
Dryland Species: 70% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/10%
S-6 Partial Wetland Wildlife Restoration with Minimized Taxes.
Wetland Species: 50% Costs: $25 61.5 4 $9.93
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 75% Water Restriction: 3 days/10%
S-7 Partial Wetland Wildlife Restoration, No Water Use
Restrictions.
Wetland Species: 50% Costs: $25 52.1 6 $0.66
Dryland Species: 50% Farmland Reduction: 100,000 acres
Esuarine Species: 75% Water Restriction: 3 days/10%
S-8 Partial Wetland Wildlife Restoration with Costs.
Wetland Species: 50% Costs: $25 42.7 8 -$8.3
Dryland Species: 50% Farmland Reduction: 100,000 acres
Estuarine Species: 75% Water Restriction: 2 days/25%
S-9 Full Wetland Wildlife Restoration with Minimized Costs.
Wetland Species: 80% Costs: $25 67.9 3 $26.63
Dryland Species: 50% Farmland Reduction:100,000 acres
Estuarine Species: 90% Water Restriction: 2 days 250
S-10 Full Wetland Wildlife Restoration with Costs and
NoWater Use Restrictions.
Wetland Species: 80% Costs: $50 52.1 5 $1.32
Dryland Species: 50% Farmland Reduction:200,000 acres
Estuarine Species: 90% Water Restriction: 3 days/10%
S-11 Full Wetland Wildlife Restoration with Full Costs
Wetland Species: 80% Costs: $50 29.9 11 -33.64
Dryland Species: 50% Farmland Reduction:200,000 acres
Estuarine Species: 90% Water Restriction: 1 day/40%
In summary, the evaluations of alternative species restoration plans support the proposition that the
average respondent had strong desires for restoration of the Everglades/South Florida ecosystem.
Responses with the species multiattribute model were generally less sensitive to the cost components
and resulted in favorable evaluations of several wetland wildlife species restoration plans. Respondents,
however, showed no inclination to restore dryland species populations.
6.3 Effects of Socioeconomic Characteristics on Evaluations of Alternative Restoration Plans
6.3.1 Evaluations of Alternative Plans with the Hydrological Multiattribute Model by
Respondents' Location and Environmental Donation Status
The same set of 13 alternative restoration plans constructed for the hydrological multiattribute model
presented in Table 6-1 were evaluated using the socioeconomic characteristics interactive model
coefficients reported previously in Table 5-4. This analysis was conducted using the same three
evaluation measures. The interactive model allows the evaluation results to be disaggregated according
to specific socioeconomic characteristics of the respondents.
Evaluation results for the 13 hydrological restoration plans for respondents in South Florida and Central
Florida are reported in Table 6-3. The results reveal some rather striking differences in the preferences
of South vs Central Floridians. While both rank H5 'Full Hydrological Restoration Without Costs'
as the most preferred alternative, a large majority (87.6 percent) of South Floridians would vote in favor
of this plan while less than a majority (48.9 percent) of Central Floridians would vote in favor.
Moreover, South Floridians' net WTP for H5 is $86.42 as compared to a net WTP of $8.24 for Central
Floridians. In fact, only two plans (H4 and H5), which both include restoration with no costs to
Floridians, are preferred by Central Floridians to the status quo and have positive net WTPs. All the
other alternative plans evaluated for Central Florida respondents would be ranked lower than the
baseline and have negative net WTPs. For some plans these negative values were quite large (e.g.
-$131.65 for H13) indicating a strong aversion to any restoration plan that would impose costs on
Floridians.
On the other hand, the percent of South Floridians who would vote in favor of a number of restoration
plans and the positive net WTPs for these plans indicates strong preferences for these alternatives. The
only plans not preferred to the status quo and with negative net WTPs (i.e. H11 and H13) would impose
high direct costs and severe water use restrictions. These results suggest that most South Floridians
would be willing to incur significant personal costs and experience at least moderate water use
restrictions to achieve hydrological restoration of the Everglades/South Florida ecosystem.
A similar evaluation of the 13 hydrological plans is reported by respondent's environmental donation
status in Table 6-4. Prior research has shown that respondents who donated to environmental causes
in the past generally express higher levels of concern about the environment and are more supportive
of policies to protect and enhance the environment (Milon et al. 1998). It would be expected that
respondents who donated to environmental causes would be more likely to have stronger preferences
for environmental restoration plans. The results reported in Table 6-4 confirm these expectations.
Respondents who donated ranked a significantly larger number of alternative plans as superior to the
status quo and the percent voting in favor across all of the plans was consistently higher for the
donations group. Moreover, the net WTPs for the donations group in Table 6-4 were negative for only
two plans (HI and H13). Compared with the average respondent's net WTPs for the same plans
(Table 6-1), the donations group expressed consistently stronger preferences for Everglades/South
Florida ecosystem restoration.
The differences in preferences between respondents who donated to environmental groups compared
to those who did not are important given that approximately 50 percent of the sample respondents had
donated (see Table 4-1). This is very similar to the percentage reporting environmental donations in
other statewide surveys (e.g. Milon et al. 1998). The results suggest that Floridians who express high
levels of concern about the environment are likely to strongly support Everglades/South Florida
ecosystem restoration.
Table 6-3. Evaluation of Selected Restoration Plans with the Hydrological Multiattribute Model by Respondent Location.
Percent in Favor Ranking Net Willingness to Pay
Plan Description
escpion South Central South Central South Central
Florida Florida Florida Florida Florida Florida
H-1 Baseline (no change) Hydrology.
Lake Okeechobee: 60% Costs: 0 NA NA 11 3 NA NA
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 50% Water Restriction: 3 days/I 0%
H-2 Partial Everglades Park Restoration.
Lake Okeechobee: 60% Costs: 0 72.9 27.8 7 4 $21.55 -$18.10
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 75% Water Restriction: 3 days/10%
H-3 Full Everglades Park Restoration.
Lake Okeechobee: 60% Costs: 0 72.9 27.8 4 6 $34.47 -$28.96
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 90% Water Restriction: 3 days/10%
H-4 Partial Hydrologic Restoration without Costs.
Lake Okeechobee: 75% Costs: 0 87.6 48.9 2 2 $51.55 $2.37
Water Conservation: 75% Farmland Reduction: 0
Everglades National Park: 75% Water Restriction: 3 days/I0%
H-5 Full Hydrologic Restoration without Costs.
Lake Okeechobee: 90% Costs: 0 87.6 48.9 1 1 $86.42 $8.24
Water Conservation: 90% Farmland Reduction: 0
Everglades National Park: 90% Water Restriction: 3 days/10%
H-6 Partial Everglades Park Restoration and Water
Conservation Area with no costs, minimum restrictions
Lake Okeechobee: 60% Costs: 0 66.7 36.7 6 7 $21.61 -$31.37
Water Conservation: 75% Farmland Reduction: 100,000 acres
Everglades National Park: 75% Water Restriction: 2 days/25%
H- 7 Partial Hydrologic Restoration with Minimized Taxes.
Lake Okeechobee: 75% Costs: $25 72.9 27.8 5 5 $6.56 -$45.28
Water Conservation: 75% Farmland Reduction: 100,000 acres
Everglades National Park: 75% Water Restriction: 2 days/25%
Table 6-3 Continued.
Percent in Favor Ranking Net Willingness to
Plan Description Pay
South Central South Central South Central
Florida Florida Florida Florida Florida Florida
H-8 Partial Hydrologic Restoration, No Water Supply
Restrictions.
Lake Okeechobee: 75% Costs: $25 64.3 16.7 8 9 $14.29 -$43.26
Water Conservation: 75% Farmland Reduction 100,000 acres
Everglades National Park: 75% Water Restriction: 3 days/10%
H-9 Partial Hydrologic Restoration with Minimized Costs.
Lake Okeechobee: 75% Costs: $25 58.1 24.4 10 10 $26.55 -$22.63
Water Conservation: 75% Farmland Reduction: 0
Everglades National Park: 75% Water Restriction:3 day s.' I)" o
H-10 Full Hydrologic Restoration with Minimized Costs.
Lake Okeechobee: 90% Costs: $25 69.0 33.3 3 8 $41.43 -$39.41
Water Conservation: 90% Farmland Reduction: 100,000 acres
Everglades National Park: 90% Water Restriction: 2 days/25%
H-11 Full Everglades Park and Water Conservation Area
Restoration
Lake Okeechobee: 60% Costs: $25 45.7 6.7 12 12 -$16.97 -$128.83
Water Conservation: 90% Farmland Reduction: 200,000 acres
Everglades National Park: 90% Water Restriction: 1 day/40%
H-12 Full Hydrologic Restoration with Costs, No Water
SupplyRestrictions
Lake Okeechobee 90% Costs: $50 58.9 16.7 9 11 $11.90 -$83.01
Water Conservation: 90% Farmland Reduction: 200,000 acres
Everglades National Park: 90% Water Restriction: 3 days/I 0%
H-13 Full HydrologicRestoration with Full Costs
Lake Okeechobee: 90% Costs: $50 43.4 13.3 13 13 -$22.11 -$131.65
Water Conservation: 90% Farmland Reduction: 200,000 acres
Everglades National Park: 90% Water Restriction: I day/40%
NA=Not Applicable
6-10
Table 6-4. Evaluation of Selected Restoration Plans with the Hydrological Multiattribute Model by Respondent Past Donations.
Percent in Favor Ranking Net Willingness to Pay
Plan Description
DONATIONS DONATIONS DONATIONS
NO YES NO YES NO YES
H-1 Baseline (no change) Hydrology.
Lake Okeechobee: 60% Costs: 0 NA NA 3 11 NA NA
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 50% Water Restriction: 3 days/10%
H-2 Partial Everglades Park Restoration.
Lake Okeechobee: 60% Costs: 0 48.4 62.6 4 8 -$2.18 $15.71
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 75% Water Restriction: 3 days/10%
H-3 Full Everglades Park Restoration.
