.3. 5:22?
:5. . 1?. ,

‘ .K.‘
. .3122...
A A.

 

I .2 ,2 ‘ . , . ‘ .

.y 21.55:?! . , . . . l I
5:... .‘ .. . . ‘ . . .7 .
zg t a ‘ . ‘ . , z- to. u“

. «.1. Ida, » .1 8

.I .o.

dill

Est;

:1}.
3.! 1;. i
, .

51.001. ..
.. Via: Mn.

 

«a... ll“
{\9! .

 

 

K‘““""

 

3 LIBRARY
20"” Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

PUBLIC PREFERENCES AND NONMARKET VALUES
FOR THE MANAGEMENT OF FOREST ECOSYSTEM
SERVICES

presented by

LAILA ANNA RACEVSKIS

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Agricultural Economics

 

 

/ / <
M

 

Major Professor's Signature

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

.. -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-_,-.-.- -.-.--.-

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

MRHiWO

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:lClRCIDatoDue.lndd-p.1

 

 

PUBLIC PREFERENCES AND NONMARKET VALUES FOR
THE MANAGEMENT OF FOREST ECOSYSTEM SERVICES

By

Laila Anna Racevskis

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Agricultural Economics

2005

ABSTRACT

PUBLIC PREFERENCES AND NONMARKET VALUES FOR
THE MANAGEMENT OF FOREST ECOSYSTEM SERVICES

By
Laila Anna Racevskis

This research uses a variety of techniques to investigate public preferences for
and attitudes towards the management of forest ecosystem services among Michigan
residents in order to provide information useful to natural resource managers and policy
makers. The study focuses on a forested area in Michigan’s Upper Peninsula to analyze
public preferences for forest management in this region. The study forest is a timber
dependent region with a history of intense timber harvesting. This area is also
characterized by high winter deer densities that are potentially affecting tree regeneration
and by the existence of important habitat for migratory forest songbirds. The interactions
between deer populations and timber harvesting activities have adversely affected habitat
for these songbirds. The presence of many complex interactions between humans, forests
and wildlife in this area makes it an ideal setting in which to examine public preferences
for tradeofi‘s between various forest ecosystem services.

The research is divided into four essays. Essay 1 is based on the results of focus
groups that were conducted as part of the qualitative research design of the survey
instrument. Analyses in essays 2, 3 and 4 are based on data collected from a mail survey
of 2,000 Michigan residents, which collected data on forest management and
environmental attitudes, demographic characteristics and stated preference contingent
valuation data.

The first essay uses qualitative research techniques to analyze the results of a

series of focus group discussions held in and near the study forest. This research finds
that residents of rural, timber-dependent communities do not hold purely anthropocentric
forest management values and urban, non-timber dependent residents do not hold purely
biocentric values.

The second essay uses factor analysis to identify an underlying structure to
environmental and forest management attitude data and uses resulting factor scores in a
regression analysis of factor scores on a set of demographic characteristics. This study
concludes that environmental and forest management attitudes can be reduced into a set
of 5 factors, which can, in turn, be interpreted according to an anthropocentric/biocentric
value framework. Regression results show that area of residence, age, membership in an
environmental organization and political views can explain factor scores that reflect
particular value orientations.

The third and fourth essays utilize the attribute-based referenda (ABR) model, a
variant of the contingent valuation method (CVM) to estimate nonmarket values for
forest ecosystem attributes of the study area. An orthogonal main effects design was
used to create choice sets presented to survey respondents. A dichotomous choice
referendum question elicited stated preference data used to estimate a series of random
effects probit models used to estimate willingness to pay (WTP) for a possible forest
easement program. Essay 3 estimates the marginal effects and marginal dollar values of
forest easement program attributes as well as the effects of a set of demographic
characteristics on WTP. Essay 4 estimates the effects of distance, environmental

attitudes and recreational use on WTP.

Copyright by
LAILA ANNA RACEVSKIS
2005

To my husband, Wolfgang, my parents, Maija and Karlis Racevskis,

my grandmother, Austra Lidacis, and in memory of my grandfather, Arvids Lidacis '

ACKNOWLEDGMENTS

Many people have helped me in the completion of this dissertation, and I would
like to acknowledge and thank them here. I would like to thank Frank Lupi, my
academic advisor, for being a wonderful mentor and for his constant support throughout
this process. His expert guidance and generosity with his time and ideas have enriched
my graduate school experience and have contributed greatly to my personal and
professional development over the years. I feel very fortunate to have had an advisor
whose help, kindness, patience and concern for my well being helped make my doctoral
program an enjoyable and rewarding experience. I would also like to thank the members
of my committee, Sandra Batie, Pat Norris and Karen Potter-Witter for the valuable
contributions they have made to help improve this work and for the support and
encouragement they have all given me along the way.

I would like to thank the Michigan Department of Natural Resources, USDA
National Research Initiative of the USDA Cooperative State Research, Educational and
Extension Service grant #MICL08292, Michigan State University Department of
Agricultural Economics, the Michigan State University Environmental Science and
Policy Program and the Michigan State University Graduate School for their financial
support that contributed to the completion of this work.

I am very grateful to Jack Liu, Mike Walters, Haijin Shi, Ed Laurent, Joseph
LeBouton and Kim Hall, the members of the Ecological Economics research team. Their
dedication to our team project was instrumental to the completion of my work. I would
especially like to thank Ed, Joseph and Kim for spending countless hours discussing
forests, deer and birds and for taking the time to review and provide feedback on many

vi

survey drafts. I greatly appreciate their time and efforts.

I would like to express my appreciation to the staff of the Department of
Agricultural Economics for all of their help throughout my time here. I would especially
like to thank Ann Robinson for her generosity in providing me with office space and
resources to help with my survey mailings.

Many thanks go to fellow graduate students, past and present. In particular, I
would like to thank Wei Zhang, David Mather, Tsitsi Makombe, Meera Manojkumar,
Miguel Zamora and Kathryn Moeller, Brady Deaton, Meeta Punjabi, Kofi Nouve and
Asfaw Negassa for their friendship, support and encouragement over the years. I would
also like to thank Melissa Gibson and Oscar Arreola for taking the time to help me with
the survey implementation process.

I thank my friends who are truly like family, Elaine Zamonski, Rebecca Watson,
Andria Shaman and Ellen Sturm for believing in me, laughing with me and keeping me
sane. I would also like to thank Zara Bennett, Michelle McGreevy, Mehnaz Safavian and
Maria Pagura for their constant support and for inspiring me to reach this goal.

I am extremely grateful to my family for all of their love, support and
encouragement. My parents deserve countless thanks for their unwavering and unending
support, devotion and love, for all the trips to East Lansing, for providing a sympathetic
and understanding car at all hours of the day and night, and for being a constant source of
inspiration to me. In remembrance of my grandfather, Arvids Lidacis, I would like to
thank him and my grandmother, Austra Lidacis, for their love, sacrifices, wisdom and
kind words that continue to inspire me in everything I do. I thank my brothers, Andrew
and Roland, my sisters-in-law, Megan and Lynnette, and my beautiful nieces, Sofia and
Rosa, for all the crazy fun, joy and laughter they have brought to my life. My deepest

vii

gratitude goes to my husband Wolfgang, for his belief in me, his eternal optimism and for

being there for me through it all.

viii

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Introduction ....................................................................................... 1
References . ...................... lO
Essay 1: Comparing Urban and Rural Perceptions of and Familiarity with the
Management of Forest Ecosysems ..1 l
1.1 Introduction and Rationale for Research - - ............... l 1
1 .2 Study Area 14
1..2 1 Ecological characteristics of the study area 15
1..2 2 Economic and demographic characteristics of the study area
and Marquette 16
1.3 Methods -. 18
1.4 Results and Discussion _- _ - .......... 21
1.4.1 Focus group participant characteristics . - 21
1.4.2 Focus group participant comments ....... 22
1.4.2.1 Forest services - 24
1.4.2.2 Forest management 28
1..4 2. 3 Land ownership . 30
1..4 2. 4 Forest management and wildlife interactions ............................ 30
1.5 Conclusions 33
References ......... - 36
Essay 2: Forest and Environmental Values and Attitudes: Differences between
Demographic Groups in Michigan - - - 41
2.1 Introduction ................................................... 41
2. 2 Methods ...... - ............................................. 45
2.2.1 Survey design and implementation ........................................................ 45
2 ..2 2 Factor analysis and factor score regression .......................................... 46
2.3 Factor analysis results ................................................................................... 49
2.4 Factor score regression results ....................................................................... 59
2.5 Discussion - ................................................................. 62
2.6 Conclusions -- - - ................................................ 63
References ......... - ................... 66
Essay 3: Valuing Forest Ecosystem Characteristics in Michigan’s Western Upper
Peninsula: An Attribute-Based Contingent Valuation Approach - ...-69
3.1 Introduction ........ 69
3. 2 Attribute-Based Referenda Model ...................................... . 71
3.2.1 Model Specification ............................................................................. 77
3.3 Welfare estimates .......................................................................................... 80
3.4 Survey design ................................................................................................ 82
3.4.] Economic attributes .............................................................................. 85
3.4.2 Birds ...................................................................................................... 87
3.4.3 Deer ....................................................................................................... 88

ix

3.5 Survey implementation .................................................................................. 89

 

3.6 Model estimation results ............................................................................... 93
3.7 Welfare analysis and discussion ................................................................... 99
3.8 Conclusions ................................................................ 102
References _ -_ -- __ - _ .- _ 104

 

Essay 4: Effects of Distance, Attitudes and Recreational Use on Willingness to Pay for

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Forest Ecosystem Services _ _- _- _ 108
4.1 Introduction - ........................................ 108

4.2 Methods ............................................... 110

4.2.1 Survey design and implementation ....... ....... 110

4.2.2 Theoretical model ..... 111

4.2.3 Model specification .................................................. 115

4.3 Model estimation ........................................................ 115

4.3.1 Distance model - -- 115

4.3.2 Distance and attitudes models 117

4.3.3 Distance, recreational use and attitudes models ....... 126

4.4 Conclusions 134
References . .......... 137
Conclusions ..................... _ ............................................................. 140
Appendix 1: Sample Design ........................................................................................... 150
Appendix 2: Survey Design ............................................................................................ 157
Appendix 3: Survey Instrument ...................................................................................... 170
Appendix 4: Survey Correspondence . ........................ - 191
Appendix 5: Survey Implementation .............................................................................. 196
Appendix 6: Survey Results - Descriptive Statistics ...................................................... 202
Appendix 7: Survey Results - Descriptive Statistics by Region ..................................... 211

LIST OF TABLES

Table 1.1 Socioeconomic Characteristics of Urban and Rural Focus Group

Participants ....................................................................................................................... 22
Table 1.2 Theme and Sub-Theme Frequencies and Comment

Extensiveness of Urban and Rural Focus Groups ........................................................... 24
Table 2.1 Survey Attitude Statements and Associated Variables Used in

Factor Analysis ................................................................................................................. 50
Table 2.2 Total Variance Explained ................................................................................ 52
Table 2.3 Rotated Factor Matrix ...................................................................................... 54
Table 2.4 Summary of Factors and Associated Variables ............................................... 58
Table 2.5 Definitions of Variables Used in Factor Score Regression Models ................ 59
Table 2.6 Factor Score OLS Stepwise Regression Results ................................ 60
Table 3.1 Variable Definitions ........................................................................................ 79
Table 3.2 Survey Attributes and Levels .......................................................................... 84
Table 3.3 Response Frequencies for Respondent Evaluation of Attribute

Importance to Choice ...................................................................................................... 92
Table 3.4 Estimation Results from Random Effects Models .......................................... 97
Table 3.5 Implicit Prices of Attributes ............................................................................ 100
Table 3.6 Marginal Rates of Substitution Between Nonmonetary Program

Attributes ......................................................................................................................... 101
Table 4.1. Survey Attributes and Levels ....................................................................... 111
Table 4.2. Distance Model Estimation Results ................................................................ 1 16
Table 4.3. Definition of Distance Zones ........................................................................ 118

Table 4.4. Proportions of Recreational Users and Non—users Within
Different Distance Zones ................................................................................................... 119

xi

Table 4.5. Cross-tabulation of Recreational Users' Forest Management
Attitudes Compared Between Distance Zones . - ...................... 120

 

 

Table 4.6. Attitude Variables Included in the Calculation of biocentric .- 122

Table 4. 7. Random Effects Probit Estimation Results of Distance

 

 

 

 

 

 

and Biocentric Models.-- ................ 125
Table 4.8. Variable Definitions for Recreational Use Models .......................................... 126
Table 4.9. Results of Recreational Use and Attitudes Models .......................................... 130
Table 4.10. Implicit Prices of Model Attributes ............................................................. 131
Table 4.11 Relative WTP of hs, other and allrec Users and Non-Users at

Different Distances from the Resource and Difi‘erent Levels of Biocentrism ................ 133
Table A1.1 Sample Frame for Stratum 1: Study Area .................................................... 153
Table A1 .2 Sample Frame for Stratum 2: Rest of UP .................................................... 153
Table A1.3 Sample Frame for Stratum 3 : Northern Lower Peninsula (NLP) ............... 155
Table A1.4 Sample Frame for Stratum 41 Southern Lower Peninsula (SLP) ................. 156
Table A2.1. Study Area Counties and Townships .......................................................... 159
Table A2.2. Study Area Township Service Sector Employment

and Estimated Tourism Spending in 2000 - - ........................... 160
Table A2.3. Study Area Zip Codes .......................................................... - - -162
Table A2.4. Forest Industry Employment Subsectors Used in

Calculation of Forest Industry Jobs in the Study Area ................................................... 163
Table A2.5. Forest Industry Employment in 2001 in Study Area Zip Codes ................ 164
Table A2.6. Migratory Forest Songbird Species of Conservation Conem

in the Study Area ....... 166
Table A5.1. Disposition of Survey Mailings - ..... - . ........... 199
Table A5.2. Regional Response Rates .................. -- .......... 199
Table A5.3. Detailed AAPOR Disposition Codes for Each Survey Mailing ................ 200
Table A6.1 Socioeconomic Characeristics of Survey Respondents .............................. 203

xii

Table A6.2 Response Frequencies for Forest Management and

 

 

Environmental Attitude Statements - ......................................................... 206
Table A6.3 Response Frequencies for Recreational Activities ...................................... 209
Table A6.4 Other Response Frequencies ............................................................................ 210
Table A7.1 Regional Response Frequencies for Forest Management

Environmental Attitude Statements - _ .......... - - .212
Table A7.2 Socioeconomic Characteristics of Survey Respondents by Region ............. 223

xiii

LIST OF FIGURES

 

 

Figure I - Ecosystem Valuation Framework - ...................................... 3
Figure 1.1 - Map of Michigan with Study Area Highlighted ............................................. 15
Figure 2.1 - Cognitive Hierarchy Framework ............... .. ................. 43
Figure 2.2 - Scree Plot of Eigenvalues and Factors ........................................................... 53
Figure 3.1 - Graph of Bid Distribution - - ........................................... 93

 

xiv

Introduction

Ecosystems provide a wide variety of services that benefit human beings,
including the production of food, fiber and water as well as other services, including
recreation, cultural values, soil building and many others (Daily 1997, Heinz 2002, MA
2003). Achievement of the production of ecosystem services, such as food and fiber,
often occurs at a cost to the provision of other services, such as wildlife habitat or water
quality (MA 2003). Humans are dependent on the services provided by ecosystems (MA
2003), and many ecosystem services are currently being degraded or are being used in an
unsustainable way (MA 2005). It is important to understand the benefits of ecosystem
goods and services to society in order to effectively evaluate the tradeoffs that are made
between the provision of different types of ecosystem goods and services (NRC 2005).
Some ecosystem services, such as the production of food and fiber, can be quantified by
placing a monetary value on them, but there are many other ecosystem goods and
services that are very important to society, the benefits of which have not been quantified
extensively (MA 2005).

Ecosystems are valued by people because they provide services that satisfy
material and nonmaterial human needs (MA 2003). Many methods have been developed
by numerous disciplines to conceptualize the values that ecosystems provide to humans
(ibid). Ecosystem functions, when conceptualized in terms of the benefits they provide to
humans, are referred to as ecosystem goods and services (de Groot et a1. 2002). The
values that these ecosystem goods and services provide to people can be placed into three

categories:

l. Ecological values, which include criteria used to determine the capacity of

ecosystems to provide these goods and services,

2. Sociocultural values, which include the social values and perceptions that

determine the importance of ecosystems to people, and

3. Economic values, which include the market and nonmarket values of

ecosystem services (de Groot et al. 2002).

Human actions affect the structure and function of ecosystems, which affect
provision of ecosystem goods and services (NRC 2005). Humans have values for these
goods and services, and these values, in turn, affect human behavior and action, which
feeds back to affect ecosystem structure and fimction (ibid). This cyclical framework is
illustrated in Figure 1. Human well being is affected by the ways in which ecosystem
goods and services are managed, and the ability to make informed management decisions
relies on information on ecosystem patterns and conditions as well as economic, political,
social and cultural consequences of management decisions (MA 2003). These
relationships illustrate the importance of incorporating knowledge of ecology, economics

and human values in natural resource allocation decisions.

Figure I. Ecosystem Valuation Framework

 

          
     

 
 

Ecosystem Structure Ecological
& Functron Production Function
Human Actions Ecosystem Goods
(Private/Public) & Services
Valuation
Approach

 

Ecological, Sociocultural
And Economic Values

 

 

 

Adapted fi'om NRC 2005

In order to assess the consequences of resource management decisions on
ecosystem service provision, economic valuation methods are used as a means of
comparing ecosystem services using a common metric (MA 2003). Economic valuation
is based in the utilitarian paradigm, within which the concept of value is based on the
idea that individuals derive utility from given ecosystem services (ibid). However,
individuals may also hold sociocultural values for ecosystem services that are not easliy
measured using the monetary metric of economic valuation methods (ibid). This research
investigates both economic and non-economic (sociocultural) values that people hold for
ecosystem goods and services.

Many different views exist across different disciplines on how to conceptualize
the values derived from ecosystems (NRC 2005). Within the social forestry literature,

the anthropooentric/biocentric continuum has been used to conceptualize the range of

3

values that people hold with respect to the management of forest ecosystem services.
Anthropocentric values in this context refer to values that emphasize the instrumental and
extractive uses of forest ecosystems for human benefit, whereas biocentric values
emphasize the intrinsic value of forest ecosystems and the importance of their use for
non-humans (Tarrant and Cordell 2002). This research implements this
conceptualization of forest ecosystem value orientations and further bases the analysis in
a cognitive hierarchy framework, in which values underlie attitudes and attitudes affect
behavior (McFarlane and Boxa112000).

Natural resource management and policy making utilize information on public
preferences for resource management. Characterizing segments of the public that tend to
hold particular resource management views can contribute to an increased understanding
of individuals’ motivations, behavior and acceptance of natural resource policies. This
work is presented as a suite of analyses that explore Michigan residents’ views, attitudes
and preferences related to the management of forest ecosystem services in an area of
Michigan’s Upper Peninsula. The analyses contribute to understanding the characteristics
of individuals that can assist in predicting acceptability of certain policy initiatives
among different segments of the population. This information can be a useful input to
educational efforts aimed at disseminating information on natural resource management
efforts as well as to natural resource policy design. This work contributes to the
literature on the integration of ecological and economic analyses in order to better
understand the connections between human actions, ecological consequences and the
values generated by ecosystem goods and services.

This research focuses on an area of Michigan in which forests are increasingly

being managed for multiple uses. Multiple use forest management decisions involve

4

many difi‘erent stakeholders who frequently have conflicting interests. For example,
resource allocation decisions require making trade-offs between managing for timber
production versus managing for improving wildlife habitat. Forests provide services that
are bought and sold. in the market (e. g. timber), and they also generate benefits that are
not traded in the market and therefore do not have market values associated with them
(e.g. wildlife habitat). A principal motivation of this research was to estimate the
nonmarket values generated by forest ecosystem services in this area of Michigan and to
gain an understanding of public views of and preferences for the tradeoffs involved in
multiple use forest management. In addition, a goal of the research was to investigate
public attitudes towards forest management in this region and to establish connections
between attitudes, demographic characteristics and nonmarket values.

To assess public preferences for multiple use forest management in an area of
Michigan’s Upper Peninsula, this study uses results from a mail survey of 2,000
Michigan residents to analyze the relationships between demographic characteristics,
environmental and resource management attitudes, and nonmarket values of forest

ecosystem attributes. The research was guided by the following objectives:

1. To gain a better understanding of public views and perceptions of forest-
human-wildlife interactions in Michigan’s western Upper Peninsula;

2. To understand how demographic characteristics and environmental and
forest management attitudes are related;

3. To estimate non-market values of forest ecosystem characteristics in

Michigan’s western Upper Peninsula; and

4. To determine how distance, attitudes and recreational use affect the results
of nonmarket valuation analysis of forest ecosystem attributes in

Michigan’s western Upper Peninsula.

The study area consists of about 500,000 ha of forested land in Michigan’s
western Upper Peninsula and includes parts of Baraga, Dickinson, Iron, Marquette, and
Menominee counties (See Figure 1.1 in Essay 1 for a map of the study area). This
research forms part of a larger project conducted by ecologists and foresters within the
same defined study area. The development of the survey used to collect socioeconomic
and non-market value data was aided by the work of ecologists done in the same location.
These researchers have collected data on tree sapling regeneration, deer densities, and
bird species diversity and have estimated the effects of forest management on these
variables as well as the effect of deer populations on tree regeneration and bird diversity.
The ability to integrate ecological and economic information in the same study area
makes this project a unique opportunity to explore the economic and ecological effects of
different forest management practices.

This region was chosen for its diverse ecological landscape, for its variety of
industrial and recreational uses of its forests, and for the presence of various forest
management methods. There are several issues of concern in this area relating to forest
and wildlife management. For example, there are concerns that timber harvesting may be
affecting wildlife habitat by removing cover in some areas through practices such as
cleareutting. There are also concerns that high deer densities are affecting regeneration of
commercially valuable tree species. The deer population, by reducing tree regeneration,

may be adversely affecting the habitat for other wildlife, such as certain types of

6

songbirds. The effects of deer on forest regeneration may decrease future timber
availability for the forest products industry as well as change the structure and
composition of forested areas used for recreation, which could impact local economies.

The effects of forest management practices on forest structure and on deer
populations, as well as the effects of deer on forest structure and wildlife, illustrate a few
of the many complex interactions within a managed forest ecosystem. Forests provide
timber, wildlife habitat, recreation, aesthetics, clean air, water filtration, and other
ecosystem functions. Forest management decisions result in different levels of these
market and non-market outputs, and trade—offs must be made between competing needs
of different stakeholders. Understanding the preferences and attitudes of the public
toward management of forest resources is an important component of forest policy
making. The study area used in this project provides an appropriate setting in which to
examine public preferences for multiple use forest management.

The research is presented in the form of four separate essays, which together have
a common focus in that they investigate individuals’ preferences for the management of
forest resources. Research methods used in the essays include qualitative and
quantitative ones and include several different analytical methods. The first essay
employs content analysis of focus group data, the second uses factor and regression
analysis and the third and fourth essays implement contingent valuation methodology.
The essays together contribute to understanding the ways in which individuals’
characteristics influence preferences for natural resource management.

The first essay, which is in press at the journal Society and Natural Resources,
reports the results of a series of focus groups held in and near the study area. This paper

focuses on the differences in perceptions of and familiarity with the management of

7

forest ecosystems in the study area between urban and rural demographic groups. The
study examines how urban and rural residents’ views fall along an
anthropocentric/biocentric continuum. This work contributes to understanding the ways
in which rural, timber dependent and urban, non-timber dependent community residents
view natural resource management.

The second essay provides a deeper exploration of the range of anthropocentric to
biocentric values that underlie individuals’ environmental and forest management
attitudes. This study uses attitudinal data collected from the mail survey to conduct
factor and regression analysis. Factor analysis is used to identify the underlying structure
of the attitudinal data and to identify whether the attitudinal data can be conceptualized
using the anthropocentric/biocentric continuum framework. The results of the factor
analysis are used as inputs to OLS regression of respondent demographic characteristics
on factor scores that reflect a range of anthropocentric to biocentric attitudes towards the
environment and forest management. The analysis in this essay contributes to
understanding the environmental and forest management attitudes of different
demographic groups and has implications for the types of values (anthropocentric versus
biocentric) held by different segments of the population.

The third essay presents a nonmarket valuation analysis of major forest ecosystem
services in the study area. The analysis uses an attribute-based referenda framework
based in random utility theory. The objective of this study is to estimate nonmarket
values for the forest ecosystem attributes of possible forest easement programs in the
study area. The attribute-based framework allows estimation of the tradeoffs individuals
are willing to make between different forest ecosystem attributes. The analysis also

investigates the demographic characteristics that have an effect on willingness to pay.

8

The fourth essay also employs nonmarket valuation of forest ecosystem attributes,
utilizing the same theoretical framework established in Essay 3. This essay explores the
effects of distance, attitudes and recreational use on willingness to pay. This research
contributes to the literature on defining the extent of the market for environmental goods
and services in nonmarket valuation literature, as well as the literature on incorporating

attitudinal data in nonmarket valuation studies.

REFERENCES

Daily, G.C. 1997. Introduction: What are ecosystem services? In Nature ’s Services:
Societal Dependence on Natural Ecosystem, edited by G. C. Daily. Washington,
DC: Island Press.

de Groot, R.S., M.A. Wilson, and R.M.J. Boumans. 2002. A typology for the
classification, description and valuation of ecosystem functions, goods and
services. Ecological Economics 41 :393-408.

Heinz (The H. John Heinz III Center for Science, Economics and the Environment).
2002. The State of the Nation 's Ecosystems: Measuring the Lands, Waters, and
Living Resources of the United States. Cambridge, UK: Cambridge University
Press.

MA (Millenium Ecosystem Assessment). 2003. Ecosystems and Human Well Being: A
Framework for Assessment. Washington, D.C.: Island Press.

 

. 2005. Ecosystems and Human Well Being: Synthesis. Washington, D.C.: Island
Press.

McFarlane, BL. and P.C. Boxall. 2000. Factors Influencing Forest Values and Attitudes
of Two Stakeholder Groups: The Case of the Foothills Model Forest, Alberta,
Canada. Society and Natural Resources 13:649-661.

NRC (National Research Council). 2005. Valuing Ecosystem Services: Toward Better
Environmental Decision Making. Washington, DC: The National Academies
Press.

Tarrant, M., and H. Cordell. 2002. Amenity values of public and private forests:

Examining the value-attitude relationship. Environmental Management 30
(5):692-703.

IO

Essay 1:
Comparing Urban and Rural Perceptions of and
Familiarity with the Management of Forest Ecosystemsl
1.1 Introduction & Rationale for Research

The focus of forest management on public and private lands has changed
considerably over the past decade due to increased public concern for wildlife, recreation
and aesthetics (Steel et al. 1994, Tarrant and Cordell 2002). As a result, management of
forests for multiple uses has emerged as an important goal for forest land managers.
Multiple use forest management involves trade-offs between frequently conflicting forest
management goals. For example, the resources provided by forest ecosystems generate
both market (e.g. timber) and non-market (e.g. wildlife) values to individuals. This
research uses the results of six focus group discussions to understand how the public
perceives these tradeoffs and to examine differences between urban and rural groups in
an area of Michigan’s Central Upper Peninsulaz.

Information on public preferences for forest management is critical for effective
resource management decision making (Bingham et al. 1995, Boxall and Macnab 2000,
Dennis 1998, Jacobson and Marynowski 1997, McFarlane and Boxall 2000, Stein et al.
1999, Tarrant and Cordell 2002, Zinkhan et al. 1997). Social acceptability is important to
the forest policy-making process (Kearney 2001), and clear communication of
management options to the public can help increase public acceptability of forest policies
(Shindler et al. 2002). This communication process begins by understanding public

preferences for and knowledge of different forest management alternatives, and this

 

' This essay is currently in press at Society and Natural Resources
2 This area is referred to in this essay as the “study area,” while in Essays 2, 3 and 4 it is referred to as the
“study forest.”

ll

process is a key to connecting management regimes with publicly acceptable ecosystem
outcomes.

Public preferences for forest ecosystem management may vary by demographic
characteristics (Bourke and Lulofi‘ 1994, Dietz et al. 1998, Jacobson and Marynowski
_ 1997, McFarlane and Boxall 2000, Reading et al. 1994, Steel et al. 1994, Tahvanainen et
al. 2001) or by the level and type of interaction with the resource (Gobster 2001).
Numerous studies have examined the influence of socioeconomic factors on forest values
and attitudes (Bourke and Luloff 1994, Dietz et al. 1998, McFarlane and Boxall 2000,
Reading et al. 1994, Solecki 1998, Steel et al. 1994). Attitudes towards and preferences
for natural resource management may differ between rural and urban groups (Brunson et
al. 1997, Ribe and Matteson 2002, Tremblay and Dunlap 1978). In addition to the urban-
rural distinction, it is also possible to differentiate communities based on whether or not
they are dependent on timber for economic stability. The “jobs versus owls” controversy
in the northwest US. is often portrayed as an urban/rural issue but may be better
described by timber-dependent versus non-timber-dependent households (Brunson et al.
1997). There may be differences in forest management preferences between timber-
dependent rural and non-timber-dependent rural populations (Xu et al. 2003). Other
studies have examined the effects of geographic location on resource management
preferences in the context of community-based ecosystem management (Bandara and
Tisdell 2003, Cordell and Tarrant 2002, Noss and Cuellar 2001, Obiri and Lawes 2002,
Stein et al. 1999).

Many studies in the social forestry literature have concluded that residents of
timber-dependent (often rural) communities tend to be more in favor of resource

extraction, and residents of communities not dependent on timber (often urban) tend to

12

favor resource protection (Brunson et al. 1997, McFarlane and Boxall 2000, Steel et al.
1994, Tarrant and Cordell 2002, Tremblay and Dunlap 1978). This theme in the social
forestry literature is sometimes referred to as the “anthropocentric/biocentric continuum,”
in which timber-dependent communities fall at the anthropocentric end and non-timber-
dependent communities fall at the biocentric end (Tarrant and Cordell 2002). In other
words, residents of communities that are dependent on forest resources for economic
stability are more likely to support forest management that promotes resource extraction
and emphasizes instrumental values of forests and are less likely to be concerned about
the intrinsic value of forests for their own sake or their ecological and life support values,
and vice versa (Steel et al. 1994, Tarrant and Cordell 2002). Drawing upon this
literature, the guiding hypothesis of this study is that rural, timber-dependent community
residents will hold strong anthropocentric views of forest management and urban, non-
timber-dependent community residents will hold strong biocentric views of forest
management. This hypothesis is examined for a managed forest ecosystem in Michigan’s
Upper Peninsula. The forest ecosystem includes rural, timber dependent communities
and draws recreational users from a nearby urban area.

Forest ecosystems provide timber, wildlife habitat, recreation, aesthetics, clean
air, water filtration and other ecological functions. Forest management decisions result in
different levels of these market and non-market outputs, and trade-offs must be made
between competing needs of different stakeholders. In the study area, prior forest
management decisions have had ecological effects on forest structure, deer populations
and forest migratory songbird habitat. Forest management also affects the local timber-
dependent economy in this area, resulting in numerous complex interactions among

humans, forests and wildlife.

l3

This study was undertaken to provide policy makers and resource managers with
information on public preferences for and familiarity with forests and forest management
in this area of Michigan. The following research objectives were identified: 1) Identify
forest ecosystem services recognized by the public; 2) Explore public familiarity with and
perceptions of forest management, forest land ownership and forest/human/wildlife
interactions in an area of Michigan’s Upper Peninsula; and 3) Determine whether urban-

rural differences exist with respect to the topics explored in Objectives 1 and 2.

1.2. Study Area

The study focuses on an area of about 500,000 ha of forested land in Michigan’s
Central Upper Peninsula and includes parts of Baraga, Dickinson, Iron, Marquette, and
Menominee counties (See Figure 1). This region was chosen for its diverse ecological
landscape, variety of industrial and recreational uses of its forests, and for the presence of
various forest management methods. In order to achieve research objective 3 and to test
the hypothesis that timber-dependent community residents will hold anthropocentric
views of forest management and non-timber-dependent community residents will hold
biocentric views, the study area is defined as a rural, timber-dependent area and the views
of its residents are compared to Marquette, a nearby urban area that is not timber

dependent.

l4

Figure 1.1 Map of Michigan with Study Area Highlighted

Marquette

 

 

 

Felch I
f
‘p I
l

/ Il L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.2.1. Ecological characteristics of the Study Area

The study area consists of forest ecosystems in which complex interactions exist
between forest management practices, forest structure and composition, deer populations,
and forest migratory songbirds. The study area provides unique and important breeding
habitat for many species of neo-tropical migratory forest songbirds (Howe et al. 1995).
Winter deer density in this area is higher in locations with high intensity timber harvests
(unpublished data, Michigan Department of Natural Resources). Decreases in
regeneration rates of some trees, such as northern white cedar and sugar maple, may be

related to these high winter deer densities (Miller 1997, Miller et al. 1990, DeCalesta
15

2000, LeBouton et al. 2003). Changes in tree regeneration may, in turn, be adversely
affecting habitat for neo-tropical migratory forest songbirds in the area (Laurent et al.
2003). In addition, these changes may affect the local economy by decreasing future
timber availability as well as changing the structure and composition of forested areas
used for recreation. The existence of these and many other complex interactions between
people, forests and wildlife within this managed forest ecosystem, make it an appropriate

setting in which to examine public preferences for multiple use forest management.

1.2.2. Economic and demographic characteristics of the study area and Marquette

The study area is also characterized by economic dependence on timber resources.
Dominated by public and private forest lands, it is a sparsely populated region made up of
natural resource-dependent communities that rely on timber, tourism and recreation for
income, employment and economic sustainability, making it similar to many other areas
of the Upper Peninsula (McDonough et al. 1999, Potter-Witter 1995). Two common
indicators of resource dependence are employment levels in a particular sector and the
percent of land devoted to a particular industry (Bailey 2004, Machlis and Force 1988).
Direct forest-based employment (logging, sawmills and wood product manufacturing)
accounts for 12 to 28 percent of total employment in the study area (U .8. Census Bureau
2004c). Two large industrial landholders and several small mills in this area produce
valuable forest products such as dimensional lumber, poles, posts, veneer, and pulp. This
area is 90 percent forested, and the land is owned primarily by forest industry (43%) and
the state (42%) (Pugh et al. 2001).

The nearby city of Marquette and its surrounding area are not timber-dependent.

In the Marquette zip code area, 4% of the land is owned by the forest industry, and 17 %

l6

is state-owned (Pugh et al. 2001). Direct forest-based employment in the Marquette area
is less than one percent of total employment (U .8. Census Bureau 2004c). In the zip code
area of Negaunee, adjacent to the Marquette zip code area, employment in forest-based
jobs accounts for 1 to 3% of total employment in that area. Moving further west to the
zip code area of Ishpeming, adjacent to Negaunee, employment in forest industry
accounts for 4 to 9% of total employment (U .8. Census Bureau 2004c). Employment in
forest products industry increases as distance from the city of Marquette increases.
Although Marquette County produces large amounts of timber, Marquette city is not
timber-dependent and is substantially different in this respect from our study area.

Marquette and the study area also differ in population and population density.
Populations of communities within the study area range from 350 to 2,064 (U .8. Census
2004d), and are defined as rural according to the US. Census definition (less than 2,500
residents)(U.S. Census Bureau 2004a). The city of Marquette is an urban area with a
population of 19,598 and is the largest city in the Upper Peninsula (U .8. Census 2004b).
The population density of the city of Marquette is 120 to 1,999 people per square mile,
and Marquette Township has a population density of 60 to 119 people per square mile
(MTA 2004). By contrast, almost all of the townships that fall within our study area have
population densities of less than 20 people per square mile, with one township having 20
to 59 people per square mile (MTA 2004).

Although Marquette is not a typical large urban center, the intent of this study was
not to compare views of residents in the study area with those of a large urban area such
as Detroit. Rather, the study sought to compare urban and rural residents’ views of a
natural resource with which both groups have some familiarity. There are substantial

differences in population, population density and economic dependency between

17

Marquette and our study area, which, along with the existing literature, suggest that the

views of their residents may also be different.

1.3. Methods

Focus groups are a research method that collects data using group discussions
based on a topic defined by the researcher (Morgan 1997, p.6). The goal of this study is
to understand differences in perceptions and the range of understanding that rural, timber-
dependent and urban, non-timber-dependent community residents have of forests and
forest/human/wildlife interactions. Focus groups are a well-accepted method of eliciting
information from participants through open-ended questions, allowing participants to
respond in a manner of their own choosing (Kreuger and Casey 2000, p.5; Morgan 1997,
p.5). By obtaining information on familiarity and perceptions of forest management in an
open-ended discussion format, results of focus groups allow us to gain an understanding
of public views of forest resources.

The goals of focus groups are to listen to people, encourage them to share their
points of view in a permissive environment, and to gather information about people’s
perceptions by listening to their discussion (Kreuger and Casey 2000, p.4). This method
allows respondents’ discussions to reveal their familiarity with a subject as well as their
preferences, beliefs and opinions. Numerous studies have used focus groups to assess
public preferences for and perceptions of natural resources and their attributes (Kaplowitz
and Hoehn 2001, Minnis et al. 1997, Pomeroy and Beck 1999, Smith and McDonough
2001, Winter and Fried 2000). They have also been used in. studies to evaluate forest
recreation preferences (Mitra 1994) and to understand why people value forests (Hull et

al. 2001).

18

To accomplish the outlined research objectives, six focus groups were conducted
in Michigan’s Central Upper Peninsula. Planning the study is an integral part of
conducting focus group research (Kreuger and Casey 2000, p.21). The structure of the
focus group discussions was carefully developed based on input from ecologists and
foresters collaborating on this project. The discussions were structured around the
following topics: perceptions of forest services and uses of forests; goals of forest
management; forest land ownership; effects of management on forests, deer and other
wildlife; and effects of deer on forests and wildlife.

The ideal size of a focus group is between 6 and 8 participants (Kreuger and
Casey 2000, p.73). It is typical to over-recruit for a focus group because of the risk of
no-shows (Morgan 1997, p.42). Therefore, fifieen individuals were recruited for each
group to ensure that 8 people would arrive and participate. Focus group recruitment was
conducted by systematic random sampling of individual names from area telephone
directories for the city of Marquette and F elch Township to form the urban and rural
groups, respectively. Potential participants were told that the discussion would be about
issues concerning Michigan’s Upper Peninsula but were given no further details on the
topic of the discussion groups.

All individuals who showed up for a focus group were paid a $40 honorarium,
and eight people were kept to participate in the discussions. Following standard practice,
when more than eight individuals arrived, some individuals were dismissed and sent
home with their honorarium (Goldman and McDonald 1987, p.34). All urban
participants resided in the city of Marquette, and all rural participants lived in small

towns in the study area (F elch, 8; Foster City, 7; Hardwood, 4; Vulcan, 1; Ralph, 3), with

19

the exception of one participant from Kingsford (population 5,666), which shares a
telephone exchange with F elch.

Discussions were structured around the previously mentioned topics and lasted
about two hours. Each discussion was tape recorded, and a systematic analysis was
conducted of the focus group data (Kreuger and Casey 2000, p.128). Discussion
questions were designed to allow participants to become familiar with topics before
launching into key questions. Information from notes taken by the assistant moderator
was incorporated in the results, and the transcribed discussions were analyzed for content.
Transcripts were coded by attaching labels to each theme every time it appeared (ibid,
p.130). Transcripts were reviewed carefully in order to ensure that participant comments
were correctly interpreted. This is a method of verifying the intent of participant
comments after the focus groups have been conducted (ibid, p. 128). Debriefing occurred
between the moderator and assistant moderator following the discussions in order to
exchange impressions and main points that emerged from the groups. Comments made
by participants were organized into theme groups to illustrate similarities and differences
between the urban and rural discussions.

The use of numbers in reporting focus group results is controversial (Kaplowitz
2000). Caution should be used when reporting numbers, and some researchers
recommend not using numbers in focus group reporting at all (Kreuger and Casey 2000,
p.141). However, others assert that frequency counts of coded comments, or “descriptive
counting,” are especially useful in studies that seek to compare different types of groups
to reveal how often different topics were mentioned in each group (Morgan 1997, p.61 ,
Shively 1992). Descriptive statistics on participants are reported in Table 1, and

frequencies and extensiveness of comments are reported in Table 2. It is important to

20

emphasize that these numbers should not be used to generalize to the population because
the sample size is too small. Importance of comments should be determined by the
specificity, detail, emotion and intensity of comments as well as extensiveness, or how
many people made each type of comment (Kreuger and Casey 2000, p.136). Therefore,
Table 2 reports comment frequencies and extensiveness as supplements to the detailed
presentation of the discussion content in order to illustrate the differences between the

urban and rural groups.

1.4. Results and Discussion
1.4.1 Focus Group Participant Characteristics

The socioeconomic characteristics of focus group participants are reported in
Table 1. The rural participants had, on average, lived in the UP. and Michigan longer
than the urban participants, though these differences are not statistically significant at the
ten percent level. Most urban participants owned zero acres of property in the study area,
while most rural participants owned ten or more acres in the study area, and this
difference is statistically significant’. Urban participant ages are more concentrated in
lower age groups while rural participant ages tend toward higher age groups. The
number of urban participants holding a college degree or higher is larger than in the rural
groups. Urban-rural differences in age and education levels are statistically significant’.
Information about participant occupations was also collected. In the urban groups, 3
participants were employed in medicine or higher education, and no participants in the

rural groups had jobs in these fields. No participants in the urban groups had forest-

 

3 The Mann Whitney U test was used to test statistical significance.
’ The Pearson Chi Square test was used to test statistical significance.

21

related jobs, while in the rural groups, two participants had jobs in logging, suggesting

that occupation types differed slightly between the rural and urban groups.

Table 1.1 Socioeconomic Characteristics of Urban and Rural Focus Group
Participants

 

 

 

Socioeconomic Characteristic Urban Rural

(n=24) (n=24)
Average years lived in the U.P. 31 35
Average years lived in Michigan 39 46

 

Property ownership in Study Area"

 

 

 

 

 

 

 

 

Median acres owned in study area 0 43

# of people owning 0 acres in study area 22 0

# of people owning 1-50 acres in study area 0 l3

# of people owning 50-500 acres in study area 1 10

# of people owning >500 acres in study area 1 1
Age

Median age group 40-49 50-59

# of people <50 years old 16 7

# of people >50 years old 8 16

Household income

Median household income level $26-50K $26-50K

# of people with <$S0,000 in annual household 13 12

income

# of people with >$50,000 in annual household 10 9

income

Education“

Median education level Associate’s Some college

degree
# of people with Associate’s Degree or lower 16 20
# of people with College deg1_’§e or higher 8 4

 

 

* =p< 0.10; ** =p< 0.05

1.4.2. Focus Group Participant Comments

Content analysis of discussion group transcripts revealed that most rural
participants were able to speak with a high degree of specificity about
forest/wildlife/human interactions in the study area, and most urban participants spoke of

these interactions in less detail and exhibited a lower level of familiarity with forest

22

management than the rural participants. Rural participants exhibited a closer connection
with the forest resources and discussed with emotion a broad range of non-market
services provided by forests. Urban participants did not display as close a connection to
the resource and expressed emotion over a smaller range of non-market services of
forests, with an emphasis on recreational uses. These results are consistent with the
findings of Stein et al. (1999) who concluded that rural community residents have a
closer connection to the resource and receive more non-market benefits of forest
resources than urban residents and, in fact, may be more likely to value these forest
services than urban residents. However, a few urban participants who had interacted
more closely with the resource (e.g. hunters or land owners in the study area) were able to
speak with greater specificity and detail about forest management and wildlife
interactions. This suggests that differences in perceptions, attitudes, and familiarity with
forest ecosystems are also related to experiences with the resource.

The following sections describe major themes that emerged from the discussions
and highlight corresponding similarities or differences between the groups. Comments
are organized around the following themes: 1) forest services, 2) forest management, 3)
land ownership and 4) forest/human/wildlife interactions. Table 2 presents frequencies
and extensiveness for comments within these themes. In the following discussion, rural
group participant comments are numbered RI through R24, and urban group participant

comments are numbered U1 through U24.

23

Table 1.2. Theme and Sub-Theme Frequencies and Comment Extensiveness of
Urban and Rural Focus Groups

 

Theme Urban (n=24) Rural (n=24)

 

Frequency Extensiveness“ Frequency Extensiveness*
Theme 1: Forest Services

 

 

 

 

 

Beauty/emotional 7 5 22 l 3
connectron
Economic values 17 14 66 21
Recreation 53 20 43 l 8
Theme 2: Forest 47 19 87 24
management
Theme 3. Land 46 18 73 23
ownershrp
Theme 4:
Forest/HumanlWildlife
Interactions
Effects of forest
management on 54 17 53 21
wildlife
Deer browse 24 12 51 20
Effects of deer on other 16 9 28 16
wrldl1fe

 

"‘ Extensiveness = The number of unique individuals who made each comment

1.4.2.1 Forest Services

One research objective was to identify forest ecosystem services recognized by
the public. Members of both urban and rural groups discussed the importance of forests,
but their discussions of why forests are important differed. In each discussion,
participants identified similar services they associated with forests, such as recreation,
water filtration, erosion control, hunting, lumber, jobs, tourism and beauty. Both groups
expressed the importance of preserving forests for future generations, and many urban
and rural participants discussed the importance of forests to wildlife. Although the

services mentioned were similar across the groups, the discussions that followed were

24

not. Forest beauty and emotional connection emerged as a strong sub-theme in the rural
discussions. Several rural participants expressed emotion and intensity when describing
the forests in the area with words like “wonderful,” “absolutely gorgeous,”
“breathtaking” and “unbelievable.” Some rural participants said the forests are a
principle reason they enjoy living there, and some discussed their emotional attachment
to the forests.

“I wouldn 't live anywhere else, ” “[We live here] because we love it " (R18 and

R22- Similar comments fiom 5 other people)

“I am much more emotionally involved with where I live here... you never feel an

emotional attachment to bricks, but you do to trees ” (RI)

Although a few urban participants expressed emotion over the beauty of forests in the
area, the frequency and extensiveness of comments within this sub-theme were lower in
the urban than in the rural groups (See Table 2).
“I came here because I want to be here, because it made an impression when I
was that young” (U8)
“They should preserve some forest for recreational beauty ” (U13 — Similar
comment from 1 person)

Concerns for local economic sustainability and the community were expressed
repeatedly and with strong emotion by most rural participants. They spoke with
specificity about the important role of forests in providing employment in recreation and
tourism as well as the forest products industry and the ways in which this affects their
community.

“When I think of the forest, I think of recreation, a lot of jobs in this area, very

important around here” (R4 —- 4 other people made similar comments)

25

“[ F orestry] is the only industry we have left here ” (R1 7 - Similar comments

made by 5 other people)

Rural groups independently identified the concept of balance, which did not
emerge from the urban discussions. Rural participants discussed difficulties of balancing
conflicting goals of forest management such as managing to provide jobs while
maintaining wildlife habitat and aesthetics. The strong emotion and specificity of these
comments were tied to stated concerns the rural participants had for community stability
and the importance to them of the multiple roles that forests play in their lives.

"It just isn 't one thing, it ’s everything together, the timber has to provide for that

and the people survive of the deer hunting and the timber and the recreation is a

big thing up here. " (R14 — Three other people made similar comments)

The economic importance of forests was discussed to a lesser extent in the urban
groups, which can be seen by comparing the frequency and extensiveness of comments in
this sub-theme in Table 2. A few participants referred to the importance of the forest
products industry to the region, though comments were not specific and did not express
strong emotion about the economic importance of forests to the local economy.

“It 's important to cut down [timber] for jobs, for industry, because we need to
have pulp ” (UI3 — Similar comments were made by 4 other people)

Urban participants referred to economic issues by mentioning recreation,
specifically the importance of snowmobile and ATV trails as tourism draws.

“... snowmobiling and 4-wheelers are getting more popular, there should be trails

for those also, it ’s a strong economic asset " (U1 - Similar comment fiom 1 other

person)

26

Concern for maintaining recreational opportunities in forests was mentioned by
most urban participants in the context of resource use for individual benefits. Urban
participants discussed the importance of forests for a variety of recreational uses.

“... you ’re not going to picnic in a clearcut, you know, there ’s really nothing afier

it ’s clearcut, there ’s not much you can do there, until like they say, 40 years fiom

now, there are some bird hunting opportunities after a few years... " (U10 — Five
other people made similar comments)

Urban comments reflect an anthropocentric view of forest management, not for
favoring resource extraction, but for favoring management that enhances individual
recreational opportunities. Rural comments, by contrast, reflected the importance of
forests for many purposes, including instrumental as well as intrinsic values, indicating
that perhaps the rural participants’ views lie not at the anthropocentric end of the
continuum but somewhere in the middle.

Among rural participants, concern was expressed for maintaining forest
recreational services to support the local economy through tourism revenues. A few
people mentioned the importance of forest aesthetics in attracting tourists.

“I know lots and lots of people make a special trip just to come and see the [fall]
colors here ” (R1 - Three other people made similar comments)
Some rural participants also discussed the importance of forest recreation through
hunting, with a strong emphasis on deer hunting and associated tourism revenues.

“... hunting is for recreation, but it ’s also economy up here, because [hunters]

spend a lot of money, generate a lot of money ” (R18 — Similar comments made by

6 other people)

27

Overall, reasons for concern about forest recreational opportunities differed
between the urban and rural groups. Both urban and rural participants’ views of
recreation can be interpreted as anthropocentric, but rural participants expressed a
stronger concern for community well-being than for their own personal interests in

recreational opportunities.

1.4.2.2 Forest Management
One objective established for the study was to explore public familiarity with and
opinions of forest management and harvesting practices. All of the rural group
participants were able to contribute something to this discussion, and most of them were
able to discuss different types of forest management practices in detail. The discussion of
forest management in the rural groups included many specific references to management
practices (See Table 2). Many rural participants exhibited familiarity with and an
understanding of different types of harvesting practices such as select cutting and
clearcutting and discussed post-harvest activities such as replanting. A few people
discussed the benefits of selective cutting.
the advantages of select cutting are tremendous... you take trees out of a
certain area that are mature, it allows the sunlight to get in, ground vegetation to
start, smaller trees to mature, it ’s just fantastic for the whole area” (R4- Five
other people made similar comments)
A few rural participants recognized the inability of loggers to practice lower impact

harvesting because of economic factors.

28

“The market won 't allow [select cutting], you ’ve got to get volume, because
prices I don ’t think have gone up that much, you just have to do it in volume”

(R1 1 — Six other people made similar comments)

The urban participants had a less detailed discussion of forest management
practices that focused on clearcutting, which some participants viewed negatively. Most
of the urban participants were not able to speak about any other harvesting practices such
as select cutting.

“Clearcutting looks like a scar " (U4 — Similar comments made by four other

peeple)

Some urban participants discussed increasing logging intensity in the area over the years.
“It seems that we ’re cutting 10 times as much out of the wood: as we used to, with
the equipment we have now. ” (U1 1 — Three other people made similar
comments)

Some urban participants made detailed comments about management, but these were not

as extensive as those in the rural groups. Urban group discussions of the effects of

intense timber harvesting on forest services was not detailed, and participants did not
mention possible negative effects of harvesting on other forest services. Many rural
participants discussed with specificity the economic need for timber extraction, however,
many of them also discussed with great emotion negative feelings towards the intensity
of extraction they see in their community. Most rural participants discussed negative
effects of intense harvesting on provision of other non-market forest services and

expressed strong concern for maintaining these other services.

29

1.4.2.3 Land Ownership
Participants were asked about land ownership and its implications for forest
management. Perceptions of land ownership and its effects differed between the groups.
Many of the rural participants view non-industrial private landowners as protectors of the
area’s forests whereas they view private industrial and public landowners as threats to
forest resource sustainability.
the individuals who manage their land, or cut their land, they do a lot better
job managing their land than the state... .a person lives here knows what ’s going
to happen if he cuts all those trees. ” (R12 — Similar comments were made by 6
other people)
Most urban participants held an opposite view of the role of landowners. They expressed
that private owners practice uninformed management that depletes forest resources while
corporate or public land management is beneficial to forest resource sustainability.
“... a private owner probably won ’t replant because they ’re getting rid of [the
trees] because they want a yard, or because they want access to something, or
they want the firewood ” (U3 — Similar comments were made by 3 other people)
Although urban and rural groups recognized that landowner type affects the way forests
are managed, perceptions of the roles of different landowners differed between the
groups.
1.4.2.4 Forest Management and Wildlife Interactions

An objective of the research was to learn about familiarity with and perceptions of
forest, wildlife and human interactions. Discussions that explored this theme are divided
into three sub-themes: 1) effects of forest management on deer and other wildlife, 2)

effects of deer on forests, and 3) effects of deer on other wildlife. Many rural participants

30

were able to discuss with specificity and detail the effects of forest management and
harvesting practices on wildlife in the area, and most participants expressed emotion
when discussing the effects of harvesting on wildlife habitat.

“We have to selective cut... if you get an area that has too much tree cover then
you don 't get the undergrowth so you don 't have the animals " (R3 — Similar
comments were made by l 1 other people)

Urban participants discussed in less detail the effects of forest management on wildlife.
Most were able to discuss clearcutting, and some mentioned its effects on wildlife. Many
were not able to speak about specific ways in which forest management can affect
wildlife.

“Certainly, anytime they change something in the forest, it alters all the habitat, I
haven 't been out in the woods doing much with wildlife, but I realize how it
afi’ects one and another ” (U18 — Similar comments were made by 6 other people)

A few urban participants did not accurately understand how timber harvesting affects
wildlife.

“T here ’s not as much wildlife in the places where they cut ” (U7 — Similar
comment made by two other people)

Perceptions of forest management and its effects on wildlife differed between
urban and rural groups. While many urban participants were not aware of the ways in
which clearcutting can affect wildlife, most rural participants were able to speak in detail
about these effects. Most rural participants spoke accurately and in detail about effects of
forest management on wildlife. A few urban participants were able to speak in detail
about these interactions, and these individuals have had more frequent contact with these

forests through recreational activities. Rural participants were more strongly aware of

31

forest/human/wildlife interactions, as evidenced by comment frequencies and
extensiveness in Table 2, and they displayed concern for the instrumental values of
forests as well as ecological values.

In the second section of human/forest/wildlife interactions, participants were
probed on their familiarity with the effects of deer on the forest. Most rural group
participants engaged in a detailed discussion of the effects of deer browse on the forest,
and a few participants had already mentioned this topic before it was initiated by the
moderator.

“Definitely, [deer aflect the forest], in a stand of hardwoods, when there ’s too

many deer. You get all this scrub brush that ’s never going to be worth anything,

you ’ll never be able to aflord the taxes on the land because deer eat the sugar
maple " (R7 - Similar comments were made by I 1 other people)
In contrast, urban participants did not initiate the deer browse discussion on their own,
and this discussion followed the predetermined script. Many were familiar with deer
browse and were able to have a general discussion about it. Of the urban participants
who discussed browse, most did not mention its specific effects on the forest and some
did not see it as a problem.
“I think the deer would have to be really overpopulatedfor quite some time in
order to have a big noticeable impact on forests " (U2 — Similar comments were
made by 4 other people
“Yeah, you can tell when there ’s large numbers and there isn ’t a lot of food,
everything fiom like 6 feet down will be stripped ” (U1 7 — Six other people made

similar comments)

32

Comments by rural participants about deer browse were very detailed and extensive,
while comments from urban participants expressed some familiarity with deer browse but
were general and not extensive (See Table 2).

In the third section on human/forest/wildlife interactions, participants were asked
about possible effects of deer on other wildlife. Members of both urban and rural groups
found it difficult to understand how deer could impact other wildlife. Neither urban nor
rural participants initiated this topic independently. Many rural group participants made
an attempt to think of ways deer could affect other wildlife while only some urban
participants contributed ideas on this topic (See Table 2). Neither group spoke on this

t0pic with much specificity or emotion.

1.5. Conclusions

This study was undertaken in order to better understand the preferences and
vieWpoints of urban and rural groups with respect to the complex interactions between
peOple, forests and wildlife in Michigan’s Central Upper Peninsula. Much of the
literature suggests that rural, timber-dependent community residents would hold strong
preferences for managing forests for human uses, and urban, non-timber-dependent
Communities would hold strong preferences for managing forests for ecological and
biological uses. The guiding hypothesis of this study was that rural, timber-dependent
community residents will hold strong anthropocentric views of forest management and
urban, non-timber-dependent community residents will hold strong biocentric views of
forest management.

The results indicate that urban/rural differences in perceptions of forests and

fOFCSt/human/wildlife interactions do not fit smoothly with the two typologies defined in

33

the literature. In contrast to some studies, the rural participants did not fall neatly on the
anthropocentric end of the anthropocentric-biocentric continuum because they do not
place importance simply on extractive, utilitarian uses of the forests. Similarly, the
findings do not place the urban participants neatly at the biocentric end of the continuum
because they did not express strong preferences for forest conservation, but they did
express strong concerns about anthropocentric forest uses such as their own recreation.

Given the limitations of qualitative research, the quantitative data in this study
should not be used to generalize the results to other populations. However, the
qualitative findings are evidence that viewpoints of rural timber-dependent and urban
non-timber-dependent community residents may not fall at expected ends of the
anthropocentric-biocentric continuum. In crafting natural resource policy, management,
and communication strategies, our findings suggest that decision makers and other
researchers should be cautious in utilizing a simple anthropocentric/biocentric continuum
or similar results of previous studies since our research suggests that relationships are
more complex.

Two unique features of this study may be fruitful avenues of research aimed at
better understanding this evidence. First, even though the urban area in this study is the
largest in the Upper Peninsula of Michigan, it is not a major metropolitan area. Second,
the urban area was close enough to the study resource to serve recreational purposes, and
the findings revealed that recreational concerns were more commonly mentioned by the
urban groups than biocentric concerns. Studies aimed at identifying the effect of these
features might improve the robustness of the anthropocentric/biocentric continuum.

The differences in resource and forest management familiarity and opinions

identified above can be useful to policy makers and resource managers interested in

34

designing management strategies to more effectively satisfy public preferences. The
information can also be used in designing communication and educational efforts to
articulate the goals and purpose of forest management. For example, although the rural
participants demonstrated a strong understanding of forest management techniques and
were clear about the need to balance the competing goals of forest management, the
urban participants did not express such understanding and might be better reached by
messages targeting their recreational and personal use of forests. Messages to rural
residents should be cognizant of the importance of balancing resource extraction and
resource protection. For example, emphasizing the degree to which voluntary forest
landowner incentive programs can help balance timber production with resource
protection could encourage forest landowners in rural areas to implement conservation

practices on their land.

35

REFERENCES

Bailey, C. 2004. Resource Dependency and Rural Development in Alabama [cited
November 10, 2004]. Available from

ht_tp://www.ag.aubum.edu/~cbailey/demndencyhtm.

Bandara, R., and C. Tisdell. 2003. Comparison of rural and urban attitudes to the
conservation of Asian elephants in Sri Lanka: empirical evidence. Biological
Conservation 110:327-342.

Bingham, G., R. Bishop, M. Brody, D. Bromley, E. Clark, W. Cooper, R. Costanza, T.
Hale, G. Hayden, S. Kellert, R. Norgaard, B. Norton, J. Payne, C. Russell, and G.
Suter. 1995. Issues in ecosystem valuation: improving information for decision
making. Ecological Economics 14:73-90.

Bourke, L., and A.E. Luloff. 1994. Attitudes toward the management of nonindustrial
private forest land. Society and Natural Resources 7:445-457.

Boxall, P.C., and B. Macnab. 2000. Exploring the preferences of wildlife recreationists for
features of boreal forest management: a choice experiment approach. Canadian
Journal of Forest Research 30: 193 1-1 941.