Lake Okeechobee: 60% Costs: 0 48.4 62.6 5 6 -$3.49 $25.15
Water Conservation: 50% Farmland Reduction: 0
Everglades National Park: 90% Water Restriction: 3 days/10%
H-4 Partial Hydrologic Restoration without Costs.
Lake Okeechobee: 75% Costs: 0 58.6 90.1 2 3 $16.29 $52.52
Water Conservation: 75% Farmland Reduction: 0
Everglades National Park: 75% Water Restriction: 3 days/I 0%
H-5 Full Hydrologic Restoration without Costs.
Lake Okeechobee: 90% Costs: 0 58.6 90.1 1 1 $27.35 $92.26
Water Conservation: 90% Farmland Reduction: 0
Everglades National Park: 90% Water Restriction: 3 days/10%
H-6 Partial Everglades Park Restoration and Water
Conservation Area with No Costs, Minimum Restrictions 40.6 73.6 7 4 $20.22 $6.51
Lake Okeechobee: 60% Costs: 0
Water Conservation: 75% Farmland Reduction: 100,000 acres
Everglades National Park: 75% Water Restriction: 2 days/25%
H-7 Partial Hydrologic Restoration with Minimized Taxes.
Lake Okeechobee: 75% Costs: $25 43.0 70.3 6 5 $41.99 $2.04
Water Conservation: 75% Farmland Reduction:100,000 acres
Everglades National Park: 75% Water Restriction: 2 days/25%
6-11
Table 6-4 Continued.
Percent in Favor Ranking Net Willingness to Pay
Plan Description
DONATIONS DONATIONS DONATIONS
NO YES NO YES NO YES
H-8 Partial Hydrologic Restoration, No Water Supply
Restrictions. 3 61.5 8 10 -$25.12 $12.82
Lake Okeechobee: 75% Costs: $25 2.8
Water Conservation: 75% Farmland Reduction 100,000 acres
Everglades National Park: 75% Water Restriction: 3 days/10%
H-9 Partial Hydrologic Restoration with Minimized Costs.
Lake Okeechobee: 75% Costs: $25 27.3 68.1 10 7 -$8.71 $27.52
Water Conservation: 75% Farmland Reduction: 0
Everglades National Park: 75% Water Restriction: 3 days/10%
H-10 Full Hydrologic Restoration with Minimized Costs.
Lake Okeechobee: 90% Costs: $25 75.8 9 2 -$30.93 $41.78
Water Conservation: 90% Farmland Reduction: 100,000 acres 39.1
Everglades National Park: 90% Water Restriction: 2 days/25%
H-11 Full Everglades Park and Water Conservation Area
Restoration. 20.3 42.9 12 13 -$90.81 -$23.75
Lake Okeechobee: 60% Costs: $25
Water Conservation: 90% Farmland Reduction: 200,000 acres
Everglades National Park: 90% Water Restriction: 1 day/40%
H-12 Full Hydrologic Restoration with Costs, No Water Supply
Restrictions. 28.1 60.4 11 9 -$55.47 $12.86
Lake Okeechobee: 90% Costs: $50
Water Conservation: 90% Farmland Reduction: 200,000 acres
E erglades National Park: 90% Water Restriction: 3 das/10%/
H-13 Full Hydrologic Restoration with Full Costs.
Lake Okeechobee: 90% Costs: $50 15.6 52.7 13 12 -$109.35 -$7.68
Water Conservation: 90% Farmland: 200,000 acres
Everglades National Park: 90% Water Restriction: I day/40%
NA=Not Applicable
6-12
6.3.2 Evaluations of Alternative Plans with the Species Multiattribute Model by Respondents'
Location and Environmental Donation Status
The set of 11 alternative restoration plans constructed for the species multiattribute model presented
in Table 6-2 was evaluated using the socioeconomic characteristics interactive model coefficients
reported previously in Table 5-5. And, the same three evaluation measures were used as in the
preceding analyses. The interactive model was used to disaggregate the species restoration plan
evaluations into responses based on respondent location (South and Central Florida) and past donations
to environmental groups.
Evaluation results for the 11 species restoration plans for respondents in South Florida and Central
Florida are reported in Table 6-5. In contrast to the hydrological attributes evaluation, the results in
Table 6-5 show relatively few differences in the preferences of South vs Central Floridians. Both rank
S3 'Full Wetland Wildlife Restoration Without Costs' as the most preferred alternative. And, a large
majority of both South Floridians (88.7 percent) and Central Floridians (98.9 percent) would vote in
favor of this plan. Both groups' net WTP for this plan are very similar ($68.82 and $72.42) and are also
comparable to South Floridians' net WTP of $86.42 for their highest ranked hydrological restoration
plan. Of the 11 species restoration plans, South Floridians prefer 4 of the plans to the status quo while
Central Floridians prefer 7 of the plans to the status quo. Neither group favored the full or partial
dryland species restoration plans (S4 and S5).
While the cost components clearly influenced the relative desirability of the plans, Central Floridians
were generally less sensitive to the cost and water restriction attributes of the plans. Both S7 and S8,
partial wetland wildlife species restoration plans with different cost attribute levels, were favored by
a majority of Central Floridians but not by South Floridians. The net WTPs for both plans were
positive for Central Floridians ($12.25 and $14.37, respectively) but negative for South Floridians
(-$6.31 and -$22.39, respectively). Indeed, the only wetland wildlife restoration plan Central
Floridians did not have a majority in favor or a positive net WTP was the full wetland wildlife
restoration plan with full costs (S 11).
A similar evaluation of the 11 species restoration plans is reported by respondent's environmental
donation status in Table 6-6. As in the earlier evaluation of the hydrological restoration plans by
environmental donation status (Table 6.4), respondents who donated to environmental causes in the past
were more likely to favor the restoration plans. Seven of the plans were ranked higher than the status
quo by those who donated compared with only 4 plans that were ranked higher by those who had not
donated. But, in contrast to the hydrological restoration plan results, the overall differences between the
donation status groups were not large. The overall percent in favor and net WTPs were relatively
similar for partial wetland wildlife restoration with a $25 annual household cost (S6). The differences
grew larger only when additional cost components were added such as in S8 and S9 which added
100,000 acre reductions in farmland and severe water use restrictions. The differences in net WTPs
for S9 for those who donated ($39.81) versus those who did not ($15.00) suggest that the former were
less averse to the prospect of severe water use restrictions as a tradeoff for full wetland wildlife
restoration.
In general, the socioeconomic interaction evaluation results in Tables 6-5 and 6-6 indicate that the
species multiattribute format tended to elicit fewer differences in preferences across socioeconomic
groups than the hydrological multiattribute format (Tables 6-3 and 6-4). This result may have occurred
because more respondents were able to understand the concept of restoring species population levels
6-13
as compared to restoring historic water levels and timing. Respondents in Central Florida and those
who had not contributed to environmental groups were the least likely to have strong preferences for
hydrological restoration plans. These results suggest that building support for Everglades/South Florida
ecosystem restoration on the basis of hydrological restoration endpoints may be difficult outside of
South Florida.
6-14
Table 6-5. Evaluation of Selected Restoration Plans with the Species Mulitattribute Model by Respondent Location.
Percent in Favor Ranking Net Willingness to Pay
Plan Description
South Central South Central South Central
Florida Florida Florida Florida Florida Florida
S-1 Baseline (no change) Wildlife.
Wetland Species: 20% Costs: 0 NA NA 5 8 NA NA
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/10%
S-2 Partial Wetland Wildlife Restoration without Costs.
Wetland Species: 50% Costs: 0 88.7 98.9 2 3 $34.41 $36.21
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 75% Water Restriction: 3 days/10%
S-3 Full Wetland Wildlife Restoration without Costs.
Wetland Species. 80% Costs: 0 88.7 98.9 1 1 $68.82 $72.42
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 90% Water Restriction: 3 days/10%
S-4 Partial Dryland Wildlife Restoration without Costs.
Wetland Species: 20% Costs: 0 7.0 34.8 8 9 -$17.62 -$3.87
Dryland Species: 60% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/10%
S-5 Full Dryland Wildlife Restoration without Costs.
Wetland Species: 20% Costs: 0 7.0 34.8 10 10 -$35.23 -$7.74
Dryland Species: 70% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/l0%
S-6 Partial Wetland Wildlife Restoration with Minimized
Taxes.
Wetland Species: 50% Costs: $25 59.9 64.1 4 7 $9.41 $11.21
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 75% Water Restriction: 3 days/10%
S-7 Partial Wetland Wildlife Restoration, No Water Supply
Restrictions.
Wetland Species: 50% Costs: $25 43.7 65.2 6 6 -$6.31 $12.25
Dryland Species: 50% Farmland Reduction: 100,000 acres
Estuarine Species: 75% Water Restriction: 3 days/10%
6-15
Table 6-5. Continued.
Percent in Favor Ranking Net Willingness to Pay
Plan Description
South Central South Central South Central
Florida Florida Florida Florida Florida Florida
S-8 Partial Wetland Wildlife Restoration with Minimized
Costs.
Wetland Species: 50% Costs: $25 21.8 75.0 9 5 -$22.39 $14.37
Dryland Species: 50% Farmland Reduction: 100,000 acres
Estuarine Species: 75% Water Restriction. 2 days/25%
S-9 Full Wetland Wildlife Restoration with Minimized Costs
Wetland Species- 80% Costs: $25 54.9 88.0 3 2 $12.02 $50.58
Dryland Species: 50% Farmland Reduction: 100,000 acres
Estuarine Species: 90% Water Restriction: 2 days/25%
S-10 Full Wetland Wildlife Restoration with Costs, No
Water Supply Restrictions
Wetland Species: 80% Costs: $50 43.7 65.2 7 4 -$12.62 $24.50
Dryland Species: 50% Farmland Reduction: 200,000 acres
Estuarine Species: 90% Water Restriction: 3 days/10%
S-11 Full Wetland Wildlife Restoration with Full Costs
Wetland Species: 80% Costs: $50 26.1 35.9 11 11 -$47.85 -$12.63
Dryland Species: 50% Farmland Reduction:200,000 acres
Estuarine Species: 90% Water Restriction: 1 day/40%
NA=Not Applicable
6-16
Table 6-6. Evaluation of Restoration Plans with the Species Multiattribute Model by Respondent Past Donations.