Brunson, M.W., B. Shindler, and BS. Steel. 1997. Consensus and dissention among rural
and urban publics concerning forest management in the Pacific Northwest. In Public
Lands Management in the West: Citizens, Interest Groups, and Values, edited by B.
Steel, pp. 83-94. Westport, CT: Praeger.

Cordell, H., and M. Tarrant. 2002. Changing demographics, values, and attitudes. Journal
of Forestry 100 (7):28-33.

DeCalesta, D.S. 2000. Deer, ecosystem damage, and sustaining forest resources 1997
[cited October 31 2000]. Available from http://wwwarec.umd.edu/mlicy/Deer—

Management-in-Magyland/decalesta.htm.

Dennis, D.F. 1998. Analyzing Public Inputs to Multiple Objective Decisions on National
Forests Using Conjoint Analysis. Forest Science 44 (3):421-429.

Dietz, T., P.C. Stern, and G.A. Guagnano. 1998. Social structural and social psychological
bases of environmental concern. Environment and Behavior 30:450-471.

Gobster, PH. 2001. Human dimensions of early successional landscapes in the eastern
United States. Wildlife Society Bulletin 29 (2):474-482.

Goldman, A.E. and 8.8. McDonald. 1987. The Group Depth Interview: Principles and
Practice. Englewood Cliffs, NJ: Prentice-Hall, Inc.

36

Howe, R.W., Niemi, G., and J .R. Probst. 1995. Management of Western Great Lakes
Forests for the Conservation of Neotropical Migratory Birds. In Management of
Midwestern Landscapes for the Conservation of Neotropical Migratory Birds, edited
by F .R. Thompson 111, pp. 144-167. Detroit, MI: USDA Forest Service North Central
Forest Experiment Station.

Hull, B.P., D. Robertson, and A. Kendra. 2001. Public understandings of nature: a case
study of local knowledge about 'natural' forest conditions. Society and Natural
Resources 14:325-340.

Jacobson, S.K., and SB. Marynowski. 1997. Public attitudes and knowledge about
ecosystem management on Department of Defense land in Florida. Conservation
Biology 1 1(3):770-81.

Kaplowitz, MD. 2000. Statistical analysis of sensitive topics in group and individual
interviews. Quality and Quantity 34:419—43 1 .

Kaplowitz, M.D., and J .P. Hoehn. 2001. Do focus groups and individual interviews reveal
the same information for natural resource valuation? Ecological Economics 36:237-
247.

Kearney, A.R. 2001. Effects of an informational intervention on public reactions to clear-
eutting. Society and Natural Resources 14:777-790.

Kreuger, R.A. and M.A. Casey. 2000. Focus Groups: A Practical Guide for Applied
Research, 3rd edition. Thousand Oaks: Sage Publications.

Laurent, E.J., J.P. LeBouton, D. Gatzioli, M.B. Walters and J. Liu. 2003. Evaluating the
effects of land cover classification indices used to investigate wildlife-habitat
relationships. Presentation. Ecological Society of America, Savannah, GA, August 3,
2003.

LeBouton, J .P., E]. Laurent, M.B. Walters, and .1. Liu. 2003. Regional vs local effects of
white-tailed deer herbivory on vegetation structure and composition in northern
hardwood forests. Ecological Society of America 88th annual meeting, Savannah,
Georgia. August 3-8, 2003.

Machlis, GE. and J .E. Force. 1988. Community Stability and Timber-Dependent
Communities. Rural Sociology 53:220-234.

McDonough, M., J. Fried, K. Potter-Witter, J. Stevens, and D. Stynes. 1999. The Role of
Natural Resources in Community and Regional Economic Stability in the Eastern
Upper Peninsula. East Lansing, MI: Michigan Agricultural Experiment Station,
Michigan State University, Research Report 568.

McFarlane, B.L., and P.C. Boxall. 2000. Factors Influencing Forest Values and Attitudes
of Two Stakeholder Groups: The Case of the Foothills Model Forest, Alberta, Canada.
Society and Natural Resources 13:649-661.

37

Miller, R. 1997. Northern White-Cedar Regeneration: Promise and Problems. Escanaba,
M1: Upper Peninsula Tree Improvement Center, Michigan State University.

Miller, R.O., D. Elsing, M. Lanasa, and M. Zuidema. 1990. Northern White-Cedar: Stand
Assessment and Management Options. Presented at Northern White-Cedar in
Michigan Workshop, Sault Sainte Marie, February 21-22, 1990.

Minnis, D.L., R.H. Holsman, L. Grice, and RB. Payton. 1997. Focus groups as a human
dimensions research tool: three illustrations of their use. Human Dimensions of
Wildlife 2 (4):40-49.

Mitra, A. 1994. Use of focus groups in the design of recreation needs assessment
questionnaires. Evaluation and Program Flaming 17 (2): 133-140.

Morgan, D.L. 1997. Focus Groups as Qualitative Research, 2nd edition. Vol. 16,
Qualitative Research Methods Series. Thousand Oaks, CA: Sage Publications.

MTA, Michigan Township Association. 2004. Michigan Township Maps 2004 [cited
October 27 2004]. Available from http://www.liaa.cc/mta/maps.

Noss, A., and R. Cuellar. 2001. Community attitudes towards wildlife management in the
Bolivian Chaco. Oryx 35 (4):292-300.

Obiri, J., and M. Lawes. 2002. Attitudes of coastal-forest users in Eastern Cape Province

to management options arising from new South African forest policies. Environmental
Conservation 29 (4):519-529.

Pomeroy, C., and J. Beck. 1999. An experiment in fishery comanagement: evidence from
Big Creek. Society and Natural Resources 12:719-739.

Potter-Witter, K. 1995. Timber and Timberland Resources. East Lansing, MI: Michigan
Agricultural Experiment Station, Michigan State University, Special Report 71.

Pugh, S.A., Reed, D.D., Pregitzer, K.S. 2001. F IAMODEL User's Guide Version 3.0.
Houghton, MI and St. Paul, MN: School of Forestry and Wood Products, Michigan
Technical University and USDA Forest Service North Central Research Station.

Reading, R.P., T.W. Clark, and SR. Kellert. 1994. Attitiudes and knowledge of people
living in the Greater Yellowstone Ecosystem. Society and Natural Resources 7
(4):349-365.

Ribe, R., and M. Matteson. 2002. Views of Old Forestry and New Among Reference
Groups in the Pacific Northwest. Western Journal of Applied Forestry l7 (4): 1 73-1 82.

Shindler, B.A., M. Brunson, and OH. Stankey. 2002. Social Acceptability of Forest

Conditions and Management Practices: A Problem Analysis. USDA Forest Service
Pacific Northwest Research Station General Technical Report PNW-GTR-537.

38

Shively, J. 1992. Cowboys and Indians: Perceptions of Western Films Among American
Indians and Anglos. American Sociological Review 57:724.

Smith, RD, and M.H. McDonough. 2001. Beyond public participation: fariness in natural
resource decision making. Society and Natural Resources 14 (3):239-249.

Solecki, W.D. 1998. Local attitudes on regional ecosystem management: A study of New
Jersey Pinelands residents. Society and Natural Resources 11 (5):441-463.

Steel, B., P. List, and B. Shindler. 1994. Conflicting values about federal forests: A
comparison of national and Oregon publics. Society and Natural Resources 7: 137-153.

Stein, T.V., D.H. Anderson, and T. Kelly. 1999. Using stakeholders' values to apply
ecosystem management in an upper midwest landscape. Environmental Management
24(3):399-413.

Tahvanainen, L., L. Tyrvainen, M. Ihalainen, N. Vuorela, and O. Kolehmainen. 2001.
Forest management and public perceptions - visual versus verbal information.
Landscape and Urban Planning 53:53-70.

Tarrant, M., and H. Cordell. 2002. Amenity values of public and private forests:
Examining the value-attitude relationship. Environmental Management 30 (5):692-
703.

Tremblay, K.R., and R. Dunlap. 1978. Rural-urban residence and concern with
environmental quality. Rural Sociology 43:474-491.

US. Census Bureau. 2004a. Urban and Rural Definitions 1995 [cited October 7 2004].
Available from http://www.censuagov/_population/cen_susdata.urdef.txt.

. 2004b. Census 2000 Urban and Rural Population for Michigan, Counties and
County Subdivisions 2000 [cited June 25 2004]. Available from
http://www.census.gov/geo/www/ua/ua 2k.html.

 

. 2004c. Zip Code Business Patterns (NAICS) 2001 2001 [cited September 14
2004]. Available from http://censtats.census.2ov/cgi-bin/sznaic/szsect.pl.

 

. 2004d. American Fact Finder 2004 [cited October 20 2004]. Available from
http://gctfinder.census.gov.

 

Winter, G., and J .S. Fried. 2000. Homeowner perspectives on fire hazard, responsibility,
and management strategies at the wildland-urban interface. Society and Natural
Resources 13:33-49.

Xu, W., B.R. Lippke, and J. Perez-Garcia. 2003. Valuing biodiversity, aesthetics, and job

losses associated with ecosystem management using stated preferences. Forest Sci.
49(2):247-257.

39

Zinkhan, F.C, T.P. Holmes and DE. Mercer. 1997. Conjoint analysis: A preference-based
approach for the accounting of multiple benefits in southern forest management.
Southern Journal of Applied Forestry 21 (4):180-186.

40

Essay 2:
Forest and Environmental Values and Attitudes:
Differences Between Demographic Groups in Michigan

2.1 Introduction

Ecosystems are valued by people because they provide services that satisfy
material and nonmaterial human needs (MA 2003). The goods and services provided by
ecosystems can have ecological, economic and sociocultural values (de Groot et al.
2002), and a wide variety of methods have been developed across different disciplines for
estimating and understanding these values (MA 2003). Information on environmental
values and attitudes is an important component of natural resource management and
policy because it can help natural resource managers and policy makers understand
which aspects of natural resource management are important to different segments of the
public (Bengston et al. 2001, Jacobson and Marynowski 1997, Weiler and O’Leary 1997,
Tarrant and Cordell 2002, McFarlane and Boxall 2000, Potter and Norville 1981,
Reading et al. 1994, Solecki et al. 1998). An important step in incorporating ecosystem
values into decision making is understanding values and attitudes towards the
management of ecosystem services and the factors that influence those values and
attitudes (McFarlane and Boxall 2000). For example, values for and attitudes towards
forest ecosystem management and the environment may be influenced by socioeconomic
characteristics. This study examines the relationship between socioeconomic
characteristics and environmental and forest ecosystem management values and attitudes
among Michigan residents.

The influence of socioeconomic and social influence factors on attitudes and

values toward resource management has been the subject of many studies (Honnold

41

1981, McFarlane and Boxall 2000, Dietz et al. 1998, Steel et al. 1994, Tarrant and
Cordell 2002). Resource dependence, urban versus rural residence location, age and
education are considered the strongest predictors of environmental attitudes (Dietz et al.
1998, McFarlane and Boxall 2000). Others have found that gender (Dietz et al 2002),
political orientation (Steel et a1 1994, Jones et al 1999) and ethnicity (Tarrant and Cordell
2002) have an important influence on environmental attitudes. Influences of other
socioeconomic variables such as income, occupation and industrial sector and religion on
environmental values have been found to be relatively weak (Dietz et al. 1998).

Understanding how different demographic groups view forest management and its
effects can facilitate a better understanding of the economic and ecological trade-offs
involved in forest management (Xu et al. 2003). In order to analyze environmental
values and attitudes, it is important to define them and put them within a structural
context. A value is defined as “an enduring conception of the good” (Rokeach 1973),
and an attitude is defined as “a learned disposition toward some object as either favorable
or unfavorable” (Fishbein and Ajzen 1975). Studies of environmental and ecosystem
service values have used a social psychological approach in order to provide a theoretical
framework for understanding the relationship between values and attitudes (McFarlane
and Boxall 2000, Stem et al. 1995, Tarrant and Cordell 2002).

The social psychological approach used to understand environmental and forest
ecosystem management values and attitudes is based within the cognitive hierarchy
framework (Rokeach 1973). Within this framework, basic values, representing
fundamental social and biological needs, influence general beliefs and held values, which
in turn affect attitudes, which ultimately influence behavior (ibid.) This framework is

illustrated in Figure 2.1. Based on this framework, general beliefs or held forest values

42

influence environmental and resource management attitudes or preferences, and these
preferences ultimately affect behavior and specific actions (Stern et al. 1995). In this
framework, an individual’s economic value is derived from behavior, either observed or
stated. For example, a willingness to pay for a good or service reveals an economic value

through behavior. Thus, attitudes and other values are antecedents to economic values.

Figure 2.1 Cognitive Hierarchy Framework

 

 

Behavior

Knowledge
Socioeconomic
Factors
General Beliefs/
Held Values

Basic Values

 

 

 

Adapted fiom McFarlane and Boxall 2000

Numerous approaches exist for conceptualizing a range for held forest values,
including distinctions between instrumental and noninstrumental or intrinsic values (Xu
and Bengston 1997) and between anthropocentric and biocentric values (McFarlane and
Boxall 2000, Steel et al. 1994, Tarrant and Cordell 2002). Instrumental or
anthropocentric values emphasize the extractive uses of forest ecosystems for human

benefit (McFarlane and Boxall 2000). This value orientation presumes that humans have

43

a central role in the natural world and nonhumans are valued only in terms of the benefits
they provide to people (Steel et al. 1994). A biocentric value orientation, on the other
hand, is based on a “nature centered” approach in which the natural world is valued in its
own right, not only for the benefits it provides to humans, and it recognizes the place of
humans within a larger ecological context (ibid).

Studies have concluded that individuals with particular demographic
characteristics should fall in predictable ways along a continuum of anthropocentric to
biocentric values (T arrant and Cordell 2002). However, Essay 1 concluded that residents
of rural, timber dependent and urban, non-timber dependent communities do not have
value orientations that fall neatly along this continuum. The qualitative data used in the
analysis in Essay 1 was taken from a very small sample, making it difficult to extrapolate
to the larger population. However, the results of Essay 1 suggest the need for a deeper
investigation of the relationship between demographic characteristics and environmental
attitudes. This essay builds on Essay 1 by using a much larger data set taken from the
results of a mail survey of Michigan residents that collects data on demographic
characteristics, general environment attitudes, and attitudes towards forest management
in a particular area of Michigan’s Upper Peninsula.

This research uses the anthropocentric/biocentric continuum approach to forest
value orientations and further bases the analysis in the cognitive hierarchy framework, in
which values underlie attitudes. Environmental attitude data are used to provide
information on the underlying held forest values of individuals. There are two guiding
hypotheses of this study. The first is that there is an underlying structure to the way in
which individuals respond to forest management and environmental attitude questions

that reveals information about held forest and environmental values. The second

44

hypothesis is that these values can be explained by a set of socioeconomic characteristics.
Exploratory factor analysis is used to determine whether there is an underlying structure
to responses to attitudinal questions. Multivariate regression analysis is used to explain

the resulting environmental value factors using a set of socioeconomic variables.

2.2 Methods
2.2.1 Survey Design and Implementation

A mail survey of Michigan residents was designed using results from focus
groups, individual interviews and interviews with forestry and agency professionals (See
Essay 1 for details of the qualitative research procedures and analysis and Essay 3 for
details of survey design). The survey focuses on an area of Michigan’s Upper Peninsula,
referred to in this essay as the study forest (See Essay 1 for a description and map of the
study forest). A 5-point likert scale was used to elicit responses to attitudinal statements
ranging from ‘strongly agree’ to ‘strongly disagree.’ Summaries of response frequencies
to all attitude questions in the survey are presented in Appendix 6.

The sample was stratified into four geographic regions of Michigan: the study
forest, the rest of the Upper Peninsula, the northern Lower Peninsula, and the southern
Lower Peninsula. Details of the sample design are provided in Appendix 1. An equal
number of households was chosen from each strata in order to ensure that a sufficient
number of responses would be attained to represent the study forest, a small, sparsely
populated area, as well as the remainder of the UP, which is also sparsely populated
relative to the rest of the state. The survey was mailed on April 15, 2005 to a stratified
sample of 2,000 Michigan residents using a modified version of Dillman’s tailored design

method (Dillman 2000). A total of 954 usable surveys were returned, yielding an overall

45

response rate of 50% (See Essay 3 and Appendix 5 for details of survey implementation).
The survey contains a set of attitudinal statements related to forest management
goals that were designed to elicit attitudes relating to trade-offs between economic and
ecological outcomes of forest management. Other attitude statements were designed to
elicit respondent opinions on community forest management participation, forest
management programs, the quality of forest management by different types of
landowners, and attitudes towards environmentally certified wood products. Nine
environmental attitude statements were included in the survey. These are based on the
New Environmental Paradigm (NEP), a popular and widely used environmental attitude
scale originally developed by Dunlap and VanLiere (1978). The original scale consisted
of 12 environmental attitude statements, which are reduced to nine statements in this
study based on Tarrant and Cordell’s adaptation of the NEP (2002). Other survey
questions elicited information about recreational use of the study area as well
socioeconomic characteristics such as age, education, income, etc. A summary of

respondent socioeconomic characteristics is reported in Appendix 6.

2.2.2 Factor Analysis and Factor Score Regression

Factor analysis is a multivariate statistical technique used to reduce a large
number of variables into a smaller, usable group of factors which can then be subjected to
further analysis (Hair et al. 1998, Churchill 1991). This method is commonly used to
reduce data into a smaller set of factors that can be used to linearly reconstruct the
original variables (STATA 2003). One principal purpose of factor analysis is to identify
an underlying structure to the data (Aaker and Day 1986). Factor analysis is represented

by the following equation:

46

y”- : Zilblj '1' 21'2sz +...+ 21°qu + 81:]: (I)

where yij is the value of the ith observation of variable j, zik is the value of observation i
on factor k, bk]- are linear coefficients referred to as factor loadings, and the e ij are unique
factors of variable j, similar to residuals (STATA 13.0). A factor loading is the
correlation between the factor and the original variable (Hair et a1 1998). The by, or
factor loadings, can be estimated using principal components factor analysis. The
estimated factor loadings provide a basis for creating a new set of variables, or factors,
that represent the original variables (ibid.). A factor can be interpreted as a construct or
unobservable variable that is inferred from the set of variables used as inputs to the
analysis and is interpreted based on factor loadings (Aaker and Day 1986).

Factor analysis can lead to infinite factor solutions, and the decision of the
number of factors to use is at the researcher’s discretion. There are a number of
guidelines that can be used to determine a limit to the number of factors to be extracted.
One criterion is the latent root (eigenvalue) criterion, which states that only factors that
have eigenvalues greater than 1 are considered significant and should be retained
(Churchill 1991, Hair et al. 1998). An eigenvalue, which is the sum of squared loadings
for a factor, represents the amount of variance explained by each factor (Hair et al. 1998).

Another criterion is the scree test, which plots the latent roots against the number of
factors extracted. A common rule of thumb is to cut off the factors at the point where the
eignenvalue plot in the scree graph becomes flat (Churchill 1991).

When the desired number of factors to be extracted is determined, the initial

factor matrix is generated. This matrix contains the factor loadings of each variable on

each factor. Each factor is orthogonal to other factors, meaning there is no correlation of

47

factor axes to each other (Hair et a1 1998). This relationship is achieved by deriving each
factor from the variance that remains after the previous factor has been extracted. This
process achieves data reduction, but in most cases, it does not provide the best
interpretation of the variables being investigated (ibid). It is important to rotate the factor
solution in order to improve the interpretation of the variables’ contributions to the
factors. In factor rotation, the reference axes of the factors are rotated to a new position,
which redistributes the variance among the factors, resulting in a simpler and more
meaningful result (ibid.). The VARIMAX rotation, which is an orthogonal rotation, is
the most commonly used method. In this type of rotation, the axes are maintained at 90
degrees.

Once the factor matrix has been rotated, it is necessary to identify the highest
factor loading for each variable. When these loadings are identified, the variables that
load highly onto the same factors are placed into groups, and these groups are interpreted
by the researcher and labeled according to the similarity of the variables within them.

Factor loadings, once calculated, can be used to compute new variables called
factor scores. The rotated factor loading matrix is used to compute factor coefficients,
which are, in turn, used to compute factor scores for each individual in the sample. These
scores can be viewed as composite measures that indicate the degree to which a person
scores highly on a particular factor based on their responses to the variables included in
that factor (Hair et al. 1998). The new variables created by factor scores can then be used
in subsequent analyses.

Survey data on environmental and forest values can be reduced to form smaller
sets of factors, which can then be converted to factor scores and analyzed using

multivariate regression techniques. This method of analysis can be used to analyze the

48

effect of demographic variables on factor scores derived from environmental attitude data
(Dietz et al. 1998, McFarlane and Boxall 2000). These regression analyses of each factor
score on a set of explanatory socioeconomic variables can provide information on the
demographic characteristics of respondents based on their factor scores.

The following equation will be estimated using OLS regression:

F}=flz+e (2)

In this equation, F ,- represents the factor score value of factor i, which is estimated based
on the results of the factor analysis, [3 is a vector of coefficients, and z is a vector of
socioeconomic variables. Explanatory variables will be presented in Table 2.5. Stepwise
regression is used to estimate the parameters of this model for all factors that result from

the factor analysis. Results of the estimation are reported in section 2.4.

2.3 Factor Analysis Results

Factor analysis was conducted on responses to the attitude statements listed in
Table 2.1. These statements were used in the survey to elicit attitudes about the goals of
forest management with respect to human and wildlife needs as well as general
environmental attitudes. The hypothesis is that there is an underlying structure to the
way in which individuals respond to these attitude statements. Factor analysis using
principal components is conducted using the attitude statement variables listed in Table
2.1 as inputs. The latent roots criterion, percentage of variance, and the scree test are

used to determine the number of factors to retain in the analysis.

49

Table 2.1. Survey Attitude Statements and Associated Variables Used in Factor
Analysis (N = 608)

 

Variable

Survey Statement

 

HUWEED

In my opinion, the Western U.P. Study Forest should be managed to meet
the needs of people.

 

WLFNEED

In my opinion, the Western U.P. Study Forest should be managed to meet
the needs of wildlife.

 

C OWEED

In my opinion, the Western U.P. Study Forest should be managed to meet
the needs of communities that are economically dependent on forests, no
matter what effect this has on the environment.

 

BALNEED

In my opinion, the Western U.P. Study Forest should be managed to
balance environmental needs with the needs of communities that are
economically dependent on forests.

 

F UT GENNEED

In my opinion, the Western U.P. Study Forest should be managed to meet
the needs of future generations.

 

MTNINDJOB

In my opinion, the Western U.P. Study Forest should be managed to
maintain forest industry jobs.

 

MTNRECJOB

In my opinion, the Western U.P. Study Forest should be managed to
maintain forest-based recreation jobs.

 

PROTRES

In my opinion, the Western U.P. Study Forest should be managed to protect
forest and wildlife resources.

 

BALIOBRES

In my opinion, the Western U.P. Study Forest should be managed to
achieve a balance between maintaining forest-related jobs and protecting
forest and wildlife resources.

 

INCRDIV

In my opinion, the Western U.P. Study Forest should be managed to
increase migratory forest songbird diversity even if there are economic
losses to forest-based industries.

 

BALINDDI V

In my opinion, the Western U.P. Study Forest should be managed to
achieve a balance between sustaining forest-based industries and migratory
forest songbird diversity.

 

PROTHAB

BALINDHAB

¥

fiUMABUSE

In my opinion, the Western U.P. Study Forest should be managed to protect
habitat for migratory forest songbird species of conservation concern even
if it results in economic losses to forest-based industries.

In my opinion, the Western U.P. Study Forest should be managed to
achieve a balance between sustaining forest-based industries and protecting
habitat for migratory forest songbird species of conservation concern.

Humans are severely abusing the environment.

50

Table 2.1 (continued)

 

 

 

 

 

 

 

Variable Survey Statement

HUMRIGHT Humans have the right to modify the natural environment to suit their
needs.

HUMRULE Humans were meant to rule over nature.

HUMLEAIRN Humans will eventually learn enough about how nature works to be able to
control it.

BALNA TURE The balance of nature is delicate and easily upset.

LIMIT We are approaching the limit to the number of people this earth can
support.

CONSEQ When humans interfere with nature, it often produces disastrous

consequences.

After forming the initial factor matrix, eigenvalues for each factor and the

percentage of variance with each successive factor were examined (See Table 2.2).

Eigenvalues are greater than one for factors one through six and become less than one

starting at factor seven. According to the latent roots criterion, factors should be retained

up to the point where eigenvalues become less than one. According to this criterion, six

factors should be retained for the analysis.

The percentage of variance extracted by successive factors is another criterion

used to determine the number of factors to retain in the analysis. This is a summary

measure that indicates the amount of total variance of all input variables explained by

each factor (Aaker and Day 1986). It is common in the social sciences to retain factor

solutions that account for at least 60 percent of total variance (Hair et al.1998).

51

Table 2.2. Total Variance Explained

 

 

 

 

 

 

 

 

Factor Initial Eigenvalues
Total % of Variance Cumulative %
1 4.494 22.468 22.468
2 3.376 16.882 39.350
3 1.580 7.902 47.252
4 1.309 6.547 53.799
5 1.086 5.432 59.231
6 1.070 5.349 64.581

 

Another criterion for factor retention requires an examination of the scree
diagram, which plots eigenvalues against factors in the order they are extracted (Hair et
al. 1998). In component factor analysis, the proportion of common variance to unique
variance changes as more factors are extracted, and the scree test provides a way to
identify the point where unique variance begins to outweigh common variance (ibid.).
The point at which this curve becomes a flat line is considered the cut-off point for the
number of factors to extract (Churchill 1991). In Figure 2.1, this cut-off point can be
seen between factors (or components, as indicated in the diagram) 5 and 6 where the line
becomes visibly flat. According to the scree test criterion, a total of 5 factors should be

retained in the solution.

52

Figure 2.2. Scree Plot of Eigenvalues and Factors

Scree Plot

 

j—I

Eigenvalue
I.»
J

 

 

 

lllllTlllllllll
67891011121314151617181920

Component Number

I r
2

j
l 5

ll
34

The latent root criterion suggests retaining 6 factors. The percentage of variance
criterion suggests retaining 6 factors as well, however, almost 60 percent of variance is
extracted by 5 factors (59.23%), and retaining a 5-factor solution would closely meet this
criterion. The scree test suggests retaining 5 factors, and therefore, 5 factors were
retained in the solution. After deciding on the number of factors to retain in the solution,
a VARIMAX orthogonal rotation was applied to the first 5 factors. The following table

reports factor loadings from the rotated factor matrix.

53

Table 2.3. Rotated Factor Matrix

 

Variable Factor 1 Factor 2 Factor 3 Factor 4 Factor 5
' ' -0.172 0.295

' 5C» .
a . Y ‘ ‘ v, -

   

HUIWVEED -0.0 1 2
WLFNEED 0.06 l

         
    
   

 

 

     
  
  
   
 

   

 

     

 

 

 

 

 

COWEED -0273 -0.268 7 -0.036 0.308
BALNEED 512:2)» , -O.261 0.110
FUTGENNEED 0.224 ”if” m .0031 0.167
MTNUVDJOB 0.17s -0.l76 0.070
MTNRECJOB 0.207

 

 
  

PROTRES
BALJOBRES

   

0.272

 

 

   

 

 

 

  

INCRDIV -0.065 0.152 -0.237 -0057
BALINDDIV I... "‘ g 0.067 0.019 0.103 0.000
PROTHAB -0.007 0.206 -0.214 -0.078
BALINDHAB 7" ' . ' 0.005 0.088 -0.045

 
 

 

  

 

 

 

 

 

 

 

 

 

 

 

HUMBUSE -0.084 0.257
HUMRIGHT -0.095 0.1 19 -0.086
HUT/[RULE -0.080 0.091 -0. 179
HUMLEARN 0.194 0.020 0.030
BALNATURE 0.119 8.75e-005 0.133
LIMIT -0.039 -0.206 0.1 19
CONSEQ 0.035 -0.104 0.158

 

 

A factor loading of 0.35 or greater indicates a significant loading value; therefore
factors are determined by variables with loading values of 0.35 or greater (Dietz et al.
1998). These variables are indicated by the highlighted cells in Table 2.3. Four
variables, WLFNEED, BALNEED, FUTGENNEED and PROTRES all load highly onto

more than one factor. It is common in factor analysis for variables to have moderate

54

loadings on more than one factor, and this result presents a challenge to the researcher in
that the interpretation of the resulting factors becomes more complex when single
variables are associated with multiple factors (Hair et al. 1998). However, in this study, a
logical interpretation can be applied to all of the factors that include variables that load
highly more than once.

When interpreting factor loadings, positive signs indicate that variables are
positively related and negative signs indicate a negative relationship between variables,
but this relationship only applies to variables within the same factor because orthogonal
factor solutions are independent (Hair et a1 1998). For example, the variable HUWEED
under Factor 1 has a negative sign. This sign indicates that responses to the HUMVEED
statement move in the opposite direction of responses to the other variables with positive
signs. In other words, a person who agrees strongly with BALNA TURE, LIMIT}
CONSEQ and HUMBUSE will tend to disagree with HUMVEED, and vice versa. This
interpretation makes intuitive sense based on the definitions of the attitude statements
(See Table 2.1).

Five factors have been extracted from the factor analysis and each factor contains
several variables with high factor loadings. Each factor needs to be described based on
these variables and given a descriptive name. The factor description is based on what is
meant by a high score response to the statements (variables) included in that factor. High
score responses are those that have scores of 4 or 5, meaning disagree or strongly
disagree, respectively. The following paragraphs describe the interpretation of each
factor.

Variables that load highly onto Factor 1 relate to achieving a balance between
economic and ecological goals, BALNEED, BALJOBRES, BALINDDI V and

55

BALHVDHAB. Factor 1 is labeled the “Balance” factor. All of the variables that load
highly onto this factor have positive signs, indicating that responses for these variables
move in the same direction. High scores (indicating disagreement) for these variables
indicate that the respondent does not believe it is important to achieve balance between
ecological and economic needs of forested communities, and low scores (indicating
agreement) indicate that the respondent feels that balancing ecological and economic
goals of forest management is important. It is ambiguous whether high scores on this
factor reveal anthropocentric or biocentric values. However, low scores on this factor
reflect a combination of anthropocentric and biocentric values.