Percent in Favor Ranking Net Willingness to Pay
Plan Description
DONATIONS DONATIONS DONATIONS
NO YES NO YES NO YES
S-1 Baseline (no change) Wildlife.
Wetland Species: 20% Costs: 0 NA NA 5 8 NA NA
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/10%
S-2 Partial Wetland Wildlife Restoration without Costs.
Wetland Species: 50% Costs: 0 92.4 93.0 2 3 $33.61 $36.68
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 75% Water Restriction: 3 days/10%
S-3 Full Wetland Wildlife Restoration without Costs.
Wetland Species: 80% Costs: 0 92.4 93.0 1 1 $67.23 $73.37
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 90% Water Restriction: 3 days/10%
S-4 Partial Dryland Wildlife Restoration without Costs.
Wetland Species: 20% Costs: 0 10.1 26.1 8 9 -$15.65 -$8.65
Dryland Species: 60% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/10%
S-5 Full Dryland Wildlife Restoration without Costs.
Wetland Species: 20% Costs: 0 10.1 26.1 10 10 -$31.31 -$17.30
Dryland Species: 70% Farmland Reduction: 0
Estuarine Species: 60% Water Restriction: 3 days/10%
S-6 Partial Wetland Wildlife Restoration with Minimized Taxes.
Wetland Species: 50% Costs: $25 63.0 60.0 4 4 $8.61 $11.68
Dryland Species: 50% Farmland Reduction: 0
Estuarine Species: 75% Water Restriction: 3 days/10%
6-17
Table 6-6 Continued.
Percent in Favor Ranking Net Willingness to Pay
Plan Description
DONATIONS DONATIONS DONATIONS
NO YES NO YES NO YES
S-7 Partial Wetland Wildlife Restoration, No Water Supply
Restrictions. 50.4 53.9 6 6 -$2.26 $4.35
Wetland Species: 50% Costs: $25
Dryland Species: 50% Farmland Reduction: 100,000 acres
Estuarine Species: 75% Water Restriction. 3 days/10%
S-8 Partial Wetland Wildlife Restoration with Costs.
Wetland Species. 50% Costs: $25 27.7 58.3 9 7 -$18.62 $3.12
Dryland Species: 50% Farmland Reduction: 100,000 acres
Estuarine Species: 75% Water Restriction: 2 days/25%
S-9 Full Wetland Wildlife Restoration with Minimized Costs.
Wetland Species: 80% Costs: $25 59.7 76.5 3 2 $15.00 $39.81
Dryland Species: 50% Farmland Reduction: 100,000 acres
Estuarine Species: 90% Water Restriction: 2 days/25%
S-10 Full Wetland Wildlife Restoration with Costs with No Water
Supply Restrictions. 50.4 53.9 7 5 -$4.52 $8.71
Wetland Species: 80% Costs: $50
Dryland Species: 50% Farmland Reduction: 200.000 acres
Estuarine Species- 90% Water Restriction: 3 days/10%
S-11 Full Wetland Wildlife Restoration with Full Costs.
Wetland Species: 80% Costs: $50 24.4 35.7 11 11 -$47.59 -$19.93
Dryland Species: 50% Farmland Reduction: 200,000 acres
Estuarine Species: 90% Water Restriction: 1 day/40%
NA=Not Applicable
6-18
SECTION 7. SUMMARY AND EXTENSIONS
7.1 Study Objectives and Methods
Efforts to restore the Everglades/South Florida ecosystem represent one of the most difficult challenges
in contemporary science and environmental decision-making. Numerous scientific questions exist over
the current state of the ecosystem and how the components of the system will respond to management
initiatives. Present and future scientific studies will help to address these uncertainties. Equally, if not
more, difficult questions arise over the choice of objectives for restoration. These questions involve
both scientific and public preference issues that require both natural and social science research. This
report has attempted to answer some of these questions by presenting the results from a survey of
Florida residents to identify their preferences and economic values for different levels of ecosystem
restoration. These results provide baseline information about Floridians' preferences that can be used
in contemporary assessments of alternative restoration plans and for comparisons with future social
science studies of public preferences.
The survey was designed as an application of multiattribute utility (MAU) theory because the
Everglades/South Florida ecosystem restoration problem requires the consideration of multiple
dimensions as well as tradeoffs between natural and social system considerations. Given the geographic
scale and ecological complexity of the region, a large number of attributes could possibly be considered.
The range of possibilities was limited, however, by the need to develop a survey instrument that could
be readily understood by the public and could be administered under normal field conditions. To
represent ecological dimensions of the problem, functional endpoints were expressed as hydrological
attributes that represented water levels and timing in three critical geographic areas: Lake Okeechobee,
the Water Conservation Areas, and Everglades National Park. An alternative characterization of the
ecological dimensions of the problem was expressed as species attributes that represented three broad
species groupings: wetland species, dryland species, and estuarine species. These alternative attribute
specifications were both measured in terms of 'percentages of historical levels' to describe how current
and possible future conditions would change under alternative management plans. Other attributes used
in the analysis captured the possible effects of restoration plans on water availability to municipal users,
changes in farmland acreage in South Florida, and the annual cost to households.
Field interviews were conducted by an independent market research firm during the first half of 1998
based on a stratified, random sampling design of households in five Florida counties: Dade,
Hillsborough, Lee, Orange and Palm Beach. Approximately 100 interviews were conducted in each
county. The interviews were conducted in respondents' homes and consisted of five parts: 1) a general
introduction, 2) questions to elicit respondents' attitudes about public expenditures and environmental
issues, 3) an informational video about the Everglades/South Florida ecosystem, 4) description of the
choice attributes and a pairwise choice task involving preselected attribute combinations, and 5)
additional questions to elicit socioeconomic characteristics. Copies of all questions and supporting
materials for the interviews are provided in the appendices to this report.
A pairwise choice response procedure was used in the interviews because protesting indicated this was
the simplest and most easily understood procedure for respondents to deal with the large number of
attributes and attribute combinations. A split sample design was used so that the hydrological and
species attribute specifications were divided evenly across respondents in each county. This design
provided a basis to compare respondents' preferences for alternative restoration plans based on different
representations of the ecological dimensions of the problem. While the interview procedure used in this
study required more time and was more difficult than a typical public opinion survey, it was clear that
respondents were able to understand the choices and express their preferences. Moreover, the survey
framework provides detailed information about respondents' preferences that could be used to address
specific components of proposed restoration plans.
7.2 Survey Results and Alternative Restoration Plan Evaluations
Responses from the pairwise choice tasks were used to estimate MAU functions for the hydrological
and species attribute sets. The MAU functions were estimated using specifications that included only
the basic attributes and with interactive variables to identify the effects of socioeconomic characteristics
on respondent preferences. The results reported in Section 5 showed that both specifications resulted
in strong statistical results indicating that respondents evaluated the attribute combination options and
expressed their preferences for different levels of the attributes.
The MAU functions for the hydrological and species models can be compared through the use of
normalized weights'2 for each attribute such as those presented in Figure 7-1. The figure shows that
respondents gave a positive weight to all three hydrological attributes. The results indicate respondents
preferred potential restoration plans that would lead to water levels and timing throughout the South
Florida region that would be more similar to historical conditions. Other attributes in the model,
however, revealed that higher levels of annual cost, water restrictions, or reductions in farmland acreage
were negatively weighted. The negative weights assigned to these attributes would temper the positive
weights given to higher levels of the hydrological attributes.
Weights for the species attributes from the species multiattribute model in Figure 7-1 show that
respondents gave a positive weight to wetland and estuarine species attributes. But, they also gave a
significantly negative weighting to the dryland species attribute indicating that restoring these species
populations to be more comparable to historical levels was not preferred. The results for restoring
wetland/estuarine species were also tempered by negative weights assigned to the annual cost, water
restrictions, and farmland attributes. Figure 7-1 also shows that the weightings assigned to these three
attributes in the hydrological and species multiattribute models were generally similar suggesting that
respondents expressed consistent preferences in the two choice settings despite the presence of other
attributes.
The statistically derived \% eights for the hydrological and species multiattribute models were used to
evaluate alternative hypothetical restoration plans. The evaluation results reported in Section 6
indicated that Floridians expressed strong preferences for restoration plans that restore the hydrological
conditions or wetland/estuarine species populations of South Florida. For example, Tables 6-1 and 6-2
showed that a large majority of respondents would be in favor of full restoration plans that imposed no
or low costs on Floridians. In addition, the highest net willingness to pay (WTP) amounts for the full
restoration plans were $58.79 annually with the hydrological multiattribute model and $69.86 annually
with the species multiattribute model. These economic values measure the maximum annual benefits
the average resident would receive if full restoration were accomplished with no costs to Floridians.
'2The normalized weights were computed by multiplying each attribute coefficient times the
mean level of the attribute and then multiplying the product times ten.
Figure 7-1. Relative Weightings of Attributes in the Multiattribute Models
Lake Okeechobee 5.8 Wetland Species 2.9
Water Conservation Areas 6.2 Drytand Species .7
Everglades National Park 2.4 Estuarine Species 8.1
Cost Per Household -3 Cost Per Household -2.4
HH Water Use I -0.1 HH Water Use 1 -0.4
HH Water Use It -2.3 HH Water Use II .1.7
Farm Land in EAA -1.9 Farm Land in EAA -0.9
-4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 10
Hydrological Model Species Model
Costs imposed by a plan would reduce these maximum amounts and vary with the specifics of each
plan. The tables in Section 6 showed how these amounts would change with different plans for the
sample as a whole and for various socioeconomic groups.