The variables with high loading values on Factor 2 relate to forest management
that benefits wildlife and future generations (WLFNEED, FUTGENNEED, PROTRES)
and to how humans affect the environment (HUMABUSE, BALNA TURE, LIMIT and
CONSEQ). Factor 2 is labeled the “Environment” factor. High scores (indicating
disagreement) for responses to all of these attitude statements indicate that the respondent
feels that wildlife resources should not be protected under forest management, forests
should not be protected for future generations, and that environmental problems are not
serious. High scores on this factor reveal an anthropocentric value orientation, and low
scores reveal a biocentric value orientation.

Factor 3 variables relate to forest-related jobs and resource dependent
communities, HUIWVEED, COMIWVEED, BALNEED, FUTGENNEED, MTNHVDJOB,
and MTNRECIOB, and all variables have positive signs. This factor is labeled the
“Management” factor. High scores (indicating disagreement) on this factor indicate
disagreement with forest management goals for humans, resource dependent

communities, for balancing economic and ecological goals, for future generations and for

56

forest-based jobs. This factor seems to reflect attitudes of individuals who believe that
forests should not be managed for human needs at all. High scores on this factor reflect a
strongly biocentric value orientation.

All of the variables that load highly onto Factor 4 are related to wildlife and
wildlife habitat: WLFNEED, PROT RES, HVCRDI V and PROTHAB. As with Factor 2,
all highly loading variables have positive signs, indicating that responses move together.
This factor is labeled the “Wildlife” factor. High scores on these variables indicate that
the respondent feels that forests should not be managed to meet the needs of wildlife or to
protect forest and wildlife resources, and that ecological goals such as increasing
songbird diversity or protecting songbird habitat should not be met if it imposes costs on
forest industries. High scores on this factor reflect an anthropocentric value orientation
and low scores reflect a biocentric orientation.

The variables that load highly onto Factor 5 concern human needs and the role of
humans with respect to the natural environment: HUMRIGHT, HUMRULE and
HUMLEARN. This factor is labeled the “Human Role” factor. These variables move in
the same direction because they all have positive signs. High scores on this factor
indicate disagreement with statements that humans have the right to modify the natural
environment to suit their needs, that humans were meant to rule over nature and that
humans will learn enough about nature to be able to control it. High scores on this factor
reflect a biocentric environmental value orientation.

The following table summarizes the 5 factors that have been identified to show

which variables fall under each factor description.

57

Table 2.4. Factors, Associated Variables and Factor Score Meanings

 

 

 

 

 

 

Factor Variables Meaning of Low Factor Scores
Included in (Opposite meaning is applicable for
Factor high factor scores)

Balance BALNEED Achieving balance between ecological and
BALJOBRES economic needs of forested communities is
BALHVDDI V important
BALHVDHAB (Ambiguous whether anthropocentric or

biocentric)

Environment WLFNEED Forest management should benefit wildlife and
F UTGENNEED future generations; Humans are abusing the
PROT RES environment and this has potentially harmful
HUMABUSE consequences
BALNA TURE (Biocentric)
LIMIT
CONSEQ

Management HUMNEED Forests should be managed to meet the needs of
COMWEED humans, the needs of resource dependent
BALNEED communities, to balance economic and ecological
F UTGENNEED goals, for future generations and for forest-based
MTNHVDJOB jobs
MTNRECIOB (Anthropocentric)

Wildlife WLFNEED Forests should be managed to meet the needs of
PROT RES wildlife, to protect wildlife resources, and should
INCRDI V improve bird habitat even at an economic cost
PROTHAB (Biocentric)

Human Role H UMRIGHT Humans have the right to modify the environment,
H UMR ULE humans were meant to rule over nature, and
HUMLEARN humans will learn enough about nature to be able

to control it
(A nthropocentric)

The factors extracted from the data from this analysis summarize forest and

environmental attitudes of survey respondents and succeed in reducing the amount of

data required for subsequent analysis. The results indicate that there is an underlying

structure to the way in which people respond to these attitude statements. The variables

58

that load highly on each factor form logical groupings, which allows logical descriptions
to be assigned to all resulting factors. The factors also reflect forest and environmental

value orientations.

2.4 Factor Score Regression Results

Factor loadings that resulted from the analysis in section 2.3 were converted into
factor scores for each respondent using SPSS software (SPSS 13.0). These factor scores
are used as dependent variables in a set of models estimated using stepwise OLS
regression. Factor scores for each of the 5 factors extracted from the data (See Table 4)
are regressed onto a set of socioeconomic variables collected from the surveys, which are

listed in Table 2.5.

Table 2.5. Definitions of Variables Used in Factor Score Regression Models

 

 

 

 

 

 

 

Variable Definition

SA Dummy variable equal to 1 if respondent lives in the study area, 0
otherwise

F ORORG Dummy variable equal to 1 if respondent is member of a forestry
organization, 0 otherwise

EN VORG Dummy variable equal to 1 if respondent is member of an
environmental organization, 0 otherwise

AGE Variable indicating respondent's age

GENDER Dummy variable equal to 1 if respondent is male, 0 otherwise

POL VIEW Dummy variable equal to 1 if reported political view is conservative,
0 otherwise

 

The stepwise regression included all socioeconomic variables collected from

survey data, but results only report those variables that were significant in the regression

59

analysis. Other variables that were included but were not significant include the

following: property ownership in study forest, membership in hunting club, education,

income, employment in resource based industry, urban/rural residence, ethnic

background and religion. The results of 5 linear multivariate stepwise regression models

estimated with the remaining socioeconomic variables are discussed below and reported

in Table 2.6. The original data set of 954 responses was reduced to a total of 563

observations in the factor score regressions. This number, 563, reflects the number of

survey respodents for whom observations were complete for all variables included in the

factor analysis and the socioeconomic variables included in the regressions.

Table 2.6. Factor Score OLS Stepwise Regression Results

 

 

 

 

 

 

 

 

 

 

 

 

Model 1 Model 2 Model 3 Model 4 Model 5
Balance Envirmt Mgmt Wildlife HumanRole
SA -0.188** 0252*"
FORORG 0582"" 0544*"
ENVORG -0.397**
AGE 0.300":-
GENDER 0405*"
POL VIEW 0308*" -0.250"'"'"‘ -0.177"
Intercept -0.324"”"* -0.1 14"
i ——
N 563 563 563 563 563
adj. R2 0.024 0.021 0.03 0.045 0.006
F value 14.891""‘ 12.889“"‘ 6888*" 1657*“ 4.149"

 

* p<0. 10,- ** p<0.05,' ***p<0.01

In Model I, GENDER is statistically significant with a positive sign, indicating

that men are more likely to disagree with forest management goals that emphasize

60

achieving a balance between economic and ecological goals than women.

In Model 2, POL VIEW is statistically significant with a positive sign. This
suggests that politically conservative individuals are more likely to disagree with attitude
statements within the “Environment” factor. Individuals who are politically conservative
are more likely to disagree with forest management for wildlife needs, future generation
needs, and protection of forest and wildlife resources. They are likely to disagree that
humans are severely abusing the environment, that the balance of nature is easily upset,
that there is a limit to the number of people the earth can support, and that human
interference with nature often produces disastrous consequences.

Three socioeconomic variables are statistically significant in Model 3: SA,

F ORORG and POL VIEW. All three variables have negative signs, indicating that
individuals with these characteristics are more likely to agree with the attitude statements
within the “Management” factor. Residents of the Study Area, members of forestry
organizations and politically conservative individuals are more likely to agree that forests
should be managed to meet human needs, the needs of communities dependent on forests,
to balance environmental needs with the needs of resource-dependent communities, to
meet the needs of future generations and to maintain forest industry and forest-based
recreation and tourism jobs.

In Model 4, SA, F ORORG, and AGE are statistically significant with positive
signs, and EN VORG is significant with a negative sign. This suggests that residents of
the Study Area, members of forestry organizations and individuals over the age of 60
tend to disagree with forest management that meets wildlife needs, protects forest and
wildlife resources, increases forest migratory songbird diversity and protects habitat for

songbirds of conservation concern. Members of environmental organizations, by

61

contrast, tend to agree with these forest management goals.

The only statistically significant variable in Model 5 is POL VIEW, which has a
negative sign. Politically conservative individuals tend to agree that humans have the
right to modify the environment to suit their needs, humans were meant to rule over
nature, and that humans will learn enough about nature to be able to control it.

Although an F-test rejects the hypothesis that all coefficients are jointly equal to
zero for all models, each model has fairly low explanatory power. Models 3 and 4 had
slightly better explanatory power than models 1, 2 and 5 as judged by the slightly higher

R2 values, which represent the percent of variation explained by the regression.

2.5 Discussion

Model results suggest that residents of the study area and members of forestry
organizations are more likely to disagree with forest management solely for the benefit of
wildlife and are likely to agree with forest management that meets human needs. This
result suggests that study area residents hold anthropocentric views of forest management
and the environment, which contradicts results of the qualitative analysis in Essay 1.
This result may be the case because the participants in focus group discussions had the
opportunity to provide richer details of their thoughts and opinions. Survey respondents
from the study area, when faced with a more impersonal mail questionnaire format, may
have been more likely than focus group participants to express strong anthropocentric
views because of their economic dependence on the timber industry.

Members of environmental organizations tend to favor forest management that
benefits forest and wildlife resources. This response suggests a biocentric value

orientation for members of environmental organizations. Older individuals do not favor

62

forest management that benefits forest and wildlife resources, and this result may reflect
a more anthropocentric value orientation of this demographic group. Men are more
likely than women to reject forest management that tries to achieve balance between
economic and ecological goals. This result reflects neither an anthropocentric nor a
biocentric value orientation clearly. Politically conservative individuals do not favor
forest management for protection of forest and wildlife resources and feel that humans
are not adversely affecting the environment. This demographic group also tends to agree
with forest management that meets human needs and feels that humans have the right to
use the environment for their own purposes and to rule over nature. These results reflect

a strong anthropocentric value orientation for this demographic group.

2.6 Conclusions

This research analyzes attitudinal data from a mail survey of Michigan residents
in order to determine whether attitudes reveal underlying forest and environmental values
and whether or not demographic characteristics can be used to describe individuals who
hold certain value orientations. There were two original research hypotheses: 1) there is
an underlying structure to the way in which individuals respond to forest management
- and environmental attitude questions that reveals information about individuals’
environmental values, and 2) these environmental values can be explained by a set of
socioeconomic characteristics.

The first hypothesis was tested by conducting exploratory factor analysis on a set
of variables that measure attitudes towards forest management in a particular area of
Michigan’s Upper Peninsula as well as general environmental attitudes. The results of

factor analysis confirm the hypothesis that there is an underlying structure to the

63

responses to these attitude statements. The five factors resulting from the exploratory
factor analysis are labeled Balance, Environment, Management, Wildlife and Human
Role. Each factor, with the exception of Balance, can be interpreted within the
anthropocentric/biocentric value framework according to the level of agreement or
disagreement with attitude statements associated with that factor.

The second hypothesis was tested by converting results of factor analysis into
factor scores for each individual in the sample and running a multivariate regression of
factor scores for each factor against a set of socioeconomic characteristics. A set of
socioeconomic characteristics was used as explanatory variables in a stepwise regression,
which reduced the number of explanatory socioeconomic variables. Each factor model
had at least one significant socioeconomic explanatory variable, which supports the
hypothesis that environmental value orientations can be explained in part by
socioeconomic characteristics. Regression analysis of computed factor scores for each
factor indicate that certain demographic groups tend to hold anthropocentric views of
forests and the environment while others tend to hold more biocentric views. Results
suggest that residents of the Study Area, members of forestry organizations, older
individuals, politically conservative individuals tend to hold anthropocentric views while
members of environmental organizations tend to hold biocentric views of forests and the
environment. The value orientation of men’s views is ambiguous.

Information on forest management and environmental attitudes and the
underlying values and characteristics of people that express these attitudes comprise an
important additional input to be considered in natural resource program and policy
evaluation. This research has established connections between certain socioeconomic

characteristics and held environmental and forest values and as such provides a means of

64

predicting attitudes and the consumer and voter behavior of different segments of the
public. Results of the regression analysis of factor scores on demographic characteristics,
however, did not demonstrate great explanatory power. This result suggests that
demographic factors may not be sufficient to fully explain which segments of the public
hold particular forest and environmental value orientations. Future research should
investigate additional factors that may differentiate groups that hold different forest and
environmental value orientations. For example, connections may exist between attitudes,
values, and pro-environmental consumer behavior or knowledge of environmental issues.
The research results do provide some guidance, nevertheless, that can inform
natural resource management and policy design. For example, when conducting a
comparison of forest easement programs to be implemented in the study area, this
research suggests that members of forestry organizations will be more likely to favor
programs that enhance opportunities for humans whereas environmentalists will tend to
prefer programs that protect wildlife and forest resources. Designing a program that
combines the preferences of both demographic groups can lead to higher public
acceptance and therefore higher success of the program. Results also suggest that
proposed programs within the study area would need to emphasize the human or
instrumental benefits of the program in order to in order to achieve broad public
acceptability in the region. Voluntary forest landowner incentive programs, such as
FLEP (Forest Land Enhancement Program), present a means of generating ecological as
well as economic benefits from forestry. Results suggest that efforts to disseminate
information in the study region to gain membership in this program would need to focus
heavily on sustaining long term economic benefits to landowners as well as achieving

ecological goals.

65

REFERENCES

Aaker, D.A., and GS. Day. 1986. Marketing Research 3rd Edition. New York: John
Wiley & Sons.

Bengston, D., G. Xu, and D. Fan. 2001. Attitudes toward ecosystem management in the
United States, 1992-1998. Society and Natural Resources 14:471-487.

Churchill, G.A. Jr. 1991. Marketing Research: Methodological Foundations 5th Edition.
Chicago: The Dryden Press Harcourt Brace College Publishers.

de Groot, R.S., M.A. Wilson, and R.M.J. Boumans. 2002. A typology for the
classification, description and valuation of ecosystem functions, goods and
services. Ecological Economics 41 :393—408.

Dietz, T., L. Kalof, and P.C. Stern. 2002. Gender, Values, and Environmentalism. Social
Science Quarterly 83 (1):353-364.

Dietz, T., P.C. Stern, and G.A. Guagnano. 1998. Social structural and social
psychological bases of environmental concern. Environment and Behavior
30:450-471.

Dillman, DA. 2000. Mail and Internet Surveys: The Tailored Design Method, 2nd
edition. 2nd ed. New York: John Wiley and Sons, Inc.

Dunlap, R.E., and K.D. VanLiere. 1978. New Environmental Paradigm. Journal of
Environmental Education. 9(4): 10-19.

Fishbein, M., and I. Ajzen. 1975. Belief Attitude, Intention and Behavior: An
Introduction to Theory and Research. Reading, MA: Addison-Wesley.

Hair, J .F . Jr., R.E. Anderson, R.L. Tatharn, and W.C. Black. 1998. Multivariate Data
Analysis 5th Edition. Upper Saddle River, NJ: Prentice Hall.

Honnold, J.A. 1981. Predictors of public environmental concern in the 19705. In
Environmental Policy Formation: The Impact of Values, Ideology, and Standards,
edited by D. E. Mann. Lexington, MA: Lexington Books, D.C. Heath and
Company.

Jacobson, S.K., and SB. Marynowski. 1997. Public attitudes and knowledge about

ecosystem management on Department of Defense land in Florida. Conservation
Biology 11 (3):?70-781.

66

Jones, R.E., J.M. Fly, and H.K. Cordell. 1999. How green is my valley? Tracking rural
and urban environmentalism in the southern Appalachian Ecoregion. Rural
Sociology 64 (3):482-499.

MA (Millenium Ecosystem Assessment). 2003. Ecosystems and Human Well Being: A
Framework for Assessment. Washington, DC: Island Press.

McFarlane, BL. and P.C. Boxall. 2000. Factors Influencing Forest Values and Attitudes
of Two Stakeholder Groups: The Case of the Foothills Model Forest, Alberta,
Canada. Society and Natural Resources 13:649-661.

Potter, H.R., and HI. Norville. 1981. Social values inherent in policy statements: An
evaluation of an energy technology assessment. In Environmental Policy
Formation: The Impact of Values, Ideology, and Standards, edited by D. E.
Mann. Lexington, MA: Lexington Books, D.C. Heath and Company.

Reading, R.P., T.W. Clark, and SR. Kellert. 1994. Attitiudes and knowledge of people
living in the Greater Yellowstone Ecosystem. Society and Natural Resources 7
(4):349-365.

Rokeach, M. 1973. The Nature of Human Values. New York: Free Press.

Solecki, W.D. 1998. Local attitudes on regional ecosystem management: A study of New
Jersey Pinelands residents. Society and Natural Resources 11 (5):441-463.

SPSS 13.0. Chicago, IL.

STATA Corporation. 2003. STA TA Base Reference Manual A-F Release 8. Vol. 1.
College Station, TX: STATA Press.

Steel, 18., P. List, and B. Shindler. 1994. Conflicting values about federal forests: A
comparison of national and Oregon publics. Society and Natural Resources
7 :137-1 53.

Stem, P.C., T. Dietz, and G.A. Guagnano. 1995. The New Ecological Paradigm in social-
psychological context. Environment and Behavior 27 (6):723-743.

Tarrant, M., and H. Cordell. 2002. Amenity values of public and private forests:
Examining the value-attitude relationship. Environmental Management 30
(5):692-703.

Weiler, K.S., and J .T. O'Leary. 1997. Demographic trends and forest resource
implications for the Lake States. In Lake States Regional Forest Resources
Assessment: Technical Papers, edited by J. M. V. a. H. H. Webster. St. Paul, MN:
USDA Forest Service North Central Forest Experiment Station General Technical
Report NC-189.

67

Xu, W., B.R. Lippke, and J. Perez-Garcia. 2003. Valuing biodiversity, aesthetics, and lob

losses associated with ecosystem management using stated preferences. Forest
Science 49 (2):247-257.

Xu, W., and UN. Bengston. 1997. Trends in national forest values among forestry
professionals, environmentalists, and the news media, 1982-1993. Society and
Natural Resources 1 0(1):43-59.

68

Essay 3:
Valuing Forest Ecosystem Characteristics:
An Attribute-Based Contingent Valuation Approach

3.1 Introduction

Forest ecosystem services, such as wildlife habitat and biodiversity, are public
goods that are not traded in a market, making it difficult to place a value on them. Public
goods are goods for which property rights are not defined and thus are available to all
individuals for consumption (Mitchell and Carson 1989). Even though no market values
exist for these goods, members of the public may hold values for them. Estimates of
these values, as well as pubic preferences for these services, can be useful to policy
makers and resource managers in evaluating the tradeoffs involved in natural resource
allocation decisions. This study estimates public preferences and values for forest
ecosystem services and the tradeoffs between them in an area of Michigan’s Upper
Peninsula (UP), referred to in this essay as the study forest (See Essay 1 for a map and
description of the study forest).

The communities in this region are economically dependent on the forests in the
area for timber, recreation and tourism industries (See Essay 1 for details). There are
numerous complex interactions between humans, wildlife and forests in the study forest,
and these interactions are, in turn, connected with ecological and economic trade-offs in
the region. For example, studies by other researchers have shown that increases in
intensive timber harvesting can increase deer density from the creation of excess browse,
which, in turn, causes deer to overbrowse tree saplings, reducing future tree regeneration
and reducing habitat for certain forest birds in this region (Laurent et al. in prep, Shi et al.

2005). An evaluation of public preferences for these trade-offs can provide useful

69

information for natural resource policy makers.

At the time this study was undertaken, the study forest, which is a heavily
forested area, was owned primarily by the state (42%) and private forest industry (43%).
However, over the past few years, hundreds of thousands of acres of land in this area
have been converted to other non-industrial private land uses (Potter-Witter 2005).

These recent changes in forest land ownership in the area have implications for the way
the forests are managed, which, in turn, affects the types of ecosystem services provided
by these forests. For example, lands transferred to non-industrial private ownership may
no longer be managed for timber, which may affect local forest industry jobs. The
conversion of previous large forest land holdings to small parcels of forest land also has
implications for wildlife habitat. For example, industrial forest lands that are sold and no
longer managed to grow a particular tree species may reduce habitat for certain types of
forest songbirds. Information on public values for ecosystem services in the study forest
is needed to understand the potential benefits and costs of the changing forest
landownership and management patterns in this region.

A variety of economic techniques exist that provide ways of estimating values for
public goods by asking individuals to state their preferences for the provision of an
environmental good (Carson 2000). Contingent Valuation (CV) is a stated preference
method that uses survey questions to elicit an individual’s willingness to pay (WTP) to
achieve an environmental improvement or to avoid an environmental injury (Mitchell and
Carson 1989, Bennett and Adamowicz 2001). The CV method presents respondents with
a hypothetical, or constructed, market that provides information about the environmental
good to be valued, how it will be provided, how it will be paid for, and asks the

respondent to make a decision about the provision of that good (Mitchell 2002).
70

CV has been used extensively to estimate benefits of environmental
improvements and to assess natural resource damages (Carson 2000). CV studies have
been used to estimate the benefits of forest ecosystem services on non-industrial private
forest lands (Stevens et al. 2000). Studies have estimated the value of protecting forest
land (vanRensburg et al. 2002, Loomis et al. 1994) and the non-market values of forest
attributes (Hanley and Ruffell 1993). This study uses ecological information to develop a
stated preference survey to estimate the economic value of forest ecosystem attributes
that are not valued in the market. The survey results yield values for the estimated
marginal utilities of forest ecosystem attributes, and these values are interpreted to

analyze the relative benefits of a set of forest ecosystem services.

3.2 Attribute-Based Referenda Model

Research on non-market values of managed forest ecosystems naturally lends
itself to a multiattribute approach because of the numerous characteristics of forests
managed for multiple uses. Attribute-based methods (ABMs) are growing in popularity
as an alternative to the traditional CVM, which has been the most commonly used
method for measuring passive use values (Adamowic and Boxall 2001, Holmes and
Adamowicz 2003, Holmes and Boyle 2005). Like the CVM, ABMs are based in random
utility theory, but they focus on sets of environmental policy-relevant attributes, along
with cost, as opposed to one total value, which is the focus of traditional CV studies
(Hanley et al. 1998, Bennett and Blarney 2001, Holmes and Boyle 2005).

Numerous studies have compared traditional CVM with ABMs and have
concluded that there are several advantages of using ABMs to estimate values of

environmental goods with multiple attributes (Boxall et al. 1996, Hanley et al. 1998).. A

71

commonly used ABM is the choice experiment (CE), which is a non-market valuation
method that is well suited for the estimation of marginal values of environmental
attributes (Boxall et al. 1996, Hanley et al. 1998, Lupi et al. 2002, Stevens et al. 2000).

Another type of ABM is the attribute-based referenda model (ABR), which is a
hybrid of contingent valuation and attribute-based method stated preference questions
(Holmes and Boyle 2005). This method uses an attribute-based description of an
environmental good or service and a referendum-style choice between the status quo and
a policy alternative to the status quo.

The ABR model used in this study is based on a contingent market for an
environmental good that is described in terms of multiple attributes. The contingent
market used in this survey is a political market that presents respondents with a decision
to vote ‘yes’ or ‘no’ to a forest and wildlife protection program for the Study Forest.
This is referred to as a referendum format, or “Take-It-or-Leave-It” approach, and it
shows the program and its costs and elicits the respondent's vote (Mitchell and Carson
1989). It is chosen because of the nature of the environmental valuation question, which
involves the protection of forest ecosystem resources in a particular geographic area. This
approach is desirable because it mimics voting choices that consumers make in the real
world, and therefore is more realistic to respondents and is not as susceptible to strategic
behavior on the part of respondents as other elicitation methods (ibid). Other reasons for
favoring this approach include that people are generally familiar with voting procedures,
voting decisions present realistic economic consequences to consumers’ households, a
ballot is easy to convey through a mail survey, and public goods are paid for collectively
(ibid).

The theoretical basis for the ABR method, like CVM and other ABMs, comes

72

from random utility theory (Adamowicz and Boxall 2001, McFadden 1974). According
to random utility theory (RUT), utility consists of an observable and an unobservable
component. It is possible to explain a large part of the essentially unobservable
consumer utility if a valid procedure for preference elicitation is used (Louviere 2001).
However, there will always be a random portion of utility that remains unexplained.
Indirect utility, u, is the maximum amount of utility that a household can derive
from income, y, given prices of goods, a vector of environmental quality variables, x,
other socioeconomic factors, 7,, and a component of individual preferences, 8, known to

the individual but not to the researcher.

u = u(y,x,z,8) (1)

In an ABR model, respondents are asked if they are willing to pay a certain
amount to achieve an environmental quality improvement. In this model, the quality
improvement is described by changes in the levels of attributes of a forested ecosystem
that will be provided by a program at a cost to the respondent. The indirect utility when

an amount p is paid is:

”I =u(xraZ’y—P,31). (2)

In this equation, u 1 represents the indirect utility function for an individual who
pays the cost of the program, and x I is a vector of forest ecosystem attributes under the
forest protection program. If the cost, p, of the program is not paid, the indirect utility

function is:

uo = U(x0,z,y, 80) . (3)

73

In this equation, “0 represents indirect utility under the status quo, and x0 is the vector of
forest attribute levels without the program. An individual will be willing to pay for the
proposed program if:

ul(xl’z’y-p981)'>'u0(x02zsy980)- (4)

The probability that a respondent is willing to pay p for the forest protection program
(probability of saying yes) is given by the probability that the utility received from the

forest protection program is greater than the utility received under the status quo:

Prryes) = Prlutrxnze— as.) > uo(xo.z.y.eo>l
=Pr[Au>0]. <5)

The indirect utility function has an unobservable, random component. Let indirect utility
of individual i from alternative j be expressed by the sum of its explainable and

unexplainable components:

uU=vij+e,-j, (6)

where vy- is the explainable component of utility to individual i from alternative j, and c is
the unexplainable, random component of utility for individual i from alternative j.

The deterministic component of indirect utility can be written as:

vyzaxj+71zi+fl(yinpj) a Yj=7mvjrm¢os (7)

where x} is a vector of k attributes associated with alternative j, a is a vector of estimable

parameters, [1 is an estimable parameter, y, is income of respondent i, pj is the price paid

74

for alternative j, z, is a vector of socioeconomic characteristics of respondent i and y, is a
vector of estimable parameters for alternative j for the effect of respondent

socioeconomic attributes. Using (6), equation (5) can be rewritten as

Pr(yes) = Pr[v,.j + 8:; > v,0 + 3,0], (3)

Substituting (7) for the deterministic component of indirect utility in (8) yields the

following result:

Pr(yes) = Prlaxj +717.) +B(y, —pj)+8ij >ox0 +7011 +13}? +810]
=P1'[w‘} "’xo +7131 ‘7011 ‘13P) >510-81‘j] (9)
=P’I“(Ax1)+731'l3pj >810 ’80 ' >

It is typically assumed that the marginal utility of income does not change between the
two states, and this assumption allows the By,- terrns to be dropped from the second part of
(9). The last equation in (9) can be estimated by making certain assumptions about the
error terms. A common assumption is that the errors follow a standard normal
distribution. When this assumption is made, the probability of an individual choosing

alternative j over the status quo is given by the standard probit model:

Pr[a(ij)+y Zr _Bpj > 810 ’55]: <D[a(Ax1)+y 3: "Ble (10)

where <1) is the cumulative distribution function (cdf) of the standard normal distribution,
assuming that o = 1. Equation (10) can be estimated using the maximum likelihood
estimation procedure for a probit model. Because of the discrete nature of the choices, 0

cannot be identified in a probit, but it is customary to normalize the coefficients so that

75

o = 1. The probability of discrete choice c, (c,- = 1 for a yes response and C,- =O for a no
response) given attributes xi is denoted as:
<D(a(Axr )+ t z: - lip,- i“ w
Pr(c,|x, z, y, p): I. Hi) - ‘ (11)
[1‘ ‘D(“(Axr )+ 7 z, ‘ 5P; )I
where <1), again, is the cdf of the standard normal distribution. The log-likelihood

function for the probit model is written as:

_ n c, log[<l>(a(ij )+ y z, - 0p; )]
L(a, Bjx, zi)’, P) - z{+(1— Ci)10g[1 _ @64ij )+ ’Y Z: _ PP} )1}. (12)

i=1

An assumption of the standard probit model is that the error component is
independent and identically distributed among individuals and across observations for
each individual. However, when an individual responds to more than one stated
preference question, it is likely that there are unobservable characteristics specific to that
individual that induce correlation across her responses. If this suspected to be the case, it
is appropriate to estimate a random effects probit model (Wooldridge 2002). In a random
effects model, the error term is treated as separable into two components: one that is
unobservable and specific to each individual and another that is unobservable and due to
random response shocks across all individuals and all responses (Boxall et al. 2003). The

random effects utility difference model can be written as:

Auy=a(ij)+yz,—Bpj+ui+a,j, (13)

where u,- is the individual-specific error term, and 8,]- is the random disturbance term

76

across all individuals and observations. The random effects probit likelihood function

can be written as follows (adapted from Boxall et al. 2003):

 

-nn. ill—J motels-i

as

(14)
In this equation, dr is the cdf of the standard normal distribution, and t indexes the choice
made by each individual i. The correlation coefficient between the responses of

individual i is defined as:

of. 6?.
9:77—7:7- (15)
(38+op o

If there is no correlation among the disturbance terms within individuals, p will not be
significantly different from zero and the standard probit model in which the data is
pooled is appropriate. If, however, p is significantly different from zero, the standard
probit model yields biased standard errors of the estimated coefficients, making it
necessary to use the random effects probit model (Boxall et al. 2003). Results of the

random effects probit models are reported in section 3.6.