There were differences in preferences for restoration between socioeconomic groups. Respondents in
Central Florida and those who had not donated to environmental groups were much less supportive of
restoration plans when these respondents expressed their preferences with the hydrological
multiattribute choice set. The same results, however, did not occur in responses with the species
multiattribute set. The differences for these groups may be that they did not relate the hydrological
attributes to ecosystem restoration and therefore gave less weight to these attributes. These results
suggest that it may be more difficult to explain and build support for an Everglades/South Florida
restoration program if the emphasis is primarily hydrological management. The fact that higher levels
of the wetland/estuarine species attributes were consistently preferred across the sample indicates that
the general public would more readily identify with a restoration program that emphasizes species
restoration.13
13 It is interesting to note that the U.S. Fish and Wildlife Service recently announced a "South
Florida Multi-Species Recovery Plan" that U.S. Department of the Interior Secretary Bruce Babbitt
describes as "an integral component in meeting the Administration's restoration plan for South
Florida" (1999, pp. I). This recovery plan is distinct from the Restudy yet many of the species
identified in the plan "are dependent upon implementation of the Restudy for survival and recovery,
and many others will benefit significantly as a result of this restoration effort" (1999, pp. ix). The
perspective of the Plan's sponsors is clearly revealed in the statement:
"South Florida cannot be considered 'restored' if the song of the Cape Sable seaside
sparrow has been silenced from the marl prairies of the Everglades, if clouds of wood
storks no longer cover the skies of the mangrove forests of southernmost Florida, if
the footprints of marsh rabbits no longer occur in the Lower Keys, or if there are no
Florida scrub-jay families defending their territories on the Central Florida Ridge"
(1999, pp. 1-2).
While the hypothetical restoration plan evaluation results indicated generally strong support for
restoration, the results in Section 6 made it evident that respondents would not support a restoration plan
that imposed high costs on Floridians. With either the hydrological or species multiattribute models,
hypothetical restoration plans that included annual costs of $50 per household ($500 over 10 years)
coupled with either farmland reductions of 100,000 or more acres or severe restrictions on municipal
water use during dry years received poor rankings and less than majority support. Moreover, the net
WTPs for these 'high cost' plans were negative suggesting a potential loss in economic welfare. Even
socioeconomic groups such as those who had made environmental donations and typically revealed the
strongest support for most plans had unfavorable evaluations of the high cost plans.
Overall the results of both the hydrological and species multiattribute models indicate that the
likelihood of support for a restoration plan will depend on the balancing of restoration objectives with
costs imposed on Floridians. The level of restoration will have to be considered in relation to the costs
imposed. Some costs were clearly not viewed as too overbearing such as moderate restrictions on water
use. But, it is clear that other costs associated with possible restoration plans such as severe water
restrictions (outdoor uses restricted to 1 day per week and indoor uses reduced 40 percent) or annual
costs greater than $25 per household would encounter significant opposition from the general public.
Thus, the task of developing an actual plan for the Everglades/South Florida region that achieves
meaningful ecosystem restoration, satisfies Floridians' obligation for cost sharing under the Water
Resources Development Act of 1996, and is supported by the public may be a daunting endeavor.
7.3 Extending the Economic Valuation Results to the Florida Population
The results presented in Section 6 for the survey sample could also be extrapolated to the Florida
population to estimate the aggregate benefits of alternative Everglades/South Florida restoration plans.
Although the sampling design for this study was selected to represent the population of five counties
(Dade, Hillsborough, Lee, Orange and Palm Beach), the comparison of the socioeconomic
characteristics of the sample and the Florida resident population in Section 4 (Tables 4-land 4-3)
indicated a strong similarity between the sample and the population. This is not surprising since each
county is located in the major population centers of the state and the five counties account for more than
one-third of the total state population. The primary areas that were not represented in the sample are
the urban and rural areas of North Florida. With these limitations of the sample data in mind,
population estimates can be calculated for the 5.82 million households14 in Florida in 1997 (Floyd et
al., 1998).
Using the net WTP estimates reported in Section 6, aggregate measures of the economic benefits to
Florida residents can be calculated by multiplying the average WTP for a specific plan by the number
of households. For example, using the net WTP value of $58.79 per household per year for full
hydrological restoration without costs (Plan H-5 in Table 6-1) from the hydrological multiattribute
model would result in an aggregate annual WTP of $342.2 million for this plan. Over the ten-year
period indicated in the cost attribute description (see Table 3-3), the aggregate WTP would be $3.42
14 Population estimates are presented for households rather than individuals because only
respondents over 18 years of age who participated in household financial decisions were selected.
Also, the cost attributes used in the multiattribute models were expressed in terms of annual
household costs.
billion in 1998 dollars." Similarly, the WTP value of $69.86 per household for full wetland/estuarine
species restoration without costs (Plan S-3 in Table 6-2) from the species multiattribute model would
result in an annual WTP of $406.5 million for this plan or $4.07 billion over ten years.16 These are
aggregate measures of the maximum economic benefits Floridians would derive from a full restoration
plan with no costs to Floridians.
A plan with direct costs to Floridians would obviously result in lower aggregate benefits for Floridians.
For example, the restoration plan H10 developed from the hydrological multiattribute model would
provide full hydrological restoration with annual costs of $25 per household, a 100,000 acre reduction
in farmland, and moderate water restrictions. This plan yielded a net WTP of $15.60 for the sample
(Table 6-1). Aggregating this economic benefit measure to the population would result in annual
benefits of $90.8 million or $907.9 million over ten years. The same approach can be used to aggregate
any of the net WTP results reported in Tables 6-1 and 6-2.
7.4 Limitations of the Study and Suggestions for Future Research
This research study was intended to provide an understanding of the preferences of Floridians for
Everglades/South Florida ecosystem restoration. In any effort to identify public preferences, especially
for complex issues such as ecosystem restoration, simplifying assumptions are always necessary. In
the survey, it was necessary to limit the dimensions of the problem to make the multiattribute tradeoffs
easily understood and consistent with the limited information that was available about different
attributes of the ecosystem. While these constraints on the choices provided to respondents may have
influenced the results, the extent and direction of this influence cannot be known until a different survey
is conducted using more extensive and complete information on the ecosystem and social attributes
influenced by restoration decisions. Similarly, other multiattribute elicitation procedures such as direct
rating and ranking of plan alternatives by respondents may influence the results. As with any social or
natural science research, replication is necessary to validate results.
Also, the geographic coverage for this survey was constrained to large population centers of Florida due
to the project budget. While the survey design resulted in a representative sample, certainly the
preferences of rural Floridians and those who live in North Florida were not included. Whether the
preferences of these Florida residents differ from the sample in this study cannot be known without
further research.
Perhaps more importantly, this research provided no information about the preferences of people who
live in other states or the millions of domestic and international visitors who come to Florida each year.
15 Growth in the number of households in Florida or changes in preferences over the ten-year
period are not considered in this calculation. Also, this calculation assumes that no additional
benefits accrue beyond the ten-year period.
16 An alternative approach to aggregating the benefits over time is to assume that the annual
benefits would continue into perpetuity. Then the capitalized value of the benefits could be
calculated by dividing the annual benefits by a capitalization (discount) rate. Assuming a real
discount rate of 3 percent, the capitalized value of the economic benefits for full restoration without
costs from the hydrological and species multiattribute models would be $11.4 and $13.5 billion,
respectively.
Certainly the Everglades/South Florida region has a unique place among the U.S.'s national treasures
(see Section 1), yet no information exists about the preferences of U.S. citizens for restoration of the
region's ecosystem. This information is particularly important in light of the 50/50 state and federal
cost sharing relationship established for Everglades/South Florida ecosystem restoration in the Water
Resources Development Act of 1996.
Finally, it is important to recognize that social science research on public preferences is an integral part
of an adaptive management approach to Everglades/South Florida ecosystem restoration (Milon et al.
1997). Uncertainty about the responses of natural systems to human intervention means that both
scientists and the public will learn about the effects of restoration as a selected plan proceeds. As
scientists update their information, so too will the public update their perceptions of the ecosystem and
the restoration plan. Therefore, the results from this study should be viewed as a benchmark for future
surveys to objectively evaluate how public perceptions and economic values may change over time.
This integration of natural and social science research in an adaptive management framework offers the
promise of continually improving the information available for both resource managers and the public.
SECTION 8. REFERENCES
Adamowicz, V., J. Louviere, and M. Williams. "Combining Revealed and Stated Preference Methods
for Valuing Environmental Amenities." Journal of Environmental Economics and Management
26(1994): 271-292.
Addelman, S. "Orthogonal Main-Effect Plans for Asymmetrical Factorial Experiments." Technometrics
4(1962): 21-46.
Beard, D.B. Wildlife Reconnaissance: Everglades National Park Project. National Park Service, U.S.
Department of the Interior, 1938.
Beggs, S., S. Cardell, and J. Hausman. "Assessing the Potential Demand for Electric Cars." Journal of
Econometrics 17(1981): 1-20.
Ben-Akiva, M. and S. Lerman. Discrete Choice Analysis. Cambridge: MIT Press, 1985.
Ben-Akiva, M., T. Morikawa, and F. Shiroishi. "Analysis of the Reliability of Preference Ranking
Data." Journal of Business Research 24(1992): 149-164.
Borcherding, K., T. Eppel and D. vonWinterfeldt. "Comparison of Weighting Judgments in
Multiattribute Utility Measurement. Management Science 37(1991): 1603-19.
Boxall, P.C., W.L. Adamowitz, J. Swait, M. Williams and J. Louviere. "A Comparison of Stated
Preference Methods for Environmental Valuation."Ecological Economics 18(1996):243-53.
Bratton, S.P. "Alternative Models of Ecosystem Restoration." In Ecosystem Health: New Goals for
Environmental Management, R. Costanza, B. Norton, and B. Haskell, eds. Washington, D.C.: Island
Press, 1992.