3.2.1 Model Specification
Equations (9) and (13) can be estimated using data collected from a stated

preference survey. If we assume that an individual has a linear conditional indirect utility

77

function, that he gains utility from the forest ecosystem in the study forest, and that his
utility can be decomposed into different ecosystem attributes, we can specify the
following utility difference function (See Table 3.1 for variable definitions):

Aug“ = 11(ij )- 5P1

= (10 + alAindiobsj + (tzArtjobsj + 113Abirddivj (16)

+ (14Abirdconsj + (tsAdeerj + [ilcostj .

(N019: Pr = ‘ 13)

where “k and [3 represent estimable parameters. Stated preference data collected from a
mail survey of Michigan residents are used to estimate the model parameters using
random effects probit estimation. Individuals are asked to state in the survey whether or
not they would vote for a program that provides specific levels of ecosystem services in
the study forest. The data from individuals' yes/no responses along with associated
attribute levels with or without the program are inputted to estimate equation (16).
Estimation results are presented in section 3.6.

Equation (16) only includes the attributes of the forest easement program that are
used to describe the program scenarios. It is also possible to include socioeconomic
characteristics, 2,, of survey respondents in order to determine whether certain
characteristics are important in predicting a yes response to the easement program. To
account for the possible effects of socioeconomic attributes, the following model will
also be estimated:

Auzj = “(Mj)+7 311 ‘13P)-
: a0 + rxlAindjobsj + rszrtjobsj + (1311birddivj

+ 014Abirdconsj + (rsAdeerj + ylage, + yzhuntclub, (17)

+ 73envorg, + y4educ, + yspolview, + yéurban, + 77rec, +13, cost j .

78

Table 3.1. Variable Definitions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Definition

indjobs Number of forest industry jobs in the study forest

rtjobs Number of forest-based recreation and tourism jobs in the
study forest

birddiv Percent of study forest with high migratory forest songbird
species diversity

birdcons Number of migratory forest songbird species of conservation
concern that are at or above their target population level (out
of 19 possible species)

deer Percent of area with deer browse high enough to affect tree
regeneration

cost Cost to household in increased annual taxes

age Dummy variable that equals 1 if respondent is over age 60, 0
otherwise '

huntclub Dummy variable equal to 1 if respondent is a member of a
hunting club, 0 otherwise

envorg Dummy variable that equals 1 if respondent is member of an
environmental organization, 0 otherwise

educ Education dummy variable that equals 1 if respondent has
college education or above, 0 otherwise

polview Dummy variable that equals 1 if respondent self-reports to be
politically conservative, 0 otherwise

urban Dummy variable that equals 1 if respondent lives in urban
area, 0 otherwise

rec Dummy variable that equals 1 if respondent has participated

in recreational activities in or near the study area, 0 otherwise

It is hypothesized that certain socioeconomic characteristics of respondents affect

the probability of voting ‘yes’ for the forest easement program. The characteristics
included in the final model include those reported in Table 3.1. Other socioeconomic

characteristics that were examined in the model but were not statistically significant were

79

then dropped from the model. These included income, religious preference, ethnicity,
resource dependence and gender. The estimated coefficients from the random effects
probit models will be used to calculate welfare estimates for the forest ecosystem
attributes and the overall forest easement program. The following section describes the

welfare estimates to be calculated.

3.3 Welfare Estimates

The results of the model estimation will provide values with which welfare
measures can be calculated. It is commonly accepted in the non-market valuation
literature to use the concepts of compensating surplus and willingness to pay to calculate
and discuss welfare estimates (Haab and McConnell 2002). The amount of money that
makes an individual indifferent between the status quo and a state with an environmental
quality improvement is the individual’s willingness to pay for the improvement (Haab
and McConnell 2002). This is equivalent to an estimate of compensating surplus (CS),
which is an estimate of welfare change that gives the change in income for which an
individual would be indifferent between two states when he has an implied right to the
status quo (Morrison et al. 1999). This value of WTP is found at the point of equality

between indirect utility with and without the program, given by (18).

u0(x02z’y980)=ul(xlrz9y_WTP981)- (18)

Substituting the deterministic components of indirect utility for the status quo and the
state with the program, which are represented by the equations in (7), yields the
following result:

”10+Byi =Gxn+7zi+li(yi-WTP). (19)

80

Solving (19) for WTP yields:

 

WTP = —[a(x"' - :7)— Y z' J. (20)

01':

WTP={30_;J]. (21)

If the resulting change in the attributes is desirable, an individual’s income needs to be
reduced by the amount of WTP in order to maintain the utility level v0, implying a
positive WTP value (Morrison et al. 1999).

The ratio of each attribute coefficient, “k or yh, to the cost parameter estimate, [3,
yields marginal dollar estimates for each individual attribute (Hanemann 1994). These
values are referred to as implicit prices (IPs) of the attributes and represent the marginal
rate of substitution between an attribute and the cost attribute (Morrison et al. 1999). The

marginal value of a particular attribute can be calculated as follows:
[Pk = {Ti—k) for xk or -[%] for z, , (22)

where k indexes each attribute used in the choice design, a is the estimated coefficient of
any model attribute and [3 is the coefficient on the cost variable. The marginal rate of
substitution can be calculated in the same way between any two model attributes to
determine how much respondents are willing to give up of one attribute to have an

additional unit of another, without changing utility.

81

3.4 Survey Design

The presentation of information in a questionnaire can affect the responses
gained, and misinterpretation of survey information can cause serious problems in the
data gathered from the responses. The validity and usefulness of results from a CV study
rely in large part on carefully defining the environmental good being valued and placing
it within the appropriate policy context (Bishop and Heberlein 1992). Careful design of
stated preference surveys with multiple attributes is necessary to ensure that choice tasks
are realistic and understandable to respondents (Kaplowitz et al. 2004).

Designing the survey instrument involves a qualitative research phase in which
focus groups and individual interviews are both integral parts of the survey design
process (Kaplowitz and Hoehn 2001). Questionnaire development has been guided by
the results of six focus groups, 21 individual pre-test interviews, and interviews with
ecologists, foresters and state agency employees (See Essay 1 for details on some of the
focus group research related to questionnaire development). The questionnaire contains
sections on the economy of the study forest, wildlife in the study forest (focusing on
migratory forest songbirds and deer) and a proposed forest management program. The
attributes were chosen based on qualitative research findings and also to ensure that the
results of this research will be suitable for integration with results from the ecological
component of this project. The survey presents respondents with information treatments
for each attribute. The questions that follow each attribute’s information treatment help
to ensure that the respondent connects with the information provided for that attribute.

The questionnaire uses a forest easement program as the policy context for the
contingent market. This type of program was chosen for several reasons. Forest

easements are a form of conservation easement whose primary purpose is to ensure the

82

protection of conservation values (Ward and Ervin 2005). A conservation easement is
defined as a legal agreement in which a landowner voluntarily limits the use of her land
for the purposes of protecting specific conservation values (Lind 2001a). Working forest
easements provide a way of conserving the ecological resources of a forest while at the
same time ensuring the continued economic and social benefits generated by forests
(Lind 2001b). It is realistic to propose a forest easement as a referendum ballot issue
because similar programs have been implemented in several other states as a result of
successful voting referenda (New Hampshire, Maine, Vermont) (Newsom 2002). Forest
easements are also flexible programs that allow landowners to have input into the process
of designing a management plan (Ward and Ervin 2005). The voluntary nature of
easement programs is a characteristic that is potentially appealing to the general public.
This type of program has not been implemented in Michigan and is therefore not a
redundancy in this region.

The forest easement program is described in the survey using a set of attributes.
These attributes, listed in Table 3.2, are allowed to take on varying levels and form the
basis of the choice sets presented to respondents in the survey. These choice sets are
established using an experimental design procedure, which is a method of manipulating
attributes and their associated levels in such a way as to permit testing specific
hypotheses (Louviere et a1. 2000). Factorial designs combine each level of each attribute
with all levels of all other attributes and have the property that all attribute interactions
are independent (ibid). However, full factorial designs can generate extremely large
numbers of choice sets. Therefore, fractional factorial designs are employed because of
their ability to reduce the number of possible choice sets into a practical amount while

retaining the statistical properties of full factorial designs (ibid.). Each of the six forest

83

easement program attributes listed in Table 3.2 is allowed to take on three levels. The

experimental design of the 6 attributes is based on a 36 main effects design plan, which

produces 18 choice sets (Addelman and Kempthome 1961).

Table 3.2. Survey Attributes and Levels

 

Variable Attribute Name

Status Quo Attribute Levels

 

 

 

 

 

 

Name Level

indjobs Number of forest industry jobs in the 675 600, 675, 710
area

rtiobs Forest-based recreation and tourism 190 170, 190, 250
jobs in the area

birddiv Percent of area with high migratory 35% 38%, 55%, 75%
forest songbird species diversity .

birdcons Number of migratory songbird 6 7, 12, 17
species of conservation concern that
are at or above their target
population (out of 19 possible
species)

deer Percent of area with deer browse 69% 67%, 58%, 49%
high enough to affect tree
regeneration

cost Cost to your household in increased $0 $20, $90, $400

annual taxes

A description of how the forest easement program works is provided, and the

respondent is asked to vote ‘yes’ or ‘no’ to a series of proposed forest easement

programs. The choices are presented as tables in which one column lists program

attributes and the remaining two columns list the levels that each attribute takes with and

without the program (See Appendix 3 for a copy of the survey instrument). Several

follow-up questions collect additional information about respondent behavior. Following

84

the choice tables, a debriefing page asks respondents to indicate the level of importance
each attribute had in their decision-making and provides a space for the respondent to list
reasons for her choices. The debriefing information is important for use in determining
whether the respondent was interpreting the tasks correctly and whether or not they
understood the tasks, as well as verifying the consistency of their responses (Hanemann

1994)

3.4.1. Economic Attributes

Results of focus group discussions indicated that the public was concerned about
timber harvest methods and their effects on the economy, wildlife and recreation in the
region. The importance of the forest products industry as well as forests for personal
recreational use and generation of tourism revenues to support the local economy came
across strongly in the focus group discussions. Focus group results suggested that job
impacts would be an important consideration for individuals’ preferences. Based on this
feedback, we have included a variable for forest industry jobs in the region as well as
forest-based recreation jobs.

The use of jobs as an attribute may not intuitively seem like an appropriate
attribute for which to estimate a non-market value, because jobs have market values.
However, in the context of a resource-dependent community such as the one in our study
area, the value of a job related to forests has an importance to community well-being and
community health overall. Several studies use jobs as attributes in non-market valuation
studies based on the rationale that jobs have a social value (Morrison et al. 1999,
Lockwood et al. 1994, Xu et al. 2003, Adamowicz et al. 1998, Morrison et a1. 2002,

Portney 1994, Blarney et al. 2000).

85

Two variables were created to account for forest-related jobs. The first variable is
defined as Number of Forest Industry Jobs in the Area and the second is Forest-based
Recreation and Tourism Jobs in the Area. A question is included in the survey that asks
the respondent whether he thinks that a change in the number of forest-related jobs in the
study forest will affect his own income. Data from these responses is used as a proxy to
ensure that respondent income is measured correctly. A majority of respondents (62%)
feel that a change in the number of forest-related jobs would not affect their own income.

Respondents are presented with information about forest industry and forest-
based recreation and tourism jobs in the survey. The information treatments are designed
to help the survey respondent understand and respond to the choice tables later in the
survey in which the forest job variables are listed as choice attributes. The baseline
number of forest industry jobs in the area was estimated using data from the US. Census
for study area zip codes and townships. Employment data for all logging and wood
product manufacturing sectors was tabulated for the zip codes that fall within the study
area boundaries. See Appendix 2 for details of these calculations.

F orest-based recreation and tourism employment encompasses many subsectors,
including retail trade, food and accommodation services. However, not all of these
services are directly related to forest-based recreation and tourism. The proportion of the
total number of arts, entertainment, recreation, accommodation and food service jobs in
the study forest that are directly attributable to forest-based recreation is calculated by
estimating the number of jobs attributable to the two most popular forest-based recreation
activities in the study forest: snowmobiling and deer hunting. The baseline number of
forest-based recreation and tourism jobs in the study forest was estimated using US.

Census employment data for retail, food, entertainment and accomodation service

86

sectors. Economic multipliers are used to estimate the number of jobs attributable to
each of the two important recreation-based industries in the area. See Appendix 2 for
details of these calculations.

3.4.2. Birds

The study forest serves as important habitat to migratory forest songbirds, and
this region has one of the highest levels of migratory forest songbird diversity in North
America because of the overlap that occurs in this part of the UP between northern and
southern migratory habitats (Howe et al. 1995). One of the variables chosen to represent
migratory forest songbirds is “percent of area with high migratory forest songbird
diversity.” This variable was defined in such a way as to allow us to connect the
economic variable later with the ecological variable in a larger integrated model.
Another variable relating to migratory forest songbirds is “number of species of
conservation concern at or above their target population level.” Because the wildlife
ecologist researchers will not have data on all of the individual species of conservation
concern in the study forest to be used in the integrated model, we chose to define the
variable in a more general way that will allow us to later connect this variable with the
ecological data in the larger integrated model.

The information treatment on migratory forest songbirds was developed with the
collaboration of the ecologists working on the project. The Partners in F light (PIF)
classification of “priority breeding species for conservation” for Region 12 (a region that
most closely approximates the study forest) was used to identify the migratory forest
songbirds in the study forest that are also birds of conservation concern (PIF 2005) (See
Appendix 2 for more details). The criteria used by PIF to determine priority for

conservation is based on numerous variables, including the following: global relative

87

abundance, global scores for breeding distribution and winter distribution, global score
for threats in the nonbreeding season, threats to successful breeding, importance of
Region 12 for breeding, population trend, priority tier, percent of species’ breeding
population in Region 12, and watch list status (PIF 2005).

The data from the PIF source was used to determine the number of migratory
songbirds of conservation concern currently in the study forest that are at or above their
target population level. This is the status quo level for the variable birdcons. The current
level of migratory forest songbird species diversity in the study area was taken from an
estimate provided by the wildlife ecologists collaborating on this project. This is the
status quo level for the variable birddiv.

3.4.3 Deer

There is evidence that deer browse is affecting the regeneration of certain species
of trees in the study forest and that this in turn is adversely affecting habitat for migratory
forest songbirds (Laurent et al. in prep, Shi et al. 2005). The concept of deer browse was
familiar to most focus group participants. The variable chosen to represent deer browse
is “percent of area with deer browse high enough to affect tree regeneration.” Deer
browse and its effect on tree regeneration is an important component of this research
project because of its complex ecological effects. We discovered through the focus
group discussions that the relationships between deer browse, forest structure and
composition, forest songbird habitat and timber harvesting activities were too
complicated to attempt to convey in this survey.

In order to maintain simplicity and to convey the issue of deer browse, only its
effect on tree regeneration was discussed in the survey. The status quo level of “deer

browse in the study area high enough to affect tree regeneration” is based on estimates

88

provided by forest ecologists collecting deer browse data in the study forest. The choice
table in the survey also showed respondents an attribute representing deer numbers. This
was intended to ensure that respondents did not confuse deer browse with deer numbers

in the area. The level of the deer number attribute did not vary across choices.

3.5 Survey Implementation

The survey was sent to a stratified random sample of Michigan households using
a modified version of Dillman’s tailored design method (Dillman 2000). The sample was
purchased from Survey Sampling International and was designed to represent four
geographic strata of Michigan households. Strata were divided to represent: 1)
households within the study area (defined by US. Census 2000 block groups), 2)
households within the Upper Peninsula but outside the study area (defined by US.
Census 2000 block groups and Michigan counties), 3) households within the counties of
the Northern Lower Peninsula and 4) households within the counties of the Southern
Lower Peninsula (See Appendix 1 for details on sample design).

Surveys were mailed to a sample of 2,000 Michigan households (See Appendix 5
for details on survey implementation). The survey was sent using four contacts: a hand-
signed, personalized prenotice letter, a first mailing of the questionnaire, a hand-signed
personalized reminder post card, and a second mailing of the questionnaire (See
Appendices 3 and 4 for copies of survey correspondence and the survey instrument).
Each questionnaire mailing included a hand-signed, personalized cover letter, a survey
booklet and a postage-paid business reply envelope. Three first class stamps were
included in the first questionnaire mailing of each group as a respondent incentive. The

use of three first class stamps as an incentive in the first wave mailing was assumed to

89

increase the response rate by as much as a third wave mailing would, and the costs of an
additional third wave mailing were not feasible within the project budget when using the
stamp incentive. Previous surveys have shown that using an incentive of three first-class
stamps can increase response rates by approximately 5% (Lupi 2005). Of the 2,000
surveys mailed, 1,899 were delivered to respondents. A total of 954 usable surveys were
returned, yielding an overall response rate of 50% (AAPOR 2004).

Summary statistics of the socioeconomic characteristics of respondents are
reported in full in Appendix 6. Most respondents (82%) do not own property in the study
area. Those that do (18%) own an average of 39 acres. A small percentage of
respondents (3%) are members of forestry organizations, while a slightly larger
percentage are members of an environmental organization (7%) and even more are
members of a hunting club (20%). Respondents on average were predominantly male
(77%), with a median age of 55. A minority (39%) have an associate's degree or higher.
A majority of respondents reported living in a rural area (71%). The respondent pool is
not very ethnically or religiously diverse with 95% white respondents and 75% protestant
or catholic respondents. Many respondents (74%) have participated in recreational
activities in or near the study area at some point in their life. This may indicate that those
who were already familiar with the study area were more likely to respond to the survey.

After the series of choice questions in the survey, each respondent was presented
with a table in which she could rank the importance of each forest easement program
attribute in her decision to vote yes or no for the programs. The fiequency of responses
for attribute importance rankings is reported in Table 3.3. Industry jobs, recreation and
tourism jobs, deer browse and deer numbers are ranked ‘important’ to ‘very important’

by a majority of respondents. Cost is considered to be important to very important by

90

about 78% of respondents. Forty-four to 45% of respondents ranked bird diversity and
protecting bird species of conservation concern as ‘important’ to ‘very important,’
respectively. Few respondents ranked the industry jobs, recreation and tourism jobs and
cost variables as unimportant (12%, 14% and 9%, respectively), while slightly more
respondents ranked the bird diversity and bird conservation variables as unimportant in
their decisions (24% and 25%, respectively). While all variables seem to rank highly for
respondents in influencing their choices, the economic variables (jobs and cost) rank the
highest in importance, followed by the deer browse and deer number attributes. The bird

diversity and bird conservation attributes rank lowest in importance overall.

91

Table 3.3. Response Frequencies for Respondent Evaluation of Attribute
Importance to Choice

 

Easement Program Attribute Percent Response N

 

VI* I SI NVI NI

 

Number of forest 1ndustry jobs in the area 22.9 38.2 27 8.2 3.7 399

 

Number of forest-based recreation and

. . . 18.5 38.5 28.8 10.8 3.4 890
tourism jobs in the area

 

Percent of area with high migratory forest

songbird species diversity 11.1 34.2 30.6 18.4 5.6 889

 

Number of migratory forest songbird
species of conservation concern at or above
their target population (out of 19 possible

species)

Percent of area with deer browse high
enough to affect tree regeneration

11.7 32.3 31.1 18.5 6.4 896

12.3 39.6 33.8 11.1 3.1 893

 

Number of deer in the area (more, same,

15.1 37.3 30.3 13.5 3.8 895
less)

 

Cost to your household in increased annual
taxes

* VI = Very Important; I = Important; S1 = Somewhat Important; N VI = Not Very
Important; N1 = Not Important At All

50.6 27.8 12.8 5.2 3.6 897

The response pattern to the different bid amounts offered in the dichotomous
choice questions can be seen in a graph of the bid distribution. The percent of
individuals who respond yes to each bid is presented in graphical form in Figure 3.1.
Responses in this graph are averaged over all responses to bid levels with the same bid
amount. The distribution of 'yes' responses to the three bids ($20, $90 and $400)
presented to survey respondents follows the expected pattern of a downward sloping

curve, which means that the percent of ‘yes’ responses decreases as the bid increases.

92

Figure 3.1. Graph of Bid Distribution

 

Bid Distribution (N = 940)

'40
35 -
30 -
25 -
20 -
15 -
10 -

 

Percent 'Yes'

 

 

 

$20 $90 $400
Bid Level

 

 

 

3.6 Model Estimation Results

The standard probit and the random effects probit models were estimated using
the forest easement program attributes as explanatory variables in order to test for the
presence of random effects. Rho, the correlation coefficient between the overall model
variance and the variance within an individual’s responses, is significantly different from
zero at the 99% level or higher, therefore it can be concluded that the standard probit
model results in biased standard errors of the estimated coefficients, and that it is
necessary to use the random effects model (Boxall et al. 2003). The model results
reported in Table 3.4 are therefore based on random effects probit model estimation.

Two models are estimated and results are reported in Table 3.4. Model 1 includes

93

only the forest easement program attributes as explanatory variables, and Model 2
includes easement program attributes as well as a set of socioeconomic characteristics of
the respondents. Many socioeconomic characteristics were originally included in the
model, but several were dropped due to a lack of explanatory power. Regardless of
different combinations of socioeconomic variables in the model, program attribute
coefficients and standard errors remained consistent. The variables that were dropped
include income, resource dependence (measured by employment of any family member
in a natural resource based industry), ethnicity, religion and gender. Income was not
found to have a significant effect on WTP in the analysis reported here‘. While income is
typically expected to have a positive and statistically significant relationship with WTP,
income elasticity of WTP is often found to be less than one in contingent valuation
studies (Carson et al. 2001, Hanemann 1994).

Expected signs of the estimated coefficients are as follows. An additional forest
industry job or forest-based recreation and tourism job should increase the probability of
an individual voting for the program. Therefore the signs on indjobs and rtiobs are
expected to be positive. The signs on birddiv and birdcons are also expected to be
positive, as an increase in bird diversity or an increase in the number of songbird species
of conservation concern should increase the probability of a ‘yes’ response because
increases in these variable represent environmental improvements. The coefficient on
deer is expected to be negative because an increase in the percent of area with high deer

browse is an effect that should decrease an individual's probability of voting ‘yes’ for the

 

' In this study, 11 individuals in the sample reported an annual household income of $200,000 or greater,
and income was significant in versions of the model where these individuals were dropped from the
analysis.

94

program because an increase in this variable represents a negative environmental effect.
Similarly, the coefficient on cost is expected to be negative, because as the price of the
program increases, the probability of a ‘yes’ response should decrease.

It is expected that recreational use, rec, will have a positive effect on willingness
to pay. Studies of environmental attitudes conclude that attitudes can be predictors of
behavior (McFarlane and Boxall 2000). Several studies that identify the effects of
socioeconomic characteristics on environmental attitudes conclude that higher levels of
education, urban residence and membership in environmental organizations are
connected to pro-environmental attitudes (Dietz et al. 1998). Studies in the
environmental attitude literature have also shown that individuals in older age groups and
politically conservative individuals tend to have attitudes that do not favor the
environment (Dietz et a1. 1998). Based on these studies and based on the assumption that
attitudes can be predictors of behavior, it is hypothesized that education, educ, urban
place of residence, urban, and environmental organization membership, envorg, will
have a positive effect on the probability of voting ‘yes’ for environmental improvements
and are therefore expected to have positive signs. Age, age, and conservative political
views, polview, are expected to decrease the probability that a respondent will vote ‘yes’
for a conservation program and therefore are expected to take on negative signs.

The estimated coefficients in both models all have the expected signs, and all
variables are significant at the 95% level or higher. All easement program attributes are
highly significant (99%) in both models with the exception of birdcons, which decreased
slightly in significance in Model 2, but is still significant at the 95% level in Model 2. Of
the socioeconomic variables, polview and urban are significant at the 95% level, and the

remaining variables, age, huntclub, envorg, educ and rec are all significant at the 99%

95

level.

Estimation results indicate that an increase in the number of forest industry jobs,
the number of forest-based recreation and tourism jobs, bird diversity and habitat for
songbirds of conservation concern increases the probability of an individual voting ‘yes’
for the easement program. An increase in the area affected by deer browse reduces the
probability of voting ‘yes’ for the program, as does an increase in the cost of the program
to a household. Older individuals, members of hunting clubs, and politically
conservative individuals have lower probabilities of voting ‘yes’ for the forest easement
program, while members of environmental organizations, individuals with higher levels
of education, people who live in an urban area, and people who have used the study area

for recreational purposes have a higher probability of voting ‘yes’ for the program.

96

Table 3.4. Estimation Results from Random Effects Models

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Model 1 Model 2
Intercept -0.9389"' "' '1‘ -0.243 8
(0.1281) (0.3556)
indjobs 0.0076“ " “ 0.0081 "' "' ‘
(0.0008) (0.0009)
rtjobs 0.0069"" 0.0071‘"
(0.0012) (0.0012)
birddiv 0.0119‘” 0.0130”"'
(0.0026) (0.0027)
birdcons 0.0240‘” 0.0235”
(0.0095) (0.0098)
deer -0.0151”‘ -0.0177""
(0.0054) (0.0056)
cost -0.0075‘ " "‘ -0.0072" " ‘
(0.0005) (0.0005)
age -0.0211""'
(0.005 I)
huntclub -0.5032‘l "' ‘
(0.1899)
envorg 0.9750" "
(0.2829)
educ 0.5226‘“
(0.1696)
polview -0.3553“I
(0. 1590)
urban 0.5322‘ "'
(0.241 8)
rec 0.4868‘ "' "
(0. 1980)
Rho 0.7902‘" 0.7572‘"
(0.0204) (0.0235)
Number of obs 3646 3264
Number of groups 940 841
Log Likelihood 4433.15 -1272.48
Prob > chi-square <0.0000 <0.0000

 

Standard errors in parentheses; "*Significant at the 99% level; ** Significant at the 95 % level

97

The chi square test of the hypothesis that all estimated model parameters are
jointly equal to zero can be rejected at the 99% level for both models. Another indicator
of the overall performance of a model is the likelihood value at convergence. Higher
values of log likelihood at convergence indicate improved performance of a model, and
likelihood ratio tests compare two different models to see if they are statistically different

(Boxall and Macnab 2000). Because likelihood tests need to be conducted on models
that use the same number of observations, another version of Model 1, Model 1a, was
estimated excluding all missing observations from the socioeconomic variables. The
hypothesis that the parameters on the socioeconomic variables in Model 2 are equal to
zero is tested using the likelihood ratio statistic. The likelihood value of Model 1a (the
restricted model) estimated using the N=841 data set is -1 303 .31 and the value for Model
2 (the unrestricted model) using the same number of observations is -1272.48. The larger
likelihood value of Model 2 indicates a better performance of that model. The likelihood

ratio statistic is calculated as follows:

i=2<LUR —LR) ~ 12

This statistic follows the chi square distribution with the degrees of freedom given by the

number of restrictions imposed. The test statistic for comparing Models la and 2 is:

2(— 1272.48 — (— 130331)): 61.66

The p-value of obtaining a chi square value higher than 61.66 with 7 degrees of freedom
is less than 0.005. Therefore, the null hypothesis that the coefficients on the 7
socioeconomic variables are equal to zero can be rejected, and the variables can be
retained in the model.

Overall, both models perform very well. Results of tests of significance of

98

parameter estimates and a likelihood ratio test indicate that Model 2 performs better than
Model la overall. All of the socioeconomic variables included in Model 2 are significant
and indicate that, in addition to program characteristics, demographic factors also have an

important influence on respondent choices.

3.7 Welfare Analysis and Discussion

The estimation results from the previous section can be used to estimate
willingness to pay for program attributes. These WTP values are referred to as implicit
prices and are calculated using equation 22, specified in Section 3.3. Their calculation is
based on the ceteris paribus assumption that all other variables are held constant (Bennett
et al. 2001). Implicit prices can provide useful information to policy makers on the
benefits of small changes in individual aspects of environmental quality (Morrison et al.
2002). The implicit prices of each attribute of the easement program from both models
are reported in Table 3.5. Implicit prices for each attribute are similar across both
models, but the values are slightly higher for all attributes in Model 2, which was
determined to be a better performing model in the previous section.

All attributes have a positive implicit price with the exception of deer, which has
a negative value for marginal WTP. This means that individuals need to be compensated
for an additional unit of deer browse damage. The amount of compensation for deer
browse damage is greater in Model 2 than Model 1. The marginal implicit prices are
slightly higher in Model 2 for indjobs, rtjobs, birddiv and birdcons. The value of an
additional forest industry job is slightly higher than that of a forest-based recreation or
tourism job in both models. People are willing to pay more for an additional songbird

species of conservation concern at or above its target population than for they are willing

99

to pay for an additional percent of bird diversity or for an additional forest industry or
forest-based recreation and tourism job. These results provide an indication of the
relative values individuals place on the different forest ecosystem attributes specified in

the forest easement program.

Table 3.5. Implicit Prices of Attributes

 

 

 

 

 

 

Variable Model 1 Model 2
indjobs $1.01 $1.13
rtjobs $0.92 $0.99
birddiv $1.59 $1.81
birdcons $3.20 $3.26
deer -$2.01 -$2.46

 

Implicit prices are marginal rates of substitution between a nonmonetary attribute
and a monetary attribute. Other marginal rates of substitution can also be calculated
between nonmonetary attributes. The marginal rate of substitution of one attribute for
another can be calculated and used to examine the trade-offs that individuals are willing
to make among program attributes, or the relative importance they place on attributes.
Because Model 2 was shown to be a better performing model than Model 1, coefficients
from Model 2 are used to calculate the trade-offs between other easement program

attributes in Table 3.6.

100

Table 3.6. Marginal Rates of Substitution Between Nonmonetary Program
Attributes

 

For a 1 unit increase in:

 

 

 

 

 

 

inab’obs rtjobs birddiv birdcons deer
Willing to give up:
indjobs 0.87 1.6 2.9 -2. l 8
rtjobs 1.14 1.83 3.31 -2.49
birddiv 0.62 0.55 1.81 -1.36
birdcons 0.34 0.30 0.55 -0.75
deer -0.46 -0.40 -0.73 -1.33

 

Individuals are willing to trade slightly more recreation and tourism jobs for an
increase in forest industry jobs than they are willing to trade the converse. Individuals
are willing to give up approximately 1 recreation and tourism job for an additional forest
industry job but will give up less than one industry job for an additional recreation and
tourism job. This suggests that forest industry jobs rank higher in importance than
recreation and tourism jobs. For additional units of bird diversity and bird species of
conservation concern, individuals are willing to give up more recreation and tourism jobs
than industry jobs. Increases in species of conservation concern are worth relatively
more in jobs than are increases in bird diversity. The largest trade-offs in nonmonetary
attributes are made for increases in migratory forest songbird species of conservation
concern that are at or above their target population.

The trade-offs made between the deer variable and other attributes require a
slightly different interpretation because an increase in the deer attribute represents a

negative effect of increased browse levels that affect tree regeneration. For increases in

101

industry jobs, recreation and tourism jobs and bird diversity, individuals are willing to
give up a less than one percent reduction in area that is heavily browsed. For an increase
in bird species of conservation concern at or above the target habitat level, individuals are
willing to give up slightly more than a one percent reduction in heavily browsed areas. If
the area heavily browsed by deer increases by one percent, people are willing to accept
an increase of about 2 industry jobs, 2 1/2 recreation and tourism jobs, slightly more than
1 percent increase in bird diversity, and an increase of less than one songbird species of

conservation concern.

3.8 Conclusions

This research uses data collected from a mail survey of 2,000 Michigan residents
to estimate the nonmarket values of forest ecosystem attributes of an area of Michigan’s
Upper Peninsula. The analysis employs the contingent valuation method, based on
random utility theory, within an attribute-based referendum format. Attributes for which
nonmarket values are estimated in this study reflect the ecological and social importance
of the study forest.

The study forest is a region that is economically dependent on forests and where
forests form a dominant part of the ecological landscape. Forests in the study area are
extremely important to the economic sustainability of the region. In addition, forest
management practices in this region have had adverse effects on wildlife habitat and
forest regeneration. This study estimates nonmarket values for forest industry jobs,
forest-based recreation and tourism jobs, forest migratory songbird species diversity,
number of forest migratory songbird species of conservation concern and the effects of

deer browse on tree regeneration. The results of this research show that ecological and

102

social attributes of forests are valued by individuals in Michigan.

Results of this research can be useful in policy and resource management
decisions that require a comparison of the relative benefits of specific environmental
attributes of the study forest. These results indicate that support for protection of
ecological and social resources in this area of the Upper Peninsula exists in Michigan.
Results also suggest that the development of programs aimed at forest and wildlife
conservation in the area need to include options for sustaining the social value of forest

jobs in the area.

103

REFERENCES

AAPOR. 2004. Standard Definitions: Final Dispositions of Case Codes and Outcome
Rates for Surveys. Edited by American Association for Public Opinion Research.
3rd ed. Lenexa, KS.

Adamowicz, V., and P. Boxall. 2001. Future directions of stated choice methods for
environment valuation. Paper read at Choice Experiments: A New Approach to
Environmental Valuation, April 10, 2001, at London, England.

Adamowicz, W.L., P. Boxall, M. Williams and J. Louviere. 1998. Stated Preference
approaches to measuring passive use values: choice experiments versus
contingent valuation. American Journal of Agricultural Economics 80 (1):64-75.

Addelman, S., and O. Kempthome. 1961 . Orthogonal Main-Effect Plans. Ames, IA: Iowa
State University, Aeronautical Research Laboratory, Office of Aerospace
Research, United States Air Force.

Bennett, J ., and R. Blarney. 2001. Introduction. In The Choice Modelling Approach to
Environmental Valuation, edited by J. Bennett and R. Blamey. Cheltenham, UK:
Edward Elgar.

Bennett, J ., and V. Adamowicz. 2001. Some Fundamental of Environmental Choice
Modelling. In The Choice Modelling Approach to Environmental Valuation,
edited by J. Bennett and R. Blamey. Cheltenham, UK: Edward Elgar.

Bennett, J ., J. Rolfe, and M. Morrison. 2001. Remnant vegetation and wetlands
protection: Non-market valuation. In The Choice Modelling Approach to
Environmental Valuation, edited by J. Bennett and R. Blamey. Cheltenham, UK:
Edward Elgar.

Bishop, R., and T. Heberlein. 1992. The Contingent Valuation Method. In Natural
Resource Damages: Law and Economics, edited by K. W. a. J. Duffield: John
Wiley and Sons.

Blarney, R., J. Rolfe, J. Bennett, and M. Morrison. 2000. Valuing remnant vegetation in
Central Queensland using choice modelling. The Australian Journal of
Agricultural and Resource Economics 44 (3):439-456.

Boxall, P., K. Rollins, and J. Englin. 2003. Heterogeneous preferences for congestion
during a wilderness experience. Resource and Energy Economics 25:177-195.

104

Boxall, P.C., and B. Macnab. 2000. Exploring the preferences of wildlife recreationists
for features of boreal forest management: a choice experiment approach.
Canadian Journal of Forest Research 30: 193 1-1941 .

Boxall, P.C., W.L. Adamowicz, J. Swait, M. Williams and J. Louviere. 1996. A
comparison of stated preference methods for environmental valuation. Ecological
Economics 18:243-253.

Carson, R.T., N.E. Flores and NF. Meade. 2001. Contingent valuation: Controversies
and Evidence. Environmental and Resource Economics 19:173-210.

Carson, R.T. 2000. Contingent Valuation: A User's Guide. Environmental Science and
Technology 34:1413-1418.

Dietz, T., P.C. Stern, and G.A. Guagnano. 1998. Social structural and social
psychological bases of environmental concern. Environment and Behavior
30:450-471.

Dillman, DA. 2000. Mail and Internet Surveys: The Tailored Design Method, 2nd
edition. 2nd ed. New York: John Wiley and Sons, Inc.

Haab, T.C., and K.E. McConnell. 2002. Valuing Environmental and Natural Resources.
Cheltenham, UK: Edward Elgar.

Hanemann, W.M. 1994. Valuing the environment through contingent valuation. The
Journal of Economic Perspectives 8 (4): 19-43.

Hanley, N. and Ruffel, R.J. 1993. The contingent valuation of forest characteristics: two
experiments. Journal of Agricultural Economics 44 (2):218-229.

Hanley, N., R.E. Wright & V. Adamowicz. 1998. Using choice experiments to value the
environment. Environmental and Resource Economics 11 (3-4):413-428.

Holmes, T.P., and K.J. Boyle. 2005. Dynamic learning and context-dependence in
sequential, attribute-based, stated-preference valuation questions. Land
Economics 81 (1):] 14-126.

Holmes, T.P., and W.L. Adamowicz. 2003. Attribute-based methods. In A Primer on
Nonmarket Valuation, edited by P.A. Champ K.J. Boyle and TC. Brown.
Dordrecht: Kluwer Academic Publishers.

Howe, R.W., Niemi, G., and J .R. Probst. 1995. Management of Western Great Lakes
Forests for the Conservation of Neotropical Migratory Birds. In Management of
Midwestern Landscapes for the Conservation of Neotropical Migratory Birds,
edited by 1. FR. Thompson. Detroit: USDA Forest Service North Central Forest

Experiment Station.

105

Kaplowitz, M.D., and J .P. Hoehn. 2001. Do focus groups and individual interviews
reveal the same information for natural resource valuation? Ecological Economics
36:237-247.

Kaplowitz, M.D., F. Lupi, and J.P. Hoehn. 2004. Multiple-Methods for Developing and
Evaluating A Stated Choice Survey for Valuing Wetland Ecosystems. In
Questionnaire Development, Evaluation, and Testing Methods, edited by S.
Presser et al. New York: John Wiley & Sons, Inc.

Laurent, E. J ., J .B. LeBouton, H. Shi, M. Walters and J. Liu. In prep. Independent effects
of managed northern hardwood stand characteristics on bird species' occurrences
in Michigan's Upper Peninsula.

Lind, B. 2001a. Trends in working forest conservation easements - A report from the
April 2001 meeting of the Land Trust Alliance Working Forest Conservation
Easements Advisory Panel. Washington, DC: Land Trust Alliance.

. 2001b. Using conservation easements to protect working forests. Exchange
Spring 2001:10-13.

 

Lockwood, M., J. Loomis, and T. De Lacy. 1994. The relative unimportance of a
nonmarket willingness to pay for timber harvesting. Ecological Economics 9: 145-
152.

Loomis, J., M. Lockwood, and T. DeLacy. 1993. Some Empirical Evidence on
Embedding Effects in Contingent Valuation of Forest Protection. Journal of
Environmental Economics and Management 24:45-55.

Louviere, J .J . 2001. Choice experiments: An overview of concepts and issues. In The
Choice Modelling Approach to Environmental Valuation, edited by J. B. a. R.
Blarney. Cheltenham, UK: Edward Elgar.

Louviere, J .J ., D.A. Hensher, and J .D. Swait. 2000. Stated Choice Methods: Analysis and
Application. Cambridge: Cambridge University Press.

Lupi, F. 2005. Personal Communication.

Lupi, F., MD. Kaplowitz, and J .P. Hoehn. 2002. The Economic Equivalency of Drained
and Restored Wetlands in Michigan. American Journal of Agricultural
Economics 84 (5):]355-1361.

McFadden, D. 1974. Conditional logit analysis of qualitative choice behavior. In
Frontiers in Econometrics, edited by P. Zarembka. New York: Academic Press.

McFarlane, B.L., and P.C. Boxall. 2000. Factors Influencing Forest Values and Attitudes
of Two Stakeholder Groups: The Case of the Foothills Model Forest, Alberta,
Canada. Society and Natural Resources 13:649-661.

106

Mitchell, RC. 2002. On designing constructed markets in valuation surveys.
Environmental and Resource Economics 22:297-321.

Mitchell, R.C., and R.T. Carson. 1989. Using Surveys to Value Public Goods: The
Contingent Valuation Method. Washington, DC: Resources for the Future.

Morrison, M., J. Bennett, and R. Blarney. 1999. Valuing improved wetland quality using
choice modeling. Water Resources Research 35 (9):2805-2814.

Morrison, M., J. Bennett, R. Blarney, and J. Louviere. 2002. Choice modeling and tests
of benefit transfer. American Journal of Agricultural Economics 84 (1):161-170.

Newsom, D. 2002. Forest certification and working forest conservation easements:
Common elements and first thoughts on a combined system: Rainforest Alliance
Sustainable Forestry Division.

Partners in Flight (PIF). 2005. Partners in Flight Bird Conservation Plan for the Boreal
Hardwood Transition (Physiographic area 20), Draft: American Bird
Conservancy.

Portney, FR. 1994. The contingent valuation debate: Why economists should care. The
Journal of Economic Perspectives 8 (4):3-17.

Potter-Witter, K. 2005. Personal Communication.

Shi, H., E.J. Laurent, J. LeBouton, L. Racevskis, K.R. Hall, M. Donovan, R.V. Doepker,
M.B. Walters, F. Lupi and J. Liu. 2005. Local spatial modeling of white-tailed
deer distribution. Ecological Modeling.

Stevens, T.H., R. Belkner, D. Dennis, D. Kittredge, and C. Willis. 2000. Comparison of
contingent valuation and conjoint analysis in ecosystem management. Ecological
Economics 32:63-74.

vanRensburg, T.M., G.A. Mill, M. Common, and J. Lovett. 2002. Preferences and
multiple use forest management. Ecological Economics 43 :23 1 -244.

Ward, J ., and J. Ervin. 2005. Land trusts and certification: Complementary tools for
forest conservation 2005 [cited September 17 2005]. Available from
http://www.greendesign.net/understory/sum99/LandTrusts.htrnl.

Wooldridge, J .M. 2002. Econometric Analysis of Cross Section and Panel Data.
Cambridge, MA: MIT Press.

Xu, W., B.R. Lippke, and J. Perez-Garcia. 2003. Valuing biodiversity, aesthetics, and job

losses associated with ecosystem management using stated preferences. Forest
Science 49 (2):247-257.

107

Essay 4:
Effects of Distance, Attitudes and Recreational
Use on Willingness to Pay for Forest Ecosystem Services

4.1 Introduction

Environmental value estimates that result from non-market valuation studies are
key inputs to cost-benefit analysis in support of policy decisions (Hanley et al. 2003). In
order to improve the usefulness of estimates for cost-benefit analysis and policy making,
it is important to determine whose values should be included in the analysis.
Determining the extent of the market for environmental goods and services can be
accomplished by estimating effects of distance, resource management attitudes and
recreational use on willingness to pay for the provision of an environmental good or
service. This study investigates the effects of these factors on willingness to pay (WTP)
for a forest easement program in an area of Michigan’s Upper Peninsula (UP).

Information on the relationship between distance and WTP can provide a
conceptualization of who benefits from environmental improvements (Hanley et a1 2003).
Several studies have analyzed the effect of distance on non-market values of natural
resources and have found a negative relationship between distance and WTP (Johnson et
al. 2001, Loomis 1996, Bateman and Langford 1997, Pate and Loomis 1997, Sutherland
and Walsh 1985), also referred to as the distance decay effect (Hanley et al. 2003). This
study estimates the effect of distance on WTP for a forest easement program in a study
forest in Michigan’s UP.

Attitudes towards resource management may also play an important role in
predicting the probability that individuals will support an environmental improvement.

The connections between environmental values, attitudes and behavior have been well

108

established in the environmental behavior literature (Dietz et al. 1998, Nordlund and
Garvill 2002, Johnson et al. 2004, Tarrant and Cordell 1997, Stem et a1. 1995, Poortinga
et al. 2004). However, relatively few studies in the nonmarket valuation literature have
examined the effects of attitudes on the nonmarket values of environmental goods and
services. Contingent valuation (CV) studies that have included attitudes in the analysis
of WTP values report a positive relationship between pro-environmental attitudes and
WTP (Bandara and Tisdell 2003, Streever et al. 1998, Stevens et al. 1991). Attitudinal
variables can improve the performance of CV models, and information on the effects of
attitudes on WTP forms an important component of nonmarket valuation that can help
understand the motivations that underlie individuals’ choices (McLelland 2001).

In this study, attitudinal data collected from a mail survey of Michigan residents
is included in a CV model to test the hypothesis that resource management attitudes can
be used to explain WTP. It is assumed that there are certain types of environmental
values that underlie individuals’ attitudes towards resource management. The
conceptualization of environmental values and attitudes in this paper is based on the
hierarchical framework established by Stern et al. (1995), in which values underlie
attitudes and attitudes predict behavior.

Consideration of use and non-use values also contributes to determining the
extent of the market for an environmental good. When the total value an individual holds
for a resource is greater than that individual’s use value, the remaining value is referred
to as non-use value (Freeman 2003). Benefit estimates taken only at the site of the
resource may result in exclusion of use and non-use values of individuals who do not live
near the resource, and if the region from which values are elicited is too broad, estimates

may be inflated (Pate and Loomis 1997, Sutherland and Walsh 1985). Results from

109

Essay 3 showed that recreational use has a significant positive effect on WTP. This study
uses data on specific types of recreational uses of survey respondents to determine how

the effects of recreational use on WTP change as distance from the resource increases.

4.2 Methods
4.2.1 Survey design and implementation

The survey instrument was designed based on results from six focus groups, 21
individual pro-test interviews and interviews with professional ecologists, forests and
state agency employees. The survey collected stated preference data using a
dichotomous choice referendum format and also collected data on attitudes towards forest
management in the study area. The questionnaire utilizes a forest easement program as
the policy context for the contingent market. The easement program is described using a
set of six program attributes that take on any of these levels (See Table 4.1). The
combinations of attributes and their levels are established according to an orthogonal
main-effects experimental design. The experimental design of the six attributes is based
on a main effects design plan for a 3‘5 design, which produced 18 choice sets (Addelman

and Kempthome 1961).

110

Table 4.1. Survey Attributes and Levels

 

 

 

 

 

 

Variable Attribute Name Status Quo Attribute

Name Level Levels

indjobs Number of forest mdustry jobs 1n the 67 5 600, 67 5’ 710
area

rtjobs Forest-based recreation and tourism 190 170’ 190, 250
jobs in the area

birddiv Percent of area wrthhrgh. m1gratory 35% 38% , 55%, 75%
forest songbird specres diversrty

birdcons Number of migratory songbird species
of conservation concern that are at or 6 7 12 17
above their target population (out of ’ ’
19 possible species)

deer Percent of area with deer browse high

0 0
enough to affect tree regeneration 69% 67 A" 58 A" 49%

 

cost Cost to your household 1n Increased $0 $20, $90, 5400
annual taxes

The survey was mailed to a stratified random sample of 2,000 Michigan
households using a modified version of Dillman’s tailored design method (Dillman
2000). Of the 2,000 surveys mailed, 954 usable surveys were returned, yielding an
overall response rate of 50% (AAPOR 2004). Details of survey design and

implementation are reported in Essay 3 and Appendices 2 and 5.

4.2.2 Theoretical model

The analysis in this essay utilizes an attribute-based referenda model (ABR),
which is a nonmarket valuation method that is a hybrid of contingent valuation and
attribute-based stated preference nonmarket valuation methods (Holmes and Boyle

2005). The analysis in this study tests the hypothesis that distance from the resource has a

111

negative effect on WTP for forest ecosystem services in an area of Michigan’s Upper
Peninsula (UP). The analysis uses data collected from a stated preference mail survey of
Michigan residents. The study forest, which forms the focus of the survey, was chosen
for the importance of its forests to deer habitat, forest migratory songbird habitat as well
as to the sustainability of the local economy.

ABR models, like contingent valuation and attribute-based methods, are based in
random utility theory (Holmes and Boyle 2005, McFadden 1974). Within the random
utility theoretical framework, utility is assumed to be composed of a deterministic
component and a random component. Indirect utility, u, is the maximum amount of
utility that a household can derive from income, y, given prices of goods, a vector of
environmental quality variables, x, other respondent characteristics, 2, and a component

of individual preferences, 8, known to the individual but not to the researcher,

u = “(EARL-‘3). (1)

In an ABR model, respondents are asked if they are willing to pay a certain
amount to achieve an environmental quality improvement. In this model, the quality
improvement is described by changes in the levels of attributes of a forested ecosystem
that will be provided by a program at a cost to the respondent. Utility to the individual

when an amount p is paid is:

u1=u(xlrzry—prsl)' (2)

In this equation, u 1 represents the indirect utility function for an individual who
pays the cost of the program; x I is a vector of forest ecosystem attributes under the forest
protection program. If the cost, p, of the program is not paid, the indirect utility fimction

is written as follows:

112

uo =u(x0,z,y,ao). (3)

In this equation, “0 represents indirect utility under the status quo, and x0 is the vector of
forest attribute levels without the program. An individual will be willing to pay for the
proposed program if:

“1(xlaz9y-p’81)2u0(x0szsys80)' (4)

The probability that a respondent is willing to pay for the forest protection program
(probability of saying yes) is given by the probability that the utility received from the

forest protection program is greater than the utility received under the status quo:

Pr(yeS) = Pr[u1(x1,Z,.V-P,31) > “0(x02z9y980)]
= p.11. > 0]. ‘5’

The indirect utility function has an unobservable, random component. Indirect utility of
individual i from alternative j, therefore, can be expressed as the sum of its explainable

and unexplainable components:

“:7'="0°+80" (6)

where vi} is the explainable component of utility to individual i from alternative j, and e is
the unexplainable, random component of utility for individual i from alternative j.

The deterministic component of utility is defined as:

vy=aj+7]zl+B(yt-pj) a Yj=7mvjsm¢og (7)

where i indexes individuals, j indexes alternatives, v is indirect utility, x] is a set of

program attributes, z, is a set of respondent characteristics, y is income, p is the cost of

113

the program and a, y and [3 are estimable parameters. An individual will vote ‘yes’ to the
program if utility with the program exceeds utility without the program. Because utility
is composed of a deterministic and a random component, the following expression

represents the probability that an individual will vote for the program:

Pr(yes) = Pr[v,j + ’31} > v,0 + 8,0] , (8)

which, when substituting (7) for indirect utility, yields

Pr(yes): Pr[a(ij)+y z, —[3pj > am - 80.]. (9)

Assuming that the error terms follow a standard normal distribution, the probit model can
be used to estimate equation 9.

An assumption of the standard probit model is that the error component is
independent and identically distributed among individuals and across observations for
each individual. However, when an individual responds to more than one stated
preference question, it is likely that there are unobservable characteristics specific to that
individual that induce correlation across her responses. If this is suspected to be the case,
it is appropriate to estimate a random effects probit model (Wooldridge 2002). In a
random effects model, the error term is treated as separable into two components: one
that is unobservable and specific to each individual and another that is unobservable and
due to random response shocks across all individuals and all responses (Boxall et al.
2003).

The utility difference function is specified using a random effects utility model

and is written as follows:

Au.=a(Ax,)+vz.-Bp,+u.+e.. (10)

114

where 11,- is the individual-specific error term, and 81] is the random disturbance term

across all individuals and observations.

4.2.3 Model Specification

To estimate the effects of distance, attitudes and recreational use on WTP,
Equations 9 and 10 are estimated using a series of random effects probit models.
Distance from the resource, attitudes and recreational use variables are included in the
model as respondent characteristics, 2,. Results of these models are presented in Section

4.3. The utility difference function is specified as follows:

Aug. = a(Ax,)+y z, -Bp,-
= 010 + a1Ainay'obs j + rtzArtiobsj + 013Abirddivj. (1 1)

+ (1,,Abirdconsj + asAdeerj + Y 11 + BICOStI ’

(Note: B, = -B)
where a is a vector of estimable parameters for each of the k program attributes, x, of
alternative j, y is a vector of estimable parameters for the effect of respondent

characteristics, 2,, and B is an estimable parameter for the program cost.

4.3 Model estimation
4.3.1 Distance model

The hypothesis that distance has a negative effect on the probability that an
individual will be willing to pay for the forest easement program is tested by estimating a

random effects probit model using binary choice data of voting ‘yes’ or ‘no’ to a forest

115

easement program as the dependent variable and program attributes and distance as
explanatory variables. The variables used in the distance model include program
atttributes, defined in Table 4.1, and the variable distance, which represents the distance
in miles from the respondent’s zip code to the zip code of the center of the study area.

Results of this model are reported in Table 4.2.

Table 4.2. Distance Model Estimation Results

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Model 1
Constant -1.1674*"
(0.1614)
inay'obs 0.00761!"
(0.0009)
rtjobs 0.0069‘"
(0.0012)
birddiv 0.0120""'
(0.0026)
birdcons 0.0242"
(0.0095)
deer -0.0158"’"
(0.0054)
cost -0.0074"*
(0.0005)
distance 0.0010"
(0.0004)
Rho 0.7919"*
(0.0210)
N 3618
# of groups 933
Log likelihood -1422.38
Prob>chi2 <0.000

 

116

The program attributes in this model all significantly affect WTP. An increase in
both forest industry jobs and forest-based recreation and tourism jobs increases the
probability of voting for the program. Likewise, increases in bird diversity and the
number of birds of conservation concern at or above their target population increase the
probability of voting for the program. As expected, an increase in the cost of the
program decreases the probability that an individual would vote for the easement
program. An increase in the level of deer browse high enough to affect tree regeneration
also decreases WTP. Results reported in Table 4.2 show that distance has a positive and
significant effect on the probability of voting ‘yes’ for the easement program. This
unexpected result leads to the rejection of the distance-decay hypothesis. However, there
may be other factors that explain this positive distance effect, and additional hypotheses

are tested to examine why distance has a positive effect in this model.

4.3.2 Distance and attitudes model

In addition to the referendum choice data, the survey instrument also collected
data on respondent attitudes towards forest management in the study area. Individuals
were presented with descriptions of the importance of the study area for migratory forest
songbird habitat as well as its importance for the provision of forest industry and forest-
based recreation and tourism jobs. Respondents were asked to respond to a series of
statements that reflect attitudes about the goals of forest management in the study area.
Analysis of these attitude statements in Essay 2 concluded that respondent attitudes can
be conceptualized using an anthropocentric/biocentric value scale. The anthropocentric-
biocentric continuum conceptualizes different types of values individuals hold that relate

to natural resources and the environment. Anthropocentric values reflect support of the

117

instrumental value of natural resources and the environment for the benefit of humans,
whereas biocentric values reflect support of the intrinsic value and ecological and life
support roles of natural resources (Steel et al. 1994, Tarrant and Cordell 2002). This
essay tests the hypothesis that differences between individuals with anthropocentric
versus biocentric forest management attitudes explain the probability of voting ‘yes’ for
the easement program.

This hypothesis is first examined by using descriptive statistics to examine how
attitudes differ between recreational users at various distances from the study area.
Survey respondents are divided into 4 different distance zone groups based on the
number of miles from the study area within each distance quartile. The following table

defines the different distance zones.

Table 4.3. Definition of Distance Zones

 

 

 

 

 

Distance Zone Miles from study area
1 0 to 37
2 38 to 117
3 118 to 368
4 369 to 503

 

Seventy-eight percent of survey respondents reported some type of participation
in recreational activities in or near the study area. The survey collected data on whether
the respondent had ever participated in any recreational activities in or near the study
area but did not collect data on the frequency of recreational use. Typically it is assumed

that recreational use is inversely related to distance (Sutherland and Walsh 1985).

118

Although Table 4.4 supports this assumption, users drop from 91% in the first quartile of
distance from the study area to 62% in the furthest quartile of distance. However, there is
still a relatively high proportion of respondents who are recreational users of the study
area living relatively far away from the study area.

Table 4.4. Proportions of Recreational Users and Non-Users Within Different
Distance Zones

 

 

 

Respondent Zone 1 Zone 2 Zone 3 Zone 4 All zones
Type (0 to 37 (38 to 117 (118 to 368 (369 to 503
miles) miles) miles) miles)
N % N % N % N % N %
Users 216 91 200 84 173 74 I46 62 735 78

 

Non-users 22 9 38 16 6O 26 90 38 210 22
Total 238 100 238 100 233 100 236 100 945 100

A comparison of attitudes between recreational users in the different distance
zones reveals that there are significant differences in biocentric/anthropocentric attitudes
between them, and that there tend to be more recreational users with biocentric attitudes
who live farther from the resource. To simplify the reporting of results, distance zones
were condensed from 4 zones into 2 zones in order to investigate how recreational users’
attitudes differ at different distances from the resource. Table 4.5 shows cross-tabulation
results of forest management attitudes for which significant differences exist between
Zone 1 and 2 users below the median distance and Zone 3 and 4 users above the median

distance.

119

Table 4.5. Summary of Recreational Users' Forest Management Attitudes
Compared Between Distance Zones

 

Attitudes Zones 1&2 or
Zones 3&4

recreational users

Tend to agree that study area forests should be managed to Zones 1&2"
meet human needs

 

Tend to agree that forests should be managed to meet the needs Zones 1&2“
of communities economically dependent on forests regardless
of environmental consequences

 

 

 

 

Tend to agree that study area forests should be managed to Zones 1&2*"
protect forest industry jobs

Tend to be aware of the importance of the study area for Zones 1&2"
songbird habitat

Tend to be concerned about forest songbird diversity in the ‘ Zones 3&4“
study area

Tend to agree that forest management in the study area should Zones 3&4"

increase songbird diversity even if it means losses to forest-
based industries

 

Tend to feel that it is important to protect habitat for forest Zones 3&4"*
songbirds of conservation concern in the study area

 

Tend to agree that forest management in the study area should Zones 3&4"*
protect habitat for songbirds of conservation concern even if it
means economic losses to forest-based industries

Pearson chi squared statistic * = p < 0.10; " = p < 0. 05; *** = p < 0.001

The results reported in Table 4.5 show that users farther from the resource tend to
hold more biocentric views of forest management, while users in closer proximity to the
resource tend to hold more anthropocentric views. Users closer to the resource tend to be
more aware of the importance of the study area for migratory forest songbirds; however,
users farther away tend to express higher levels of concern for protection of songbird

habitat in the study area. These results suggest that forest management attitudes may

120

play an important role in explaining the probability that an individual will support the
forest easement program.

In order to test the hypothesis that forest management attitudes affect WTP, a
variable was created that incorporates several attitude variables from the survey, each of
which reflects a range of anthropocentric to biocentric attitudes towards forest
management in the study area. An index of several attitude statements is preferred to
using a set of individual attitude statements because of the potential colinearity among
the separate attitude indicators (McClelland 2001). Attitude statements in the following
table are therefore summed to form an attitude index variable, biocentric, for which high
scores reflect biocentric attitudes and low scores reflect anthropocentric ones. The
sample average for each individual attitude statement was used to impute values for
attitude variables with missing values. The index variable, biocentric, was calculated by

summing values for the attitude variables listed in Table 4.6.

121

Table 4.6. Attitude Variables Included in the Calculation of biocentric

 

Variable

imptcons

Attitude statements/questions and scale

How important is it to you to protect habitat for migratory forest songbird
species of conservation concern in the Western U.P. Study Forest?
1 (Not at all important) H 4 (Very important)

 

prothab

In my opinion, the Western U.P. Study Forest should be managed to protect
habitat for migratory forest songbird species of conservation concern even if it
results in economic losses to forest-based industries.

1 (Strongly disagree) H 5 (Strongly agree)

 

incrdiv

In my opinion, the Western U.P. Study Forest should be managed to increase
migratory forest songbird diversity even if there are economic losses to forest-
based industries.

1 (Strongly disagree) H 5 (Strongly agree)

 

concernhab

humneed

commneed

How concerned are you about migratory forest songbird diversity in the
Western U.P. Study Forest?
1 (Not at all concerned) H 4 (Very concerned)

In my opinion, the Western U.P. Study Forest should be managed to meet the
needs of people.
1 (Strongly agree) H 5 (Strongly disagree)

In my opinion, the Western U.P. Study Forest should be managed to meet the
needs of communities that are economically dependent on forests, no matter
what effect this has on the environment.