Cairns, J. "Restoration Ecology: The New Frontier." In Rehabilitating Damaged Ecosystems, J.
Cairns, ed. Boca Raton, FL: CRC Press, 1988.
Davis, S.M. and J.C. Ogden, eds. Everglades The Ecosystem and Its Restoration. Delray Beach,
FL: St. Lucie Press, 1994.
De Palma, A., G. Myers, and Y. Papageorgiou. "Rational Choice Under an Imperfect Ability to
Choose." American Economic Review 84(1994):419-440.
Douglas, M.S. The Everglades: River of Grass. Sarasota, FL: Pineapple Press, 1947 (revised ed.,
1988).
Ehrhardt, N.M. "The Lobster Fisheries off the Caribbean Coast of Central America." In Spiny Lobster
Management, B. Phillips. J. Cobb and J. Kittaka, eds., pp. 133-143. Cambridge, MA: Blackwell, 1994
Everglades Coalition. The Greater Everglades Ecosystem Restoration Plan. Washington, DC, 1994.
Fennema, R.J., C.J. Neidrauer, R.A. Johnson, T.K. MacVicar, and W.A. Perkins. "A Computer Model
to Simulate Natural Everglades Hydrology." In Everglades The Ecosystem and Its Restoration. S.M.
Davis and J.C. Ogden, eds., pp. 249-289. Delray Beach, FL: St. Lucie Press, 1994.
Floyd, S., E. Irwin, and D. Evans. 1998 Florida Statistical Abstract. Gainesville, FL: Bureau of
Economic and Business Research, University of Florida, 1998.
Giocoechea, A, D.R. Hansen and L. Duckstein. Multiobjective Decision Analysis With Engineering and
Business Applications. John Wiley, New York, 1982.
Greene, W. Econometric Analysis, Third Edition. Englewood Cliffs: Prentice Hall, 1997
Green, C.H. and S.M. Tunstall. "Is the Economic Evaluation of Environmental Resources Possible?"
Journal of Environmental Management 33(1991): 123-141.
Harwell, M.A., W. Cooper, and R. Flaak. "Prioritizing Ecological and Human Welfare Risks from
Environmental Stresses." Environmental Management 16(1992): 451-464.
Harwell, M.A. and J.F. Long. US MAB Human-Dominated Systems Directorate Workshop on
Ecological Endpoints and Sustainability Goals. United States Man and the Biosphere Program.
Rosensteil School of Marine and Atmospheric Science, University of Miami, May 1992.
Hausman, J. and P. Ruud. "Specifying and Testing Econometric Models for Rank-Ordered Data."
Journal of Econometrics 34(1987): 83-104.
Hoehn, J.P. and J.B. Loomis. "Substitution Effects in the Valuation of Multiple Environmental
Programs." Journal of Environmental Economics and Management 25(1993): 56-75.
Holling, C.S. "Simplifying the Complex: The Paradigms of Ecological Function and Structure."
European Journal of Operational Research 30(1987): 139-146.
Holling, C.S., L.H. Gunderson, and C.J. Walters. "The Structure and Dynamics of the Everglades
System: Guidelines for Ecosystem Restoration." In Everglades The Ecosystem and Its Restoration.
S.M. Davis and J.C. Ogden, eds. Delray Beach, FL: St. Lucie Press, 1994.
Hunt, J. "Status of the Spiny Lobster Fishery in Florida." In Spiny Lobster Management, B. Phillips.
J. Cobb and J. Kittaka, eds., pp. 158-168. Cambridge, MA: Blackwell, 1994
Hwang, C. and K. Yoon, 1981. Multiple Attribute Decision Making. Springer-Verlag, Berlin.
Johnson, F.R. and W.H. Desvousges. "Estimating Stated Preferences with Rated-Pair Data:
Environmental, Health, and Employment Effects of Energy Programs." Journal of Environmental
Economics and Mianagement 34(1997):79-99.
Keeney, R.L., 1980. Siting Energy Facilities. Academic Press, New York, 413 p.
Keeney, R.L. and H. Raiffa. Decisions with Multiple Objectives: Preferences and Value Tradeoffs.
New York, Wiley, 1976.
Labisky, R., D. Gregory, and J. Conti. "Florida's Spiny Lobster Fishery: An Historical
Perspective." In Spiny Lobster Series, G. Davis, ed., pp. 28-37. Bethesda, MD: American Fisheries
Society, 1980.
Light, S.S. and J.W. Dineen. "Water Control in the Everglades: A Historical Perspective." In
Everglades The Ecosystem and Its Restoration. S.M. Davis and J.C. Ogden, eds. Delray Beach, FL:
St. Lucie Press, 1994.
Light, S.S., L.H. Gunderson, and C.S. Holling. "The Everglades: Evolution of Management in a
Turbulent Ecosystem." In Barriers and Bridges to the Renewal of Ecosystems and Institutions. L.H.
Gunderson, C.S. Holling, and S.S. Light, eds. New York: Columbia University Press, 1995.
Louviere, J. Analyzing Decision Making: Metric Conjoint Analysis. Newbury Park, CA: Sage
Publications, 1988.
Louviere, J. "Experimental Choice Analysis: Introduction and Overview." Journal ofBusiness Research
24(1992): 89-96.
Mackenzie, J. "A Comparison of Contingent Preference Models." American Journal ofAgricultural
Economics 75(1993): 593-603.
Maltby, E. and P.J. Dugan. "Wetland Ecosystem Protection, Management, and Restoration: An
International Perspective." In Everglades The Ecosystem and Its Restoration, S.M. Davis and J.C.
Ogden, eds. Delray Beach, FL: St. Lucie Press, 1994.
May, R. Stability and Complexity in Model Ecosystems. Princeton: Princeton University Press, 1973.
Mazzotta, M. and J. Opaluch. "Decision Making When Choices are Complex: A Test of Heiner's
Hypothesis." Land Economics 71(1995):500-515.
McFadden, D. "The Choice Theory Approach to Market Research." Marketing Science 5(1986): 275-
297.
McIvor, C.C., J.A. Ley, and R.D. Bjork. "Changes in Freshwater Inflow from the Everglades to Florida
Bay Including Effects on Biota and Biotic Processes: A Review." In Everglades The Ecosystem and
Its Restoration, S.M. Davis and J.C. Ogden, eds. Delray Beach, FL: St. Lucie Press, 1994.
McNaughton, S. "Diversity and Stability." Nature 333(1988): 204-205.
Miller, G.A. "The Magical Number Seven, Plus or Minus Two: Some Limits on our Capacity for
Processing Information." Psychological Review 62(1956): 81-97.
Milon, J. W., C.M. Adams, and D.W. Carter. Floridian 's Atiiides About the Environment and Coastal
Marine Resources. Florida Sea Grant Technical Paper No. 95. Gainesville, FL: Florida Sea Grant
College Program, August 1998.
Milon, J.W., C.F. Kiker, and D.J. Lee. "Ecosystems and Social Conflict: Lessons from the Florida
Everglades." In Conflict and Cooperation on Trans-Boundary Water Resources, R. Just and S.
Netanyahu, eds., pp. 231-247. New York: Kluwer Academic Publishers, 1998.
Milon, J.W., C.F. Kiker, and D.J. Lee. "Adaptive Ecosystem Management and the Florida Everglades:
The Role of Social Scientists." Journal ofAgricultural and Applied Economics 29(1997):99-107.
Milon, J.W., C.F. Kiker, and D.J. Lee. "Adaptive Ecosystem Management and the Florida Everglades:
More Than Trial-and-Error?" Water Resources Update, Issue No. 113, Autumn 1998:37-46.
Milon, J.W., S. Larkin, D.J. Lee, K.J. Quigley, and C.M. Adams. The Performance of Florida's Spiny
Lobster Trap Certificate Program. Florida Sea Grant Report No. 116. Gainesville, FL: Florida Sea
Grant College Program, December 1998.
Mitchell, R. and R. Carson. Using Surveys to Value Public Goods: The Contingent Valuation Method.
Washington, D.C.: Resources for the Future, 1989.
Mollaghasemi, M. and J. Pet-Edwards, Making Multiple-Objective Decisions, IEEE Computer Society
Technical Briefing, Inst. Electrical and Electronics Engineers.
Moore, W. and M. Holbrook. "Conjoint Analysis on Objects with Environmentally Correlated
Attributes: The Questionable Importance of Representative Design." Journal of Consumer Research
16(1990):490-497.
Nijkamp. P., P. Rietveld and H. Voogd. Multicriteria Evaluation in Physical Planning, North-Holland,
Amsterdam, 1990.
National Research Council. Restoring and Protecting Marine Habitat The Role of Engineering and
Technology. Washington, DC, 1994.
Nash, S. "What Price Nature." BioScience 41(1991): 677-680.
Ogden, J.C. "A Comparison of Wading Bird Nesting Colony Dynamics (1931-1946 and 1974-1989)
as an Indication of Ecosystem Conditions in the Southern Everglades." In Everglades The Ecosystem
and Its Restoration, S.M. Davis and J.C. Ogden, eds. Delray Beach, FL: St. Lucie Press, 1994.
Opaluch, J.J., S.K. Swvallow, T. Weaver, C.W. Wessells, and D. Wichelns. "Evaluating Impacts from
Noxious Facilities: Including Public Preferences in Current Siting Mechanisms." Journal of
Environmental Economics and Management 24(1993): 41-59.
Payne, J.W., J.R. Bettman, and E.J. Johnson. "Adaptive Strategy Selection in Decision Making."
Journal of Experimental Psychology: Learning, Memory, and Cognition 14(1988): 534-552.
Rife Market Research, Inc. "Focus Group Sessions on Public Preferences for Restoration of the Florida
Everglades: Summary Report." Miami, FL, January 1997.