1 (Strongly agree) H 5 (Strongly disagree)

 

mtnindjob

In my opinion, the Western U.P. Study Forest should be managed to maintain
forest industry jobs.
1 (Strongly agree) H 5 (Strongly disagree)

 

biocentric

Attitude index created by summing responses to all attitude statements listed
above
7 (Anthropocentric attitudes) H 33 (Biocentric attitudes)

The hypothesis that forest management attitudes affect WTP is tested by

estimating a random effects probit model using program attributes, cost and the attitude

index variable biocentric as explanatory variables. Model 2 results, reported in Table

4.7, show that biocentric has a positive and highly significant effect on WTP. This

indicates that individuals who hold biocentric attitudes towards forest management in the

study area are more likely to vote for the forest easement program than individuals who

122

hold anthropocentric attitudes. However, this result applies to individuals holding these
attitudes regardless of where they live in relation to the resource. To investigate how
attitudes and distance together affect WTP, a random effects probit model is estimated
again using program attributes, cost, distance and biocentric as explanatory variables.
The results of this model, Model 3 in Table 4.7, show that biocentric again has a highly
significant and positive effect on WTP and that the effect of distance on WTP is no
longer significant. This suggests that the biocentric variable is explaining part of the
positive distance effect that was found in Model 1.

Model 4 reports results of estimation of a random effects probit model using
program attributes, cost, distance, biocentric, as well as the interaction of biocentric with
distance. These results show that biocent‘dist has a positive and significant effect, while
distance now has a negative and significant effect. This suggests that individuals who
hold biocentric attitudes and live farther away from the resource are more likely to be
WTP than individuals who hold biocentric attitudes and live closer to the resource.
Biocentric is again highly significant and positive, indicating the individuals, regardless
of location, with biocentric attitudes are more likely to be WTP than individuals with
anthropocentric views. When these attitude effects are controlled for, the effect of
distance is negative.

It can be seen from the model results reported in Table 4.7 that distance, while it
was shown to have a positive and significant effect when included alone as an
explanatory variable, no longer has the same effect when biocentric and biocent‘dist are
included as explanatory variables. The biocentric attitude index has strong explanatory
power and has a highly significant positive effect on the probability that an individual

will vote ‘yes’ for the program. In general, individuals with biocentric attitudes are more

123

likely to support the program than individuals with anthropocentric attitudes. This effect
also holds as distance from the resource increases. This suggests that individuals with
biocentric attitudes that live farther away from the resource have a higher willingness to
pay for the program than individuals with biocentric attitudes who live closer to the
resource. The inclusion of the variable that indexes attitudes changes the effect of
distance, and when the interaction of attitudes and distance is included, the distance
decay hypothesis is supported by the results. Likelihood values indicate that Model 4
performs better than the other models, which suggests that distance, biocentric and
biocent‘dist should be retained in the model.

In the models discussed above, the program attribute variable coefficients are
almost identical to those estimated in Model 1, and changes in the estimated coefficients
for the program attributes estimated between Models 2, 3 and 4 are almost negligible.

Results for program attribute variables, therefore, are very stable across the models.

124

Table 4.7. Random Effects Probit Estimation Results of Distance and Attitudes

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Models
Variable Model 2 Model 3 Model 4
Constant —4.0770"'** -4.0940"'" -3.4340”*
(0.3850) (0.3848) (0.5230)
indjobs 0.0079" 0.0075“: 0.00750”
(.0008) (0.0008) (0.0008)
rtjobs 0.0069“* 0.0069“* 0.0069”"‘
(0.0012) (0.0012) (0.0012)
birddiv 0.0125‘" 0.0126‘“ 0.01270“
(0.0026) (0.0026) (0.0026)
birdcons 0.0229" 0.0228" 0.0228”
(0.0094) (0.0094) (0.0094)
deer -0.0170""' 0.01700" ~0.0170““
(0.0054) (0.0054) (0.0054)
0031 -0.0070"'" -0.0070"* -0.0071"'"
(0.0005) (0.0005) (0.0005)
distance . 0.0002 -0.0032"
(0.0004) (0.0020)
biocentric 0.162?" 0.1609*” 0.1265‘"
(0.0181) (0.0182) (0.0260)
biocent‘dist 0.0002"
(0.0001)
Rho 0.7429m 0.7415m 0,7412...
(0.0224) (0.0225) (0.0224)
N 3618 3618 3618
# of groups 933 933 933
Log likelihood -1377.88 -1377.713 -1376.09
Prob>chi2 0.000 0.000 0.000

 

* =p < 0.10; " =p < 0.05; =p < 0.01

125

4.3.3 Distance, recreational use and attitude models

The previous section showed that forest management attitudes can be used to
explain why WTP for the easement program increases as distance from the resource
increases. However, an examination of recreational use of the resource may provide
additional information about the factors that explain this positive distance effect.

Variables used in the following analyses are defined in Table 4.8.

Table 4.8. Variable Definitions for Recreational Use Models

 

 

 

 

 

 

 

 

 

 

 

Variable Definition

user Dummy variable equal to one if respondent is a recreational user of
the study area

user *dist Interaction of user and distance

nonuser Dummy variable equal to one if respondent is not a recreational user
of the study area

nonuser ’dist Interaction of nonuser with distance

hs Dummy variable equal to one if respondent has participated in only
hunting and/or snowmobiling in the study area

hs‘dist Interaction of hs with distance

other Dummy variable equal to one if respondent has participated in any
other recreational activities in the study area and has not participated
in hunting or snowmobiling

other *dist Interaction of other with distance

allrec Dummy variable equal to one if respondent has participated in
hunting and/or snowmobiling in the study area and has participated in
other recreational activities in the study area

allrec *dist Interaction of allrec with distance

 

This section tests the hypothesis that recreational users who live farther away

from the resource are more likely to support the program than recreational users who live

126

closer to the resource. To test this hypothesis, a dummy variable for recreational users is
interacted with distance and included as an explanatory variable in the model, along with
distance. Model 5 in Table 4.9 shows the estimation results. In this model, distance is
not significant and user‘dist is positive and highly significant, indicating the users who
live farther from the resource are willing to pay more than users who live closer to the
resource.

It is generally expected that, all else equal, use values decline with distance, and it
is therefore surprising that users who live farther from the resource would be WTP more
than users who live closer to the resource. A deeper look at the causes of this is
wan'anted, and one way to do this is to look at different types of recreational uses.
Recreational activities are often categorized as consumptive, e.g. hunting or firewood
gathering, versus non-consumptive uses, e.g. birdwatching or biking (Li et al. 2003,
Wilson and Tisdell 2001, Benson 2001). Studies have shown that non-consumptive
recreational users tend to have stronger conservationist attitudes than consumptive
recreational users (Porter and Bright 2003). This provides the rationale for the
hypotheses that certain types of recreational users will have higher WTP than other types
of users and that WTP for certain types of recreational users will increase as distance
from the resource increases.

In order to determine what type of recreational use distinctions were appropriate
for this research, survey data on 22 different recreational uses was investigated for
correlations between activities. High correlation (>0.50) was found among deer hunters,
bird hunters, other hunters and snowmobilers. High correlation was also found among
recreational users who participate in all other recreational activities, such as berry

picking, birdwatching, boating, firewood gathering, fishing and hiking. Because these

127

correlations do not distinguish themselves based on consumptive versus non-
consumptive uses, this study distinguishes between people who participate in only
hunting and snowmobilig (hs), those who participate in other recreational uses but not in
hunting or snowmobiling (other), and those who participate in hunting, snowmobiling
and at least one other recreational activity (allrec).

The recreational use dummy variables, other and allrec are included in the model
as well as a dummy variable that represents those who are not recreational users of the
study area, nonuser. Parameter estimates of these dummy variables measure the effect of
each dummy variable relative to recreational users who participated in hunting and or
snowmobiling in the study area (hs). Interactions of these dummy variables with distance
are included to capture the effect of distance on the probability of voting ‘yes’ by non-
users and users. All of the distance/recreational use interaction terms together account
for the distance effect.

Model results are reported in Table 4.9. The effect of nonuser and other on WTP
is positive but not significant, and the effect of allrec is positive but not significant.
These results weakly support the hypothesis that there are certain types of recreational
users who are WTP more than other types of users. The interaction of recreational use
variables with distance is only significant for allrec *dist, supporting the hypothesis that
there are certain types of recreational users whose WTP increases with increasing
distance from the resource. Individuals who participate in a combination of hunting,
snowmobiling and other activities and live farther from the resource have higher WTP
than those types of users living closer to the resource. Nonuser‘dist has a negative effect
on WTP but is not significant. The effects of other and allrec users compared to hs users

are positive but not significant, which weakly supports the hypothesis that certain types

128

of recreational users are willing to pay more than other types of users.

In order to examine what happens when attitude variables are included with
recreational use variables to estimate the combined effects of distance, attitudes and
recreational use on WTP, Model 7 is estimated including all of these variables. The
interaction of nonuser and distance has a significant and negative effect, indicating that
nonusers living farther from the resource have lower WTP than those living closer to the
resource, which suggests that the distance decay effect holds for non-users. Results are
again not significant for the interaction of recreational use variables with distance.
However, nonuser, other and allrec have positive and significant effects on WTP, which
indicates that these groups have a higher WTP than hs users, supporting the hypothesis
that certain types of recreational users have higher WTP than other users of the resource.
The effect of biocentric is again highly significant and positive, and its interaction with
distance again shows that individuals with biocentric views who live farther away are
more likely to vote ‘yes’ than those who hold biocentric views and live closer to the
resource.

As was the case in the distance and attitudes models, program attribute
coefiicients in the models discussed above display stability across all models. Likelihood
ratio tests of the three models show that the inclusion of the attitude variables in Model 7
significantly improved performance of the model. These results demonstrate that WTP
differs for different types of recreational users, and the attitudes of respondents remain an

important explanatory factor in estimating WTP.

129

Table 4.9. Results of Recreational Use and Attitudes Models

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Model 5 Model 6 Model 7
Constant -1.1666"'Ml -1.2188"'" -4.4853”*
(0.1574) (0.3679) (0.5364)
indjobs 0.0077‘" 0.0076*" 0.0074‘"
(0.0009) (0.0008) (0.0008)
rtjobs 0.0070“* 0.0070"* 0.0069*"
(0.0012) (0.0012) (0.0012)
birddiv 0.0121 ‘" 0.0122‘“ 0.0128*"
(0.0026) (0.0026) (0.0026)
birdcons 0.0238" 0.0238" 0.0222"
(0.0095) (0.0095) (0.0093)
deer -0.0161”"' -0.0162"” -0.017l”“'
(0.0054) (0.0054) (0.0053)
cost -0.0073‘” -0.0073‘“ -0.0070*"”"
(0.0005) (0.0005) (0.0005)
distance -0.0007
(0.0007)
user‘dist 0.0023“*
(0.0007)
nonuser ‘dist —0.0008 -0.0038* "
(0.0023) (0.0017)
hs *dist 0.0001 -0.001 1
(0.0023) (0.0022)
other *dist 0.0010 -0.0026
(0.0001) (0.0017)
allrec *dist 0.0017“ * -0.0015
(0.0007) (0.4036)
nonuser 0.0905 1 .0923 "
(0.4821) (0.6219)
other 0.3507 1 .2748"
(0.4368) (0.6060)
allrec -0.0034 0.9383 ‘
(0.3773) (0.5577)
biocentric 0.1277111“

(0.0222)

 

 

Table 4.9 (continued)

 

 

 

 

 

 

 

Variable Model 5 Model 6 Model 7
biocentric ’dist 0.0001 * '1‘
(0.0001)
Rho 0.7858"" 0.7789*" 0.7327"“
(0.0210) (0.0218) (0.0226)
N 3618 3618 3618
# of groups 933 933 933
Log likelihood -1416.94 -l415.46 -1369.47
Prob>chi2 0.000 0.000 0.000

 

" =P < 0.10; ** =p < 0.05; ““ =17 < 0-01

Implicit prices of attributes can be calculated by dividing parameter estimates by

the estimated coefficient on the cost variable, as shown in the following equation:

 

ring).

The following table shows implicit prices calculated based on results from Model 7.

Table 4.10. Implicit Prices of Model Attributes

 

 

 

 

 

 

Attribute Implicit Price
inay'obs $1 .06

rg’obs $0.99

birddiv $1 .83

birdcons $3.14

deer -$2.44

 

131

Willingness to pay for an additional forest industry job is slightly higher than a
forest-based recreation and tourism job. Individuals are WTP more for an additional unit
of bird diversity than for an additional job, either forest industry or recreation and
tourism-based. WTP is higher still for an additional songbird species of conservation
concern at or above its target population level. Individuals need to be compensated for
an additional unit of deer browse high enough to affect tree regeneration.

Because of the interactions present in some of the variables, implicit prices
calculated from the recreational use and attitude variables are more difficult to interpret
without reference to the levels of the variables being interacted. To address this, the
implicit prices for the variables nonuser, other and allrec, were calculated as if distance
was zero, and these values were then used to calculate relative WTP for hs, other, and
allrec users and non-users at different distances from the resource and at different levels
of biocentrism. Establishing hs users (the base case for the other, allrec and nonuser
dummy variables) who live closest to the resource and have low levels of biocentrism as
a baseline, WTP values relative to this baseline were calculated to investigate relative

WTP values. Table 4.11 reports these relative WTP values.

132

Table 4.11. Relative WTP of hs, other and allrec Users and Non-Users at Different
Distances from the Resource and Different Levels of Bioeentrisml

 

 

 

 

 

 

 

 

 

 

Distance Level of hs other allrec N on-

(miles)z Biocentrism3 Users Users Users Users
38 Low $0 $175 $132 $141
38 Medium $122 $297 $255 $263
38 High $244 $419 $377 $386
118 Low $2 $160 $131 $112
1 18 Medium $1 3 1 $290 $260 $242
1 18 High $261 $419 $390 $371
369 Low $7 $115 $125 $21
369 Medium $160 $268 $278 $174
369 High $313 $421 , $431 $326

 

’ The table presents the relfive difl’erence in WTP, holding all else equal, across
gradients of distance, biocentrism, and user type relative to the WTP for hs users that
have low biocentrism and live 38 miles fi'om the study area.

2 The distances represent the first quartile, median and third quartile for distances in the
sample.

3 The levels of biocentrism represent the first quartile, median and third quartile in the
sample for the biocentrism index that ranges fiom 1 to 33.

Table 4.11 shows the relative WTP of hs, other and allrec users and non-users at
different distances from the resource and at different levels of biocentrism. A low level
of biocentrism corresponds to anthropocentric attitudes. The values in the table illustrate
relative differences in WTP amounts between different types of recreational users and
non-users and how these differences are affected by distance from the resource and
environmental attitudes. At 38 and 118 miles from the resource, other users have higher
relative WTP at each level of biocentrism than hs, allrec and non-users. At 369 miles

from the resource, allrec users have the highest relative WTP. Relative WTP of other,

133

 

allrec users and non-users is much higher than that of hs users at distances of 38 and 118
miles from the resource, but at 369 miles from the resource, the relative WTP of hs users
and non-users is similar.

The relative WTP of strongly biocentric other and allrec users who live farther
from the resource is slightly higher than for those types of users who are strongly
biocentric and live close to the resource. Relative WTP is much higher for strongly
biocentric other and allrec users than those types of users living close to the resource
who have low to medium biocentrism. The relative WTP of anthropocentric and
biocentric non-users decreases as distance increases, supporting the distance decay
hypothesis.

These results suggest that environmental attitudes have a strong effect on WTP.
Individuals with biocentric values and attitudes have a higher WTP relative to individuals
with anthropocentric values. This relative effect exists even as distance from the
resource increases. Recreational users who participate in activities other than hunting
and snowmobiling and those who participate in hunting, snowmobiling and other
recreational activities have higher WTP at all distance gradients and biocentrism levels
than people who only hunt or snowmobile. The results demonstrate that although types
of recreational use and non-use affect WTP values, the predominant factor affecting WTP

in this analysis is the presence of biocentric attitudes.

4.4 Conclusions
This study sought to understand the effects of various factors on WTP in order to
determine the extent of the market for the provision of a forest easement program in an

area of Michigan’s Upper Peninsula. The analysis used stated preference and attitudinal

134

data collected from a mail survey of Michigan residents to estimate an attribute-based
referenda model. A series of random effects probit models were estimated to identify the
effects of distance, forest management attitudes and recreational use on predicting WTP.

Model results revealed a positive distance effect when distance was included
without attitudinal or recreational use variables. This results in the rejection of the
distance decay hypothesis, which is contrary to the results found throughout the
nonmarket valuation literature on distance effects. The distance effect, however, changed
when attitudinal and recreational use variables were included in the model.

The inclusion of an attitudinal index variable that reflects a range of
anthropocentric to biocentric attitudes towards resource management eliminated the
positive distance effect apart from its positive interaction with biocentric attitudes. The
strong explanatory power of the attitude variable shows that attitudes are significant
predictors of WTP. Results of recreational use models showed that use values increase as
distance from the resource increases, which was a counterintuitive result. However, the
inclusion of the attitude variable in this model once again changed the positive effect of
use interacted with distance.

Recreational users who participate in activities other than hunting and
snowmobiling, users who participate in a combination of all activities and nonusers are
WTP substantially more than users who participate in only hunting and snowmobiling.
Results show that a significant negative distance effect exists for individuals who are not
users of the resource when attitude variables are included in the model. Attitudes have a
strong influence on WTP values, and individuals with biocentric attitudes have a higher
WTP than individuals with anthropocentric attitudes. The WTP of individuals with

biocentric attitudes increases as distance from the resource increases.

135

The results of this research show that the market for the implementation of a
forest easement program in Michigan’s UP extends into areas of the state that are a large
distance away from the resource. The results confirm the hypothesis that attitudes are
predictors of WTP and support the results of other studies that have shown that it is
appropriate to include attitudes in nonmarket valuation analyses. Results also indicate
that individuals who live far away from the resource and who value it can be identified
by their environmental attitudes or by their recreational use of the area. Results show
that non-use values decline with distance but values of recreational users increase with
distance. This may be due to increasing use values with distance that offset the distance
decay effect found for non-users. However, this also may be due to increasing non-use
values with distance among users. Further research on the effects of distance and use on
WTP is warranted.

This study showed that biocentric and anthropocentric values that underlie
environmental attitudes play an important role in estimating nonmarket values of
environmental goods and services. Results provide insight into the factors that influence
individuals’ choices and behavior, which can be a useful input to the policy making
process. Results show that there are strong interactions between distance, attitudes and
types of recreational use/non-use that influence WTP. Information on the types of beliefs
and values that motivate individuals to support certain policy actions can help predict
which members of the public will be more willing to support particular conservation

initiatives.

I36

REFERENCES

AAPOR. 2004. Standard Definitions: Final Dispositions of Case Codes and Outcome
Rates for Surveys. Edited by American Association for Public Opinion Research.
3rd ed. Lenexa, KS.

Addelman, S., and O. Kempthome. 1961. Orthogonal Main-Effect Plans. Ames, IA: Iowa
State University, Aeronautical Research Laboratory, Office of Aerospace
Research, United States Air Force.

Bandara, R. and C. Tisdell. 2003. Comparison of rural and urban attitudes to the
conservation of Asian elephants in Sri Lanka: empirical evidence. Biological
Conservation 1 10:327-342.

Bateman, I.J., and I.H. Langford. 1997. Non-users' willingness to pay for a national park:
An application and critique of the contingent valuation method. Regional Studies
31 (6):571-582.

Benson, DE. 2001. Wildlife and recreation management on private lands in the United
States. Wildlife Society Bulletin 29 (1):359-371.

Boxall, P., K. Rollins, and J. Englin. 2003. Heterogeneous preferences for congestion
during a wilderness experience. Resource and Energy Economics 25:177-195.

Dietz, T., P.C. Stern, and G.A. Guagnano. 1998. Social structural and social
psychological bases of environmental concern. Environment and Behavior
30:450-471.

Dillman, DA. 2000. Mail and Internet Surveys: The Tailored Design Method, 2nd
edition. 2nd ed. New York: John Wiley and Sons, Inc.

Freeman, A.M.III. 2003. The Measurement of Environmental and Resource Values:
Theory and Methods. Washington, D.C.: Resources for the Future.

Hanley, N., F. Schlapfer, and J. Spurgeon. 2003. Aggregating the benefits of
environmental improvements: distance-decay functions for use and non-use
values. Journal of Environmental Management 68:297-304.

Holmes, T.P., and K.J. Boyle. 2005. Dynamic learning and context-dependence in
sequential, attribute-based, stated-preference valuation questions. Land
Economics 81 (1):] 14-126.

Johnson, C.Y., J.M. Bowker, and H.K. Cordell. 2004. Ethnic variation in environmental
belief and behavior: An examination of the New Ecological Paradigm in a social
psychological context. Environment and Behavior 36 (2):157-186.

137

Johnson, F ., R. Dunford, W. Desvouges, and M. Banzhaf. 2001. Role of knowledge in
assessing nonuse values for natural resource damages. Growth and Change 32:43-
68.

Li, C.L., H.C. Zinn, and SC. Barro. 2003. A cross-regional comparison of recreation
patterns of older hunters. Leisure Sciences 25 (1)21-16.

Loomis, J. 1996. How large is the extent of the market for public goods: evidence from a
nationwide contingent valuation survey. Applied Economics 28:779-782.

McClelland, E. 2001. Measurement issues and validity tests for using attitude indicators
in contingent valuation research, Working Paper #01-01. Washington, D.C.: U.S.
Environmental Protection Agency National Center for Environmental Economics.

McFadden, D. 1974. Conditional logit analysis of qualitative choice behavior. In
Frontiers in Econometrics, edited by P. Zarembka. New York: Academic Press.

Nordlund, A.M., and J. Garvill. 2002. Value structures behind proenvironmental
behavior. Environment and Behavior 34 (6):740-756.

Pate, J ., and J. Loomis. 1997. The effect of distance on willingness to pay values: a case
study of wetlands and salmon in California. Ecological Economics 20:199-207.

Poortinga, W., L. Steg, and C. Vlek. 2004. Values, environmental concern, and
environmental behavior: A study into household energy use. Environment and
Behavior 36 (1):70-93.

Porter, R., and AD. Bright. 2003. Non-Consumptive Outdoor Recreation, Activity,
Meaning and Environmental Concern. In Proceedings of the 2003 Northeastern
Recreation Research symposium, edited by J. Murdy. USDA Forest Service
Northeastern Research Station, General Technical Report NE—317. Newtown
Square, PA.

Steel, B., P. List, and B. Shindler. I994. Conflicting values about federal forests: A
comparison of national and Oregon publics. Society and Natural Resources
7:137-153.

Stern, P.C., T. Dietz, and G.A. Guagnano. 1995. The new ecological paradigm in social-
psychological context. Environment and Behavior 27 (6):723-743.

Stevens, T.H., J. Echeverria, R.J. Glass, T. Hager, and T.A. More. 1991. Measuring the
existence value of wildlife: What do CVM estimates really show? Land
Economics 67 (4):390-400.

Streever, W.J., M. Callaghan-Perry, A. Searles, T. Stevens, and P. Svoboda. 1998. Public
attitudes and values for wetland conservation in New South Wales, Australia.
Jounral of Environmental Management 54:1-14.

138

Sutherland, R.J., and R.G. Walsh. 1985. Effect of distance on the preservation value of
water quality. Land Economics 61 (3):281-291.

Tarrant, M., and H. Cordell. 2002. Amenity values of public and private forests:
Examining the value-attitude relationship. Environmental Management 30
(5):692-703.

Wilson, C., and C. Tisdell. 2001. Sea turtles as a non-consumptive tourism resource
especially in Australia. Tourism Management 22 (3):279-288.

Wooldridge, J .M. 2002. Econometric Analysis of Cross Section and Panel Data.
Cambridge, MA: MIT Press.

139

Conclusions

This research contributes to the understanding of sociocultural and economic
values of ecosystem services in a forested area of Michigan’s Upper Peninsula. The first
essay is based on results of focus group discussions that were also used to develop a mail
survey instrument. Essays 2, 3 and 4 are based on results of the mail survey of 2,000
Michigan residents.

The work establishes links between environmental attitudes, underlying
environmental value orientations, socioeconomic characteristics, and economic values.
The results contribute to a broader understanding of the factors that influence individuals’
behavior with respect to environmental conservation issues. The connections established
in this research follow the cognitive hierarchy framework presented in Essay 2, Figure
2.1, in which economic values can be linked to behavior, which can in turn be predicted
by attitudes, which can be linked to underlying environmental value orientations.

The first research objective was to gain a better understanding of public views
and perceptions of forest-human-wildlife interactions in an area of Michigan’s Upper
Peninsula. Essay 1 reports results from a series of focus group discussions held in and
near the study forest. Content analysis of focus group data reveals that the value
orientation of rural, timber-dependent individuals and urban, non-timber dependent
individuals does not lie as expected along an anthropocentric/biocentric continuum. The
rural participants did not express purely extractive, utilitarian views of forest
management, as would be expected according to much of the literature on value
orientations of resource dependent community residents. Findings with respect to urban

participants also contradict much of the literature that suggests that urban, non-timber

140

dependent residents hold strong biocentric views of resource management. The urban
participants in this study did not express strong opinions in favor of ecological forest
values but did express strong concerns for anthropocentric forest uses such as their own
recreation.

The second objective of this research was to understand how demographic
characteristics and environmental and forest management attitudes are related. This was
accomplished in Essay 2 by examining survey respondent attitudes towards the
environment and towards forest management in the study area. Factor analysis was used
to reduce a set of attitude variables into a set of 5 factors that reflect a range of
anthropocentric to biocentric value orientations. The five factors extracted from the
analysis are labeled Balance, Environment, Management, Wildlife and Human Role.
Factor scores were created from the factors for each respondent in the sample and a series
of stepwise regressions regressed factor scores for each of the 5 factors onto a set of
demographic characteristics.

Results of Essay 2 show that area of residence, age, membership in an
environmental organization and political views had significant effects in explaining
factor scores. The factors extracted from the analysis reflected anthropocentric and
biocentric value orientations. Results suggest that residents of the study area, members
of forestry organizations, older individuals and politically conservative individuals tend
to hold anthropocentric views of the environment and forest management in the study
area while members of environmental organizations were found to hold biocentric views.
The results of this research demonstrate that there is a connection between demographic
characteristics and environmental and forest management attitudes and value

orientations. The low explanatory power of the regression models, however, suggests

141

that demographic characteristics do not provide sufficient information to adequately
explain variation in attitudes and values towards forest ecosystem management and the
environment.

The third objective of the research was to estimate non-market values of forest
ecosystem characteristics in Michigan’s western Upper Peninsula. An attribute-based
referenda (ABR) model was used to estimate willingness to pay for a forest easement
program. The forest easement program was described in the survey using sets of
attributes that were created based on an orthogonal main-effects experimental design.
The binomial choice stated preference data, attribute levels and demographic
characteristics were used as inputs to a random effects probit model. Results report the
estimated effects of these variables on public support for the forest easement program and
show that individuals hold values for the nonmarketed attributes of forest ecosystems in
the study area.

The attributes for which nonmarket values were estimated reflect the social and
ecological importance of forests in the study area. Attributes in the analysis include
forest industry jobs, forest-based recreation and tourism jobs, forest migratory songbird
species diversity, number of forest migratory songbird species of conservation concern
and the effects of deer browse on tree regeneration. Results show that people value both
the economic (job related) and ecological attributes of a possible forest easement
program in the study area. Resource management and policy decisions affecting the
study area that cause reductions in forest-based employment will be less acceptable to the
public. Programs that improve bird habitat, reduce deer browse and reduce forest-based
employment will be less acceptable than programs that improve ecological attributes

while also improving or maintaining forest-based employment opportunities.

142

Results also indicate that older individuals, members of a hunting club and
politically conservative individuals are less likely to vote for and have lower willingness
to pay for the forest easement program while members of environmental organizations,
individuals with higher levels of education, residents of urban areas and individuals who
have participated in recreational activities in the study area are more likely to vote ‘yes’
for the easement program and have higher willingness to pay.

The fourth objective of the research was to determine how distance, attitudes and
recreational use affect the results of nonmarket valuation analysis of forest ecosystem
attributes in Michigan’s western Upper Peninsula. Essay 4 reports model results from
maximum likelihood estimation of random effects probit models using the stated
preference data collected from the survey. Models included various combinations of
distance, attitude and recreational use variables. Results show that while distance is
found to have a positive effect on WTP, which is a counterintuitive result, this effect is
mitigated by the inclusion of attitudinal and recreational use variables. Attitudes that
reflect anthropocentric/biocentric value orientations significantly affect WTP, and a
significant effect was also found for the interaction of distance with attitudes. Particular
types of recreational users (those who have participated in a variety of recreational
activities in the study forest) are more likely to vote ‘yes’ for the program than
recreational users that only participate in hunting and/or snowmobiling.

Information gained from the analyses in Essay 4 provides an indication of the
extent of the market for a forest easement program in the study area. Results indicate that
there are subpopulations of individuals who live further from the resource who are more
likely to support the forest easement program than individuals living closer to the

resource. In addition, the positive effect of distance on WTP is explained by the value

143

orientation of individuals. Those individuals holding biocentric attitudes towards
resource management have higher WTP relative to individuals who hold anthropocentric
views. While recreational use and non-use were found to affect relative WTP values,
changes in relative WTP between different types of recreational users and non-users at
different distances from the resource was predominantly affected by biocentric attitudes.
The results indicate that including environmental attitudes in nonmarket valuation studies
can substantially improve one's ability to explain preferences and WTP.

The work as a whole illustrates the connections that exist between and
sociocultural and economic ecosystem service values, socioeconomic variables and
attitudes. The essays informed each other in the investigation of these links. Results of
Essay 1 reported urban-rural differences in perceptions of foreSt management but were
limited to a small sample of individuals living in close proximity to the resource. These
results, however, pointed towards the need for a deeper investigation of the effects of
demographic characteristics on environmental attitudes. Essay 2 built upon these results
by analyzing the underlying structure of environmental attitude data and connecting it to
sociocultural ecosystem service values within an anthropocentric/biocentric framework.
Environmental attitudes were further connected to demographic characteristics to
describe segments of the public who tend to hold particular sociocultural value
orientations. The explanatory power of the demographic characteristics, however, was
not strong, suggesting that other characteristics should be explored in future research to
explain environmental attitudes. For example, connections could be investigated
between attitudes and pro-environmental consumer behavior or knowledge of

environmental issues. Essay 2 results, nevertheless, suggested a deeper investigation of

I44

the environmental value/attitude connection and how this is linked to behavior, which
was undertaken in Essay 4.