Roe, B., K. Boyle, and M. Teisl. "Using Conjoint Analysis to Derive Estimates of Compensating
Variation." Journal of Environmental Economics and Management 31(1996):145-159.
Russell, C.S. "Moving from 'Principle' to 'Practical' in the Valuation of Changes in Ecological
Systems." Research and Exploration 9(1993): 272-273.
Russel. C.S.. "Old Lessons and New Contexts in Economic-Ecological Modeling." In Integrating
Economic and Ecological Indicators, J.W. Milon and J.F. Shogren, eds., pp. 9-25. New York: Praeger
Publishers, 1995.
Saaty, T.L., 1980. The Analytic Hierarchy Process. McGraw-Hill, New York.
SAS Institute. The FACTEX Procedure, SAS Procedures Guide, ch. 4, pp. 209-35, Cary, NC.
Science Sub-Group. Report of Federal Objectives for the South Florida Restoration. J. Weaver and
B. Brown, eds. Report of the Science Sub-Group to the South Florida Management and Coordination
Working Group, November 15, 1993.
South Florida Water Management District. Water Shortage Plan. West Palm Beach, FL, February 1991.
South Florida Water Management District and Florida Department of Environmental Protection.
Everglades Program Implementation Program Management Plan, Preliminary Draft. West Palm
Beach, FL, September 13, 1994.
South Florida Water Management District. Lower East Coast Regional Water Supply Plan. West Palm
Beach, FL, March 1997.
Stewart, D.W. Focus Groups: Theory and Practice. Newbury Park, CA: Sage, 1992.
Suter, G. and L. Barnthouse. "Assessment Concepts." In Ecological Risk Assessment, G.W. Suter, ed.
, pp. 21-47. Boca Raton, FL: Lewis Publishers, 1993.
Swallow, S., T. Weaver, J. Opaluch, and T. Michelman. "Heterogeneous Preferences and Aggregation
in Environmental Policy Analysis: A Landfill Siting Case." American Journal of Agricultural
Economics 76(1994): 431-443.
Szidarovszky, F., M.E. Gershon and L. Duckstein, 1986. Techniques for Multiobjective Decision
Making in Systems Management. Elsevier, Amsterdam, 506 p.
U.S. Army Corps of Engineers. Comprehensive Review Study: Reconnaissance Report, Volumes I and
II. Central and South Florida Project, Jacksonville District, 1994.
U.S. Army Corps of Engineers. Final Integrated Feasibility Report and Programatic Environmental
Impact Statement. Central and South Florida Project Comprehensive Review Study, Jacksonville
District, April 1999.
U.S. Fish and Wildlife Service. Action Plan, South Florida Ecosystem Restoration. Draft. 1993.
U.S. Fish and Wildlife Service. South Florida Multi-Species Recovery Plan. Atlanta, GA: May 1999.
Vogel, C.C. "Central & Southern Florida Project Comprehensive Review Study: Road Map or
Roadblock for the Future? A Case Study in Water Resource Planning in the Age of Ecosystem
Management." Water Resources Update Issue No. 111(Spring 1998):83-97.
von Winterfeldt, D. and W. Edwards. Decision Analysis and Behavioral Research. Cambridge:
Cambridge University Press, 1986.
Westman, W. Ecology, Impact Assessment, and Environmental Planning. New York: John Wiley and
Sons, 1985.
Westman, W. "Ecological Restoration Projects: Measuring Their Performance." The Environmental
Professional 13(1991):207-215.
Wilson, E.O. Biophilia. Cambridge: Harvard University Press, 1984.
Wilson, E.O. The Diversity ofLife. Cambridge: Harvard University Press, 1992.
APPENDIX A
QUESTIONNAIRE AND INTERVIEWER GUIDE
PUBLIC PREFERENCES FOR CHANGES IN THE SOUTH FLORIDA
WATER MANAGEMENT SYSTEM
QUESTIONNAIRE
AND
INTERVIEWER GUIDE
I. RESPONDENT SCREENING
S-1. READ: Are you 18 years of age or older? If NO, may I speak with someone in the
household who is?
Age of respondent:
Yes 1
No 2
Refused 88
S-2. READ: Are you currently a resident of the State of Florida?
Yes 1
No 2
Refused 88
S-3. READ: Are you usually involved in the financial decision making for this household?
Yes 1
No 2
Refused 88
S-4. READ: This interview is being conducted as part of a research project by the University of
Florida. You will not have to answer any questions you do not wish to answer and
you may discontinue your participation at any time. The interview will take about
45 minutes and you will be paid $10 for your time.
In order to proceed with the interview, I need you to read this consent form and agree to participate
in this interview. (IF RESPONDENT AGREES, CODE S-5 AND S-6)
S-5. CODE: Gender of Respondent
Male 1
Female 2
S-6. CODE: Type of Household Structure
Single-Family Detached Home 1
Townhouse, Row House 2
Attached Apartment or Condominium 3
Mobile Home 4
Duplex 5
Other (Specify) 6
II. BASE SURVEY QUESTIONS
II-A. SPENDING PRIORITIES AND ENVIRONMENTAL ATTITUDES
II-B. VIDEO VIEWING
II-C. DESCRIPTION OF WATER/WILDLIFE MANAGEMENT PLAN ATTRIBUTES AND
ATTRIBUTE LEVELS.
II-D. SELECTION OF EXAMPLE WATER/WILDLIFE MANAGEMENT PLANS
II-E. SELECTION OF SURVEY WATER MANAGEMENT PLANS.
II-F. SOCIO-DEMOGRAPHIC QUESTIONS
II-G. INTERVIEW EVALUATION
SECTION II-A.SPENDING PRIORITIES AND ENVIRONMENTAL
ATTITUDES
READ: Thank you again for participating in this survey. I am going to start this survey by asking
you some general questions about spending on programs by the State of Florida. Please
bear in mind that eventually all government spending comes out of the taxes you and other
Floridians pay. There are no right or wrong answers.
SHOW CARD A THAT LISTS STATE PROGRAMS.
READ: As I mention each program, tell me whether the amount now being spent should be
increased a lot, increased somewhat, kept at the present level, decreased somewhat or
decreased a lot.
SHOW CARD B THAT LISTS RESPONSES TO SPENDING ON STATE PROGRAMS.
CIRCLE ONE CODE FOR EACH QUESTION. RE-READ EACH QUESTION AS NECESSARY.
A-1. READ: Do you think that state spending should increase, be kept at the present level, or
decrease for programs to combat crime?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-2. How about for public schools (K-12)?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-3. To protect the environment?
Increase A Lot 1
A-3
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-4. How about state spending for industrial development and to attract new industry?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-5. To acquire land to protect endangered species?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-6. For health care services?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-7. How about spending for colleges and universities?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-8. For state highways and road systems?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-9. For low income families with children?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-10. How about state spending for the elderly?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-l 1. For state prisons and correctional facilities?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
A-12. How about state spending to promote tourism?
Increase A Lot 1
Increase Somewhat 2
Same 3
Decrease Somewhat 4
Decrease A Lot 5
Don't know 88
WITHDRAW CARD B.
A-13. Of those program areas which you feel should receive increased spending, which one
would you give as your top priority for an increase?
REFER TO CARD A AGAIN: PRIORITY SPENDING PROGRAM
Programs to combat crime 1
Public schools (K-12) 2
Protect the environment 3
Industrial development and to attract new industry 4
Acquire land to protect endangered species 5
Health care service 6
State colleges and universities 7
State highways and road systems 8
Low income families with children 9
Elderly 10
Prisons and correctional facilities 11
Promote tourism 12
No increased spending on anything 13
Don't know 88
WITHDRAW CARD A,
READ: Now on want to ask you a few questions about your household water use.
A-14. Do you currently get your water for household uses such as drinking water and washing
from a water utility or do you have a private well?
Water Utility 1
Private Well 2
Don't Know 88
A-15. How about water for watering the yard? Do you use water from a water utility or do you
have a private well?
A-5
Water Utility 1
Private Well 2
Don't Know 88
A-16. Do you currently use water filters in your home?
Yes 1
No 2
Don't Know 88
A-17. Do you buy bottled drinking water from places such as the grocery store or a water
company to use in your home?
Yes 1
No 2
Don't Know 88
IF RESPONDENT ANSWERS "YES" TO A-17, GO TO QUESTION A-17a. IF NOT, GO TO
QUESTION A-18.
A-17a. IF YES, about what percentage of your total water uses for things like drinking, cooking,
or making coffee and tea comes from bottled water? (ASSIGN RESPONSE TO
CLOSEST CATEGORY)
100% 1
75% 2
50% 3
25% or less 4
Don't Know 88
A-18. Now I want to read you some statements and I'd like you to tell me whether you agree or
disagree with each statement. Please indicate whether you Strongly Agree, Somewhat
Agree, Somewhat Disagree or Strongly Disagree.
SHOW CARD C. AGREE/DISAGREE STATEMENTS
READ QUESTIONS AND CHOICES FIRST TIME, CIRCLE ONE RESPONSE FOR EACH; RE-READ QUESTIONS AS
NECESSARY.
READ: The first statement is: In 10 years, there will not be enough water available for everyone
who wants to move into my community. Do you ...
Strongly Agree 1
Somewhat Agree 2
Somewhat Disagree 3
Strongly Disagree 4
Don't Know 88
A-19. In 10 years, there will not be enough water available for everyone who wants to move into
Florida.
Strongly Agree 1
Somewhat Agree 2
Somewhat Disagree 3
Strongly Disagree 4
Don't Know 88
A-20. Public officials should give first priority to the water supply needs of communities and
economic development before they consider other \' after uses such as maintaining wetlands
and protecting wildlife.
A-6
Strongly Agree 1
Somewhat Agree 2
Somewhat Disagree 3
Strongly Disagree 4
Don't Know 88
A-21. The environment is very delicate and easily upset.
Strongly Agree 1
Somewhat Agree 2
Somewhat Disagree 3
Strongly Disagree 4
Don't Know 88
A-22. When people interfere with the environment it often produces disastrous results.
IF NECESSARY, Do you strongly agree, somewhat agree, somewhat disagree, or strongly
disagree?
Strongly Agree 1
Somewhat Agree 2
Somewhat Disagree 3
Strongly Disagree 4
Don't Know 88
A-23. Public officials have done all they can do to protect the people and buildings in my
community from flooding caused by heavy rains and hurricanes.
Strongly Agree 1
Somewhat Agree 2
Somewhat Disagree 3
Strongly Disagree 4
Don't Know 88
A-24. People have the right to change the natural environment to suit their needs.
Strongly Agree 1
Somewhat Agree 2
Somewhat Disagree 3
Strongly Disagree 4
Don't Know 88
A-25. Everyone should use water conservation practices such as lo\ -flow showers and toilets
and landscaping their yards with plants that need little or no irrigation.
Strongly Agree 1
Somewhat Agree 2
Somewhat Disagree 3
Strongly Disagree 4
Don't Know 88
GO TO SECTION II-B.
A-7
SECTION II-B. VIDEO VIEWING
READ: In this part of the survey I am going to show you a video that will give you some
information on the historical and present water management system in South Florida.
Included in the video is a description of the various uses for water in South Florida and
how those uses compete for water resources in South Florida. After the showing, I am
going to ask you some questions about some of the issues raised in the video. The video
will take about 10 minutes to show.
SHOW THE VIDEO.
AFTER THE VIDEO IS OVER GO TO SECTION II-C.
SECTION II-C. DESCRIPTION OF
WATER MANAGEMENT/WILDLIFE MANAGEMENT
PLAN ATTRIBUTES AND ATTRIBUTE LEVELS
INTERVIEWER NOTE:There are two types of attributes that describe the water management plans
that the respondent will be asked to choose from. SET I Attributes are comprised of
HYDROLOGICAL attributes and these are listed and described in Appendix III-B. SET II
Attributes are comprised of WILDLIFE attributes and a listing and description of these attributes
are given in Appendix III-C.
READ: As you heard in the video, several water management plans are now under consideration.
Each possible water management plan has six attributes. In this part of the interview I am
going to describe to you the various attributes of a water management plan. These
attributes can be best described as representing the types of features one would compare
between 2 automobile makes or models. For example, when buying a car, features to
compare would include the price of the car, safety, horsepower or engine size, fuel
economy, and seating capacity.
SHOW CARD D. ATTRIBUTES OF TWO CAR MODELS
READ: Each car has these same features or attributes, but the amount or level of each of these
attributes will differ according to the make of car. For example, some cars will cost less
than others, but may have fewer safety features as well. Or, one car may get great gas
mileage, but has less seating capacity than another make. People choose the model of the
car they want based on the amount or level of the attributes they value most. The choice
process involves making tradeoffs.
I am now going to describe each attribute of a water management plan, and the corresponding
attribute levels, to you in detail.
INTERVIEWER NOTE: If you are describing SET I HYDROLOGICAL attributes, see (Appendix
III-B). If describing SET II WILDLIFE attributes, see Appendix III-C). The listing of the respective
attributes, as well as a written and graphical description of the attributes and their levels are provided
in NOTEBOOK form.
READ THE ATTRIBUTE LISTING AND EACH OF THE ATTRIBUTE DESCRIPTIONS WITH THE RESPONDENT,
HOLDING THE NOTEBOOK SO THAT BOTH THE RESPONDENT AND YOU CAN EASILY SEE IT. EACH
ATTRIBUTE DESCRIPTION CONSISTS OF TWO PARTS: A VERBAL DESCRIPTION AND A GRAPHIC
DESCRIPTION IN THE FORM OF A BAR CHART. READ THE WRITTEN DESCRIPTION OF EACH ATTRIBUTE
ALOUD, ALONG WITH THE CORRESPONDING ATTRIBUTE LEVELS GIVEN BELOW THE WRITTEN
DESCRIPTION. NEXT, REFER TO THE BAR CHART THAT GIVES A GRAPHICAL REPRESENTATION OF THE
LEVELS FOR THE ATTRIBUTE.
A-8
AFTER READING THE ATTRIBUTE DESCRIPTION AND SHOWING THE BAR CHART FOR EACH ATTRIBUTE,
ASK THE RESPONDENT IF HE OR SHE HAS ANY QUESTIONS.
AFTER COMPLETING THE SET OF ATTRIBUTE DESCRIPTIONS, GIVE THE RESPONDENT COPIES OF THE BAR
CHARTS (LAST PAGE OF NOTEBOOK) FOR EACH ATTRIBUTE TO REFER TO DURING THE REMAINDER OF
WATER MANAGEMENT PLAN SELECTION EXERCISE.
GO TO SECTION II-D.
SECTION II-D. EXAMPLE WATER MANAGEMENT PLAN PAIRS
INTERVIEWER NOTE: THE EXAMPLE PAIRS OF WATER MANAGEMENT PLANS WHICH
WILL BE USED FOR PRACTICE EXERCISES ARE DIVIDED INTO TWO SECTIONS.
SECTION 1 HAS THE SET I HYDROLOGICAL ATTRIBUTES AND THREE PAIRS OF
EXAMPLE PLANS TO CHOOSE FROM ARE SHOWN IN THE NOTEBOOK.
SECTION 2 INCLUDES SET II WILDLIFE ATTRIBUTES AND THREE PAIRS OF EXAMPLE
PLANS TO CHOOSE FROM ARE SHOWN IN THE NOTEBOOK.
THE THREE PAIRS OF EXAMPLE PLANS TO CHOOSE FROM FOR SET I AND SET II
ATTRIBUTES ARE PROVIDED IN THE NOTEBOOK.
WHETHER YOU ARE ADMINISTERING A QUESTIONNAIRE USING SET I
HYDROLOGICAL ATTRIBUTES OR SET II WILDLIFE ATTRIBUTES, START WITH
PRACTICE 1.
READ: We are now going to go through an example of a way of choosing between two alternative
water management plans. Recall what I said earlier that, like automobiles, two
management plans will differ according the effects each one will have on the levels of the
attributes. Each management plans has six attributes, as earlier described.
SHOW PRACTICE 1: CHOOSE BETWEEN MANAGEMENT PLANS 1 AND 2.
READ: As you can see, Column 1 lists the six attributes for each of the two management plans.
READ THE LIST OF ATTRIBUTES TO THE RESPONDENT.
READ: Column 2 shows the attribute level achieved with Management Plan 1, and Column 3
shows the attribute level achieved with Water Management Plan 2.
POINT OUT TO THE RESPONDENT THE SIMILARITIES AND DIFFERENCES IN THE
ATTRIBUTE LEVELS BETWEEN THE TWO MANAGEMENT PLANS. INDICATE WHICH
ATTRIBUTES ARE AT THEIR "PRESENT LEVELS" AND REMIND THE RESPONDENT
THAT HE/SHE CAN REFER TO THE BAR CHART GIVEN TO HIM/HER INDICATING THE
POSSIBLE LEVELS THE SIX ATTRIBUTES.
ASK THE RESPONDENT IF HE/SHE HAS ANY QUESTIONS.
READ: Please read the levels of the attributes that Plan 1 and Plan 2 can achieve and select the
Plan you most prefer. Take your time to make a selection. Tell me the number of the Plan
you most prefer.
AFTER THE RESPONDENT HAS SELECTED ONE OF THE PLANS CIRCLE THE RESPONSE
ON THE EXAMPLE PLAN ANSWER SHEET.
DEBRIEFING: ASK THE RESPONDENT WHY HE'SHE SELECTED A PARTICULAR PLAN?
WHAT ATTRIBUTES WERE IMPORTANT? THE PURPOSE OF THE DEBRIEFING IS TO
MAKE SURE THE RESPONDENT UNDERSTANDS WHAT THE DIFFERENT LEVELS OF
A-9
THE PLAN ATTRIBUTES REALLY MEAN (e.g. that restrictions on water use become
increasingly severe).
IF A RESPONDENT SELECTS A PLAN WHICH IMPLIES A HIGH LEVEL OF TAXATION,
LOSS OF AGRICULTURAL LAND, OR SEVERE RESTRICTIONS ON INDOOR AND
OUTDOOR WATER USE, ASK IF THIS IS THE CHOICE THE RESPONDENT REALLY
WANTS.
READ: We are now going to do a second example exercise.
SHOW EXAMPLE PLAN'S 3 AND 4.
READ: The format for Management Plans 3 and 4 are the same as the previous example.
However, there are now four attributes with different levels in Plans 3 and Plan 4.
INDICATE HOW THE ATTRIBUTES OF THE PLANS 3 AND 4 DIFFER. REMIND THE
RESPONDENT THAT HE/SHE CAN REFER TO THE BAR CHART HANDOUTS
DESCRIBING THE DIFFERENT POSSIBLE LEVELS OF ATTRIBUTES.
READ: Plans 3 and 4 are different from Plans 1 and 2. You are not to compare Plan 3 or Plan 4
with either Plan 1 or Plan 2. The only choice to make is between Plans 3 and 4. Please
read the levels of the attributes of each plan carefully and select the one you most prefer.
Tell me the number of the plan you most prefer. Please ask if you have any questions.
AFTER THE RESPONDENT HAS SELECTED ONE OF THE PLANS MARK THE CHOICE ON
THE EXAMPLE PLAN ANSWER SHEET.
DEBRIEFING: ASK THE RESPONDENT WHY HE/SHE SELECTED A PARTICULAR PLAN?
WHAT ATTRIBUTES WERE IMPORTANT? THE PURPOSE OF THE DEBRIEFING IS TO
MAKE SURE THE RESPONDENT UNDERSTANDS WHAT THE DIFFERENT LEVELS OF
THE PLAN ATTRIBUTES REALLY MEAN (e.g. that restrictions on water use become
increasingly severe).
IF A RESPONDENT SELECTS A PLAN WHICH IMPLIES A HIGH LEVEL OF TAXATION,
LOSS OF AGRICULTURAL LAND, OR SEVERE RESTRICTIONS ON INDOOR AND
OUTDOOR WATER USE, ASK IF THIS IS THE CHOICE THE RESPONDENT REALLY
WANTS.
STOP HERE AND PROCEED TO THE ACTUAL CHOICES (A thru N IF THE RESPONDENT
SHOWS THAT HE/SHE UNDERSTANDS THE TASKS AND TRADEOFFS FOR CHOOSING
A MANAGEMENT PLAN. IF NOT, CONTINUE ON AND DO THIRD EXAMPLE 3.
READ: We will now do a third example selection.
SHOW EXAMPLE PLANS 5 AND 6
READ: The format for the presented plans is the same as the previous two examples. Plan 5 and
Plan 6 are different from one another because some of the levels for some attributes are
different. Do not compare these plans with the previous example plans. Choose only
between Plans 5 and 6. In this example, the levels of four attributes are different, but the
differences are not the same as in the previous example.
INDICATE TO THE RESPONDENT HOW THE ATTRIBUTE LEVELS DIFFER BETWEEN
THE TWO PLANS. REMIND THE READER THAT HE/SHE CAN REFER TO THE BAR
CHART HANDOUTS IN MAKING A CHOICE.
READ: Please read the levels for each of the components of each Plan carefully and select the Plan
you most prefer. Tell me the number of the Plan you most prefer. Please ask if you have
any questions.
A-10
AFTER THE RESPONDENT HAS SELECTED ONE OF THE PLANS MARK THE ANSWER
ON THE EXAMPLE ANSWER SHEET.
GO TO SECTION II-E.
SECTION II-E. SELECTION OF
WATER/WILDLIFE MANAGEMENT PLANS
INTERVIEWER NOTE: THERE ARE TWO BASIC SETS OF WATER MANAGEMENT PLAN
SELECTION EXERCISES, CORRESPONDING TO THE HYDROLOGICAL AND WILDLIFE
ATTRIBUTES. EACH SET IS COMPOSED OF TWO FORMS. EACH FORM HAS SEVEN
PAIRS OF MANAGEMENT PLANS. THE DIFFERENCE IN THE TWO FORMS FOR EACH
SET IS THE ATTRIBUTE LEVELS. ONLY ONE FORM WILL BE ADMINISTERED FOR AN
INDIVIDUAL RESPONDENT. THE FORM TO BE ADMINISTERED WILL BE DETERMINED
PRIOR TO THE INTERVIEW. THE TWO SETS AND FOUR FORMS INCLUDE:
SET I: HYDROLOGICAL FORM 1 PLANS SET II: WILDLIFE FORM 1 PLANS.
SET I: HYDROLOGICAL FORM 2 PLANS SET II: WILDLIFE FORM 2 PLANS.
WITHIN EACH FORM THERE ARE SEVEN PAIRS OF WATER MANAGEMENT PLANS, STARTING WITH PAIR
A-B AND GOING THROUGH M-N. AS IN THE PRACTICE EXERCISE, YOU ARE TO ASK RESPONDENTS TO
CHOOSE ONE OF TWO PRESENTED PLANS.
READ: I am now going to ask you to choose between two alternative Management Plans. Each
Plan is a unique combination of the levels for each of the six attributes described earlier.
The first set is made up of Plans A and B.
SHOW PLANS A AND B.
READ: Please read the levels of each of the levels of the Plan A and B carefully, and tell me the
letter of the Plan you most prefer. Take your time and refer to the bar charts as necessary.
Please ask if you have any questions.
WAIT UNTIL RESPONDENT HAS SELECTED ONE OF THE PLANS AND CIRCLE THE PLAN INDICATED BY THE
RESPONDENT ON THE PLAN ANSWER SHEET. INDICATE THE SET NUMBER ADMINISTERED AT THE TOP OF
THE ANSWER SHEET.
READ: I am now going to give you the second set of management plans to consider. These are
Plans C and D.
SHOW PLANS C AND D.
READ: These are plans C and D, and the levels for each of these plans also differ. Plans C and D
are different from Plans A and B. You are choosing only between Plans C and D and are
not to compare them to Plans A and B. Please consider each plan carefully and go at your
own pace. Tell me the letter of the Plan you most prefer. If you have any questions,
please ask.
WAIT UNTIL THE RESPONDENT HAS SELECTED ONE OF THE PLANS AND CIRCLE THE PLAN INDICATED BY
THE RESPONDENT ON THE CODING ANSWER SHEET.
READ: I am now going to ask you to select a management Plan from a few more sets of Plans,
starting with E and F. Remember, you are selecting only between these two Plans and are
not to compare them with previous Plans. Once you have selected a Plan, tell me the letter
of the Plan you most prefer. As you go along, please ask me any questions you may have.
A-Il
SHOW PLANS E AND F. AFTER RESPONDENT HAS CHOSEN PLAN E OR F AND CIRCLE THE PLAN
INDICATED BY THE RESPONDENT ON THE PLAN ANSWER SHEET. CONTINUE WITH ALL REMAINING PLANS
THRU M AND N. CIRCLE THE PLAN INDICATED BY THE RESPONDENT ON THE PLAN ANSWER SHEET.
PROCEED TO SECTION II-F.
SECTION II-F. SOCIO-DEMOGRAPHICS
F-1. READ: I have a few final questions for you so that can analyze your answers along with the
answers of others.
Have you voted in a state or local election within the past 3 years?
Yes 1
No 2
Don't know 88
F-2 What was the highest grade or year in school you completed?
0-8 years 1
Some high school 2
Completed high school 3
Some college 4
Completed college 5
Graduate or professional school 6
Don't Know 88
F-3 Were you born in Florida
Yes 1
No 2 4 Go to F-3a
F-3a How many years have you lived in Florida?
0-99
-88 Refused
F-4 What year were you born? (CODE LAST TWO YEARS ONLY e.g. 1950 is 50)
0-80
88 Refused
F-5 Including yourself, how many people live in your household?
1-20
88 Refused
F-6 How many children under age 18 do you have living with you?
0-10
88 Refused
SHOW RESPONDENT CARD E.
F-7 Please indicate how you would you describe your racial or ethnic background?
White (Non-Hispanic) 1
Black (Non-Hispanic) 2
A-12
White (Hispanic) 3
Black (Hispanic) 4
American Indian 5
Asian 6
Pacific Islander 7
Other 8
Refused 88
F-8 Generally speaking, do you usually think of yourself as a Republican, a Democrat, an
Independent, or what?
Republican 1 -= Go to F-9
Democrat 2 Go to F-10
Independent 3 D Go to F-11
Other party 4 = Go to F- 1
No preference 5 =P Go to F-11
F-9 Would you consider yourself a strong, moderate or a not very strong Republican?
Strong 1
Moderate 2
Not very strong 3
Refused 88
F-10 Would you consider yourself a strong, moderate or a not very strong Democrat?
Strong 1
Moderate 2
Not very strong 3
Refused 88
F-11 Do you make donations to environmental groups?
Yes 1 ~ Go toF-12
No 2 -: Go toF-13
Refused 88 Go to F-13
F-12 On average how much do you donate annually?
Less than $100 1
Between $100 and $500 2
More than $500 3
Refused 88
F-13 Now consider your family's household income from all sources. Here is a card that lists
various categories of income.
SHOW CARD F. CATEGORIES OF INCOME.
Please indicate the number of the category that best describes your total household income in 1996.
Less than $10,000 1
$10,000 to $20,000 2
$20,000 to $30,000 3
$30.000 to $40,000 4
$40,000 to $50,000 5
$50,000 to $60,000 6
$60,000 to $80,000 7
$80,000 to $100,000 8
A-13
over $100,000
Refused
THIS IS THE END OF THE SURVEY. THANK THE RESPONDENT FOR HIS/HER TIME AND
COOPERATION.
SECTION II-G. INTERVIEW EVALUATION QUESTIONS
INTERVIEWER NOTE: PLEASE ANSWER QUESTIONS G-1 THROUGH G-4.
G-1. What was the reaction of the respondent as you read through the attribute
includes the descriptive material, including the bar charts).
EXTREMELY VERY
SOMEWHAT
SLIGHTLY
descriptions (this
NOT AT NOT
ALL
a. How distracted was
the respondent
b. How attentive was
the respondent
c. How well did the
respondent understand
the material
G-2. Did the respondent say anything suggesting that he or she had any difficulty understanding
either the attributes or the attribute levels.
1 YES (if YES, go to question G-2a).
2 NO
G-2a. Describe the difficulties.
G-3. Did the respondent have
management or wildlife plan?
any difficulty understanding the process of choosing a water
1 YES (if YES, go to question G-3a).
2 NO
G-3a. Describe the difficulties.
A-14
SURE
G-4. How serious was the consideration the respondent gave to the decision about what water
management or wildlife plan to choose?
1 Extremely Serious
2 Very Serious
3 Somewhat Serious
4 Slightly Serious
5 Not At All Serious
6 Not Sure
A-15
' UNIVERSITY OF
FLORIDA
INTERVIEW CONSENT
I have received a copy of the interview consent
form and I agree to participate in this interview.
NAME
DATE
A-16
___ I_ __ __ ____ ;
___
--- ------
i
j
I
i ---
i
- ------~-~
r
CARD A
Programs to Combat Crime
Public Schools (K-12)
Protect the Environment
Industrial Development and Attract New
Industry
Acquire land to Protect Endangered
Species
Health Care Services
Spending for Colleges and Universities
Highways and Road Systems
Low Income Families with Children
State Spending for the Elderly
State Prisons and Correctional Facilities
State Spending to Promote Tourism
A-17
CARD B
Increase A Lot
Increase Somewhat
Same
Decrease Somewhat
Decrease A Lot
A-18
|