The third essay sought to estimate preferences for forest ecosystem services and
to connect socioeconomic characteristics to economic ecosystem service values. Results
showed that certain demographic characteristics were significant in explaining these
values, connecting individuals’ socioeconomic characteristics to their stated preferences
and their behavior. Essay 4 continued to explore the effects of respondent characteristics
on economic values by including distance, environmental attitudes and types of
recreational use in the analysis. Results showed that economic values are influenced not
only by demographic characteristics but also by environmental attitudes and that in fact,
attitudes have stronger explanatory power in nonmarket valuation than socioeconomic
characteristics alone. Essay 4 results connect attitudes and underlying environmental
value orientations to economic values, thereby establishing a connection between
sociocultural and economic ecosystem service values.

The research has several limitations. The sample used in the survey research was
stratified into four regions of Michigan. A limitation of the sample design is that the
division of Michigan into four geographic strata may not have been the best way to
capture regional differences across the state. Perhaps the division of the sample into
more regional strata would capture differences between regions within Michigan more
successfully. For example, strong regional differences may exist within the Upper
Peninsula, and dividing the UP sample into more than 2 regions may have better captured
these differences.

Response rates differed by sample strata, but there is not enough information

available to determine whether there was a non-response bias based on regions.

145

Responses were relatively higher from the study area and the rest of the Upper Peninsula
than for the Lower Peninsula strata.

The nonmarket valuation component of the research resulted in a large number of
‘no’ responses. Therefore, when mean WTP is computed over the unrestricted interval
(-oo, oo), i.e. when negative WTP is allowed, the results yield negative aggregate WTP
values for some forest easement program scenarios. We can identify respondents who
are indifferent to the choice based on survey questions that asked whether the individual
would ever vote for the program regardless of the price. This problem of a large number
of ‘no’ responses could be investigated further by estimating a spike model to account for
the effect of indifferent respondents.

It was expected that distance from the study site would be positively correlated
with income, however, the correlation between distance and income was not significant.
Distance and biocentric were found to be significantly positively correlated, but the
correlation is very low. Distance was positively related to WTP in some of the models,
and although negative distance effects are typically found in the nonmarket valuation
literature, this positive distance effect may not be counterintuitive. According to studies
in the social forestry literature, urban residents tend to have strong biocentric attitudes.
There is a much higher concentration of urban areas in the southern portion of Michigan,
and individuals from the southern LP strata are more likely to live in or near a major
urban center than residents of the northern LP or the UP. This may explain the positive
distance effect and the positive relationship between distance and biocentric attitudes.

Descriptive analysis of regional differences in environmental attitudes (reported
in Appendix 7) reveals that individuals across all regions tended to agree with statements

about the importance of managing forests to benefit wildlife and the importance of

146

achieving balance between ecological and economic goals of forest management. These
results suggest that some of the individuals identified as anthropocentric in this research
may be better characterized as having a mixture of anthropocentric and biocentric
attitudes. The regional comparison of attitude statements indicates that UP residents
tended to disagree with forest management that improves ecological attributes at a cost to
the local economy. Individuals living closer to the resource will bear more of the costs of
economic losses to forest-based industries in the area than people living farther away.
Therefore, it is logical that people in the UP would express attitudes that favor forest
management to meet human needs. However, the UP respondents’ attitudes towards the
importance of achieving balance is not significantly different than those of LP
respondents. This warrants a deeper investigation of attitudes towards balance of
ecological and economic forest management goals. Individuals living close to or in the
study area may have biocentric views of resources in other parts of Michigan, where
changes in forest ecosystems may not directly affect the economic well being of their
communities.

The finding of significant preferences for economic as well as ecological forest
ecosystem services demonstrates public support for multiple use forest management.
Due in part to increasing public concern for the sustainable use of forest resources, forest
management for multiple uses has been increasing in its application throughout the US
over the past few decades. Information on preferences for multiple use forest
management as well as the types of people who hold those preferences can contribute to
connecting forest management activities with publicly acceptable outcomes.

The work provides information useful to policy makers and resource managers on

the characteristics of individuals who hold particular environmental attitudes and

147

examines how these characteristics influence individuals’ economic ecosystem service
values, which, in turn, can assist in predicting the behavior of different segments of the
public and can help predict acceptability of conservation initiatives. For example, essay
4 showed that values of individuals who hold biocentric attitudes increase with increasing
distance from the resource, indicating that public support for conservation initiatives in
the study area exists in areas of Michigan at great distances from the study forest.

Results of this research demonstrate public support for forest management that
provides multiple uses. For example, the public cares not only about forest industry jobs
but also about forest recreation and tourism jobs in the study area. The management of
forests for multiple uses contributes to the sustainable management of forest resources,
and information on public preferences for this type of management can help ensure its
continued implementation.

The results also indicate that the public values the ecosystem services provided by
forests, such as songbird species diversity. A manager could use these results to inform
legislators of public concern for these ecosystem services and of the values people hold
for them in order to provide support for conservation initiatives to protect the provision
of these services. Results also show that there is public support for a particular type of
conservation initiative, the forest easement, although the support is not overwhelmingly
strong.

The essays as a whole explain different aspects of individuals’ motivations and
behavior with respect to natural resource management. The research contributes to
understanding the characteristics of people that influence their acceptance of particular
types of natural resource policies. The results of this research provide information on the

sociocultural and economic values that people hold for ecosystem goods and services.

148

These values are further linked to characteristics and attitudes, which establishes
connections that can be used to help understand and predict human actions that feed back

to affect the provision of ecosystem goods and services.

149

 

APPENDIX 1:

SURVEY SAMPLE DESIGN

150

The survey sample was designed as a stratified sample of 1,750 households from
each of four regions of Michigan for a total sample size of 7,000. The research aims to
understand public perceptions of and values for forest resources within a particular area
of Michigan’s Upper Peninsula (referred to in this appendix as the study area). It was
therefore important to obtain a representative sample of residents of the study area, as
well as residents of the Upper Peninsula, in addition to residents of Michigan’s Lower
Peninsula. Therefore, the sample was stratified to represent four regions: the Study Area
(SA), the Rest of Upper Peninsula (RUP), the Northern Lower Peninsula (N LP) and the
Southern Lower Peninsula (SLP). From the initial stratified sample of 7,000 Michigan
residents, 500 households were randomly chosen from each strata for form a total sample
size of 2,000 for the survey mailing.

The survey sample was purchased from Survey Sampling International and was
formed by randomly choosing households from U.S. Census 2000 Block Groups and
Michigan counties. The first stratum, SA, was selected from the Census Block Groups
that fall most closely within the ecologically defined study area boundaries (See Table
A.1.1). The RUP stratum was randomly selected from the remaining Census Block
Groups that fall within study area counties but are not in the study area, as well as all
remaining Upper Peninsula counties (See Table A.1.2). The NLP stratum was randomly
selected from all counties in the Northern Lower Peninsula (See Table A. l .3). The SLP
stratum was randomly selected from all counties in the Southern Lower Peninsula but
was adjusted to ensure that large urban areas would not be over-represented in the sample
(See Table A.1.4).

The Southern LP includes three very populous counties: Wayne, Macomb and

Oakland. Each of these counties contains 24, 9 and 14 percent, respectively, of the entire

151

population of the Southern LP (US Census 2005). The total number of households in the
SLP in 2002 was 2,838,765, and 39.6 percent of these households were located in
Wayne, Macomb and Oakland counties (US Census 2005). In order to avoid over-
representing these very populous counties in the sample, the number of households
sampled from these three counties was designed to be proportionate to the three counties’
share of Southern LP households. Therefore, 693 households were sampled from Wayne,
Macomb and Oakland counties (the proportion of SLP households in the most populous
counties, 39.6%, multiplied by the total SLP sample size, 1,750), and 1,057 households
were sampled from the remaining counties in the SLP (the proportion of SLP households

in the remaining counties, 60.4%, multiplied by the total SLP sample size, 1,750).

152

Table A.1.1. Sample Frame for Stratum 1: Study Area

 

 

 

 

 

 

 

 

County Census Tract Census Block Group
Dickinson 9501 1

Dickinson 9502 All

Iron 9801 1

Marquette 0022 1

Marquette 0023 1

Marquette 0026 All

Menominee 9601 1

 

Table A.1.2. Sample Frame for Stratum 2: Rest of U.P.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

County Census Tract Census Block Group
Alger All All
Chippewa All All
Delta All All
Dickinson 9503 All
9504 All
9505 All
9506 All
9507 All
Gogebic All All
Houghton All All
Iron 9801 2, 3 & 4
9802 All
9803 All
9804 All
9805 All

 

153

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

County Census Tract Census Block Group
Keewenaw All All
Luce All All
Mackinac All All
Marquette 0001 All
0002 All
0003 All
0004 All
0005 All
0006 All
0007 All
0008 All
0009 All
0010 All
0011 All
0012 All
0013 All
0014 All
0015 All
0016 All
0017 All
0018 All
0019 All
0020 All
0021 All
0022 2 & 3
0023 2 & 3
0024 All

 

154

 

 

 

 

 

 

 

 

 

County Census Tract Census Block Group
0025 All
Menominee 9601 2
9602 All
9603 All
9604 All
9605 All
Ontanagon All All
Schoolcraft All All

 

Table A.1.3. Sample Frame for Stratum 3: Northern Lower Peninsula (NLP)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Counties

Alcona Iosco Ogemaw
Alpena Isabella Osceola
Antrim Kalkaska Oscoda
Arenac Lake Otsego

Bay Leelanau Presque Isle
Benzie Manistee Roscommon
Charlevoix Mason Wexford
Cheboygan Mecosta

Clare Midland

Crawford Missaukee

Emmet Montrnorency

Gladwin Newaygo

Grand Traverse Oceana

 

155

Table A.1.4. Sample Frame for Stratum 4: Southern Lower Peninsula (SLP)

 

 

 

 

 

 

 

 

 

 

 

 

 

Counties

Allegan Ingham Ottawa
Barry Ionia Saginaw
Berrien Jackson Sanilac
Branch Kalamazoo Shiawassee
Calhoun Kent St. Clair
Cass Lapeer St. Joseph
Clinton Lenawee Tuscola
Eaton Livingston Van Buren
Gennessee Macomb Washtenaw
Gratiot Monroe Wayne
Hillsdale Montcalm

Huron Oakland

 

156

APPENDIX 2:

SURVEY DESIGN

157

This appendix describes the design of the survey instrument. The hypothetical
forest easement program presented in the survey consisted of a set of attributes. A
baseline level was calculated for each attribute, and the following sections describe how
these baselines levels were estimated. The attributes and their varying levels were
grouped into choice sets according to an experimental design plan. The last section of

this appendix describes the experimental design.

Forest-based recreation and tourism jobs

One of the two economic attributes included in the survey is forest-based
recreation and tourism jobs in the study area. This section describes how the baseline
number of jobs in forest-based recreation and tourism in the study area was calculated.
The calculation was based on estimates of tourism spending for the most popular
recreational activities in the study area, snowmobiling and deer hunting.

Snowmobiling is an important recreational activity throughout Michigan, and it
draws many recreational users in the UP (Stynes et al. 1998). According to a study of the
economic impact of snowmobiling in Michigan, 11% of all tourism spending in the
western Upper Peninsula (UP) is attributable to snowmobiling (Stynes et a1. 1998). Our
study area lies almost entirely within the western UP, and drawing upon the Stynes et a1.
(1998) study, it was assumed that 11% of all tourism spending within the study area is
also attributable to snowmobiling.

To calculate how many forest-based recreation and tourism jobs are due to
snowmobiling activity in the study area, it is necessary to first estimate the total amount
of spending on tourism in the study area. US Census data for the study area can be most

closely approximated by township-level data. Table A2.1 lists the townships that make

158

 

up the study area along with the counties within which they are located.

Table A2.1. Study Area Counties and Townships

 

County

Baraga

Townships

Spurr

 

Dickinson

Breen

 

Felch

 

Norway

 

Sagola

 

Waucedah

 

West Branch

 

Iron

Crystal Falls

 

Mansfield

 

Marquette

Ely

 

Ewing

 

Forsyth
Humboldt

 

Republic

 

Tilden

 

Turin

 

Wells

 

Menominee

Harris

 

Meyer

 

Spalding

159

Tourism spending estimates are not available at the township level from US
Census data. Therefore, spending on tourism in study area townships is extrapolated
from county-level tourism spending data. Employment attributable to tourism spending is
approximated by data on employment in the arts, entertainment, recreation, accomodation
and food service sector, and Table A2.2 presents data on employment in the study area

townships in this sector as a percent of total county employment.

Table A2.2. Study Area Township Service Sector Employment and Estimated
Tourism Spending in 2000

 

 

 

 

 

 

County Study area Service Tourism Spending Estimated Study
Sector Employment as (millions of S) Area Township
Percent of Total County ‘ Tourism Spending
Employment

Baraga _ 1% 14.1 $141,000

Dickinson 1% 26.6 $266,000

Iron 1 .6% 24.1 $385,600

Marquette 1 .7% 82.7 $1,405,900

Menominee 3.5% 23.5 $822,500

 

Sources: US Census 2005a, US Census 2005b, Stynes 2005

Summing study area township tourism spending estimates in Table A2.2. yields a
total study area tourism spending estimate of $3,021 ,000. If we assume that 11% of all
tourism spending in the study area is attributable to snowmobiling, then $3,021,000*1 1%
= $332,310 is spent on snowmobiling in the study area annually. According to the
Stynes et al. economic impact of snowmobiling study, in the Western UP, the multiplier
for number of jobs per $1 million in sales from snowmobiling is 57. This implies that

one tourism job is attributable to $17,543 in annual snowmobiling spending in the study

160

area (Stynes et al. 1998). Based on the estimated spending on snowmobiling in the study
area calculated above, the total number of jobs created from snowmobiling in the study
area in 2000 was 19 jobs (332,310/ 17,543 = 19).

The number of jobs created from deer hunting in the study area was calculated by
estimating the number of deer hunters that hunted in the study area in 2003. Based on the
Michigan Deer Harvest Survey Report, approximately 12,000 deer hunters visited the
Deer Management Units (DMUs) that fall within the study area in 2003 (F rawley 2004).
This number was then multiplied by $768, the average per trip spending by Michigan
hunters (US Dept of Interior et al. 2001). Total estimated spending atttributable to deer
hunting in the study area is $9,216,000 (12,000“ $768). Estimated expenditures in
Michigan due to deer hunting in 2001 equaled $281,774,267, and the number of jobs
attributable to these retail sales in 2001 was 5,386 (IAF WA 2002). The estimated ratio
of deer hunting expenditures to jobs is $52,300 (281,774,267/5,386). Using this number,
the total number of jobs due to deer hunting in the study area can be calculated by
dividing total deer hunting expenditures in the study area, $9,216,000, by $52,300 to
yield a total of 176 jobs attirubtable to deer hunting in the study area.

The baseline number of forest-based recreation and tourism jobs used in the

survey is 190, and the attribute was allowed to take on levels of 170, 190 and 250.

Forest Industry Jobs

The finest level of forest industry employment data available from US Census
data is at the zip code level. Therefore, the zip codes that fall within the study area
boundaries were identified. Table A2.3 lists the study area zip codes used in the

following calculation of forest industry jobs in the study area.

161

Table A2.3. Study Area Zip Codes

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zip Code Name Zip Code
Crystal Falls 49920
Michigamme 49861
Republic 49879
Champion 49814
Ishpeming 49849
Gwinn 49841
Little Lake 49833
Perronville 49873
Spalding 49866
Powers . 49874
Hermansville 49847
Foster City 49834
Vulcan 49892
Charming 4981 5
iagola 49881

 

Source: MIC 1994

Industry codes as defined by the North American Industry Classification System
(NAICS) used to calculate forest industry jobs in the study area include sector 11, the
forestry, fishing, hunting and agriculture sector, and sector 31, the manufacturing sector.
Forest industry employment was calculated by collecting employment data from

subsectors of Sector 11 and Sector 31, and these are listed in Table A2.4.

162

Table A2.4. Forest Industry Employment Subsectors Used in Calculation of Forest
Industry Jobs in the Study Area

 

Sector 11 subsectors Sector 31 subsectors

Logging Sawmills

 

Other millwork

 

Reconstituted wood product manufacturing

 

Cut stock, resawing lumber and planing

 

Wood window and door manufacturing

 

Other misc wood product manufacturing

 

Prefab wood building manufacturing

Employment in study area zip codes is reported in Table A2.5. Two of the study
area zip codes included in forest industry employment data include townships that do not
fall within study area boundaries. Because employment data at the township level is not
disaggregated into subsectors, it was not possible to use township data to calculate forest
industry jobs. However, the number of Sector 1 1 jobs in the townships that are included
in study area zip code boundaries but that clearly do not lie in the study area are
subtracted from the employment data collected at the zip code level. After making this
adjustment, the total number of jobs attributable to forest-based industries in the study

area is 677.

163

Table A2.5. Forest Industry Employment in 2001 in Study Area Zip Codes

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Zip Code Total Total Total Empl. Emplmt. Adjusted Total
Empl. Empl. in in forest- not in Forest-based
Logflg based mfg. study area Employmentl
Michigamme 61 0 0 0
Crystal Falls 764 17-36 21-53 46 17
Republic 74 26-5 1 1 —4 26.5
Champion 124 0 0 0
Ishpeming 3449 8-21 100-249 1 89 0
Gwinn 1 392 1-4 100-249 177
Charming 55 20-38 1-4 31.5
Foster City 19 8-21 0 14.5
Sagola 273 60-1 18 121-302 300.5
Vulcan 107 0 20-49 34.5
Little Lake 25 0 0 0
Perronville 0-19 0 0 0
Hermansville 324 1-4 5-9 9.5
Powers 194 0 21-53 37
Spalding 148 10-19 10-19 29
TOTAL 677

 

Source: US Census 2005c
1Total employment calculated using median values for ranges reported in each category

The basline number of forest industry jobs was set at 675 and this attribute was

allowed to take on levels of 600, 675 or 710.

164

Birds

The attribute levels chosen for the bird attributes were developed with the
collaboration of ecologists working on this project. Of the migratory forest songbird
species present in the study area, a subset was identified that are considered to be of
conservation concern. Our study area falls within the Boreal Hardwood Transition
(physiographic region 20), as defined by the organization Partners in Flight. At a finer
scale, the study area falls within Region 12 of the Boreal Hardwood Transition. Scores
defined by Partners in Flight for Region 12 were used to identify which songbird species
in our study area are considered “priority breeding species for conservation.” The
criteria used by PIF to determine priority for conservation is based on numerous
variables, including the following: global relative abundance, global scores for breeding
distribution and winter distribution, global score for threats in the nonbreeding season,
threats to successful breeding, importance of Region 12 for breeding, population trend,
priority tier, percent of species’ breeding population in Region 12, and watch list status
(PIF 2005). Table A2.6. lists the migratory forest songbird species of conservation

concern identified in the study area.

165

Table A2.6. Migratory Forest Songbird Species of Conservation Conern in the
Study Area

 

Species of Conservation Concern

Black and White Warbler" Eastern Wood Pewee Rose-breasted Grosbeak

 

 

 

 

 

 

Black-billed Cuckoo Hermit Thrush“ Scarlet Tanager
Blackbumian Warbler Least Flycatcher Veery

Black-throated Blue Warbler Magnolia Warbler White-throated Sparrow
Black-throated Green Warbler Mourning Warbler Winter Wren“
Blue-headed Vireo Nashville Warbler“

Chestnut-sided Warbler“ Ovenbird"

 

* = Species that are at or above target population level

Of the species of conservation concern listed in Table X, 6 species were identified
whose populations are currently at or above their target level based on an evaluation of
PIF criteria (personal communication, Kim Hall). The total number of birds of
conservation concern that could be at their target habitat level is 19. Therefore, the
baseline level for the birds of conservation concern attribute was set at 6 and was allowed
to take on levels of 7, 12 and 17.

The second bird attribute was chosen to represent the biodiversity of songbirds in
the study area. The levels of this attribute were suggested by the ecologists working on
the project based on their estimation of current percent of the study area with high
migratory forest songbird species diversity and the maximum level they believe it could
reach with certain changes in forest management practices (personal communication, Ed
Laurent and Kim Hall). The baseline level for this attribute was defined to be 35% and

was allowed to take on the following levels: 38%, 55% and 75%.

166

Deer

Current levels of deer browse have been estimated by foresters collaborating on
this project. Data from their research shows that approximately 69% of the study area
currently has deer browse levels that are high enough to affect tree regeneration. These
researchers estimate that this level could be reduced to approximately 49% of the study
area if the appropriate forest management practices are implemented (personal
communication, Joseph LeBouton). Therefore, the baseline value for the deer attribute

was set a 69% and was allowed to take levels of 67%, 58% and 49%.

Experimental Design

The experimental design of the choice sets included in the survey resulted in five
choice sets presented to each respondent. A 3‘ orthogonal design generated 18 choice
sets (Addelman and Kempthome 1961). From these 18 sets, eleven versions of the
survey instrument were designed in which each respondent is presented with five choice
sets. Each survey respondent was presented with an introductory ‘easy’ choice that was
not part of the experimental main effects design. This choice was constructed to have no
change in the job variables and very small improvements in the ecological variables. The
remaining four choice sets in each version were chosen based on the main effects design.

Six of the eleven survey versions include choice sets where the respondent sees
the same cost across all choices and the remaining five versions include choice sets in
which the respondent sees different costs across the choices. Versions with the same cost
were created by assigning random numbers to each of the 18 possible choice sets that had
the same cost. Because the cost took on three levels, there were six choice sets for each

cost level. For example, all six choice sets in which the cost is $20 were assigned

167

random numbers and then placed in sequence. The first four sets were then assigned to
version 1, the last four to version 2. This was done for the remaining two price levels
creating four choice sets with the same price for six of the survey versions. Adding these
four choice sets to the introductory choice generated five choice sets for each of these six
versions.

The remaining five survey versions present respondents with varying prices.
These were created by taking all choice sets with the same price level, randomizing and
then sequencing them and then assigning one choice set from each price level group to
each of versions 7 through 11. This assigned three choice sets to each of versions 7
through 11. After randomizing and assigning the choice sets to these 5 versions, 3 choice
sets (one for each price level) remained. These three were randomized, sequenced and
assigned as the fifth choice to versions 7, 8 and 9. To assign a fifth choice to versions 10
and 11, all choice sets (excluding the previous three that were left over and assigned to
7,8 and 9) were randomized and sequenced, and the first two sets were assigned
respectively to versions 10 and 11. Along with the easy introductory choice, this created

five total choice sets for each version.

168

REFERENCES

Addelman, S., and O. Kempthome. 1961. Orthogonal Main-Effect Plans. Ames, IA: Iowa
State University, Aeronautical Research Laboratory, Office of Aerospace
Research, United States Air Force.

Frawley, B.J. 2004. Michigan deer harvest survey report - 2003 seasons, Wildlife Report
No. 3418: Michigan Department of Natural Resources.

International Association of Fish and Wildlife Agencies (IAFWA). 2002. Economic
Importance of Hunting in America. Washington, D.C.: International Association
of Fish and Wildlife Agencies.

Michigan Information Center (MIC). 1994. Zip codes of Michigan: A geographic
directory: Michigan Information Center Department of Management and Budget.

Partners in Flight (PIF). 2005. Partners in Flight Bird Conservation Plan for the Boreal
Hardwood Transition (Physiographic area 20), Draft: American Bird
Conservancy.

Stynes, D. 2005. Michigan Tourism Spending by County, 2000 - Update 2000 [cited July
14 2005]. Available from http://www.prr.msu.edu/miteim/michtsm00.htm.

Stynes, D.J., C.M. Nelson, and J .A. Lynch. 1998. State and regional economic impacts of
snowmobiling in Michigan. East Lansing, MI: Michigan State Univeristy,
Department of Park, Recreation and Tourism Resources.

US Census. 2005c. Zip Code Business Patterns WAICS). US Census Bureau 2001 [cited
July 14 2005c]. Available from http://censtats.census.gov.

 

. 2005b. 2000 County Business Patterns (NAICS). US Census Bureau 2005 [cited
July 14 2005b]. Available from http://censtats.gov.

 

. 2005a. General Demographic Characteristics, 2000 Census Demographic
Profile. Northwest Michigan Council of Governments 2005 [cited July 14 2005a].
Available from www.nwmcog.org.

US. Dept of the Interior, Fish and Wildlife Service and U.S. Dept of Commerce, U.S.

Census Bureau. 2001. 2001 National Survey of Fishing, Hunting and Wildlife-
Associated Recreation - Michigan.

169

APPENDIX 3:

SURVEY INSTRUMENT

170

Survey printed on 8 1/2 x 11 paper

 

 

Forests, Wildlife and People in Michigan ’s
Upper Peninsula

A Survey of Your Opinions

 

Forests provide many benefits to people, including wood products, wildlife
habitat, recreation, water filtration, clean air, beauty. and erosion control.
All of these things are affected by forest management decisions.
Your opinions matter!

By completing this survey, you are helping forest land managers get the
information they need in order to design forest policies that better reflect the
views and concerns of all Michigan citizens.

MICHIGAN STATE
0 N l v E R s I T Y

 

 

I71

 

 

 

 

The Western Upper Peninsula Study Fore-t

Thissurveyasksyour ophrlonsabout forestsmdwildlifeinmarea ofMlchlgan‘s
UpperPenhmdatlatunMflrefertoasfireWutemUP.$mdyForm.1hhmk
mpresentedbythedudedpanofflremapimndmfirbpageandonmebadtmof
thissurvey. PleasereiertothismaerenMngquaflonsthliheunvey.

 

l'iavewuevervlsltedtheWestemUPfimdyf-‘orest? Pleaserefertotheshoded
areaofthemaploundonthispogeoronlhebackcwerofthlswruey.

0Y6
0N0
DNotsure

 

 

 

 

 

 

 

 

 

 

 

I72

 

 

Your Opinions on the Western U.P. Study Forest

 

 

This is a heavily forested area where forest policies have significant effects. N

Statefmestpolkymakersneedtokrmyouropinionsondconcemsinaderto
designforsdpoliclesforthisarea.

The Western U.P. Study Forest is economically important. and forest management
canalfedforedinduhyiobsaswalasfaed-buednaeaflonmdtmuhnlobs.

The Western U.P. Study Forest is also important for wildlife habitat, and forest
managemntcanafiect habitatforsongbirdsanddeerin theme.

l-thirnpadtimberharvstinghasbeenpracmdhthisareafamanyyears.
affecting timber production, recreation. tourism and wildlife habitat.

The information in this survey is backed by scientific research being done in the
Western U.P. Study Forest.

 

J

Please provide your opinion by indicating your agreement or disagreement with the
lollowing statements. Mark one response for each statement.

In my opinion, the Western U.P. Study Forest
should be managed to:

 

Meet the needs of people.

 

Meet the needs of wildlife.

 

Meet the needs of communities that are economically
dependent on forests. no matter what effect this has on
the environment.

 

Balance environmental needs with the needs of
communities that are economically dependent on forests.

 

 

 

Meet the needs of future generations.

 

 

 

 

 

173

 

 

 

The Economy

 

Foresthducts

WWestemUPfitudyForestisanimportantsourceolwoodproducu.

Jobsinforestry.logglng,andwoodpmductandpapermanulacnuingarean
importantsourceofernplcymentinthisma.

1.12001 therewereabout675iorestlndustryiobsintl'reWestern U.P.StudyForest.
Recreationendtourisar

Many people visit the Western U.P. Study Forest to puticipate in deer hunting,
snowmobilingand fall colortours.

0 ReaeafionaladiviuesinmeWedernU.P.SmdyForastauradvisitorswhospend
moneylntheaeaandsupportthelocaleconomy. '

ln2m1maewereaboutl90ionst-basedreaeauonandtourismjobsinthe
WesternU.P.StudyForest.

 

J

3. Pkasepmvideyowopmlonbymdicafingyuuageementmdbagreanentmthe
following statements. Markone response for each smternent.

    

 

   

 

 

 

 

 

 

 

 

lamyoplnlon.theWestesn U.P.StudyForest g; g 1.3" 8 ‘UE
should be managed to: 5m . g 5 Q; "m 5
a. Maintain forest indusz jobs. it: 0 D5 0 ”U h
b. Maintain forest-based recreation jobs. _ Q _ 0 it" Cl ‘0
c. Protectforestandwildlife resources. ..D" 0 £011 0 «5.13;
d. Adrieveabalancebetweenmaintainingforw-related i 5‘ 5:5»; ‘ f2,
jobsandprotectingforestandwildlite resources. Offl D 501] O in;

 

 

 

 

 

4. Doyouthinlrthatachangeinthenumberofforest-basedjobsintheWestemUP.
StudyForestcouldaffectyourowninoome?

0 Yes
D No
D Notsure

 

174

 

 

 

Wildlife in the Western U.P. Study Forest

Migratory Forest Songbirds \

Migratory forest songbirds breed and nest in the northern U.S. and Canada and
have winter homes in southern area.

ScientificresearchhasshownthattheWestern U.P. StudyForestprovides
Important habitat forthese typesofblrds.

Onewaytomeasureblrd diversityisbycounting the numberofdifferentbirdsa
personcanhearby standif.nganwherelntheorest

FortheWestern U.P.StudyForest, "high" migratoryforestsongbird m
meansbeingabletohearlOormorebirdswhflestandinganyMiereinflrefaed.

Currently. 35% of the Western U.P. Study Forest has high migratory forest
songbird diversity.

With changes in forest management, up to 90% of the Western U.P. Study Forest
could have high migratory forest songbird diversity. J
Were you aware of the importance of the Western U.P. Study Forest for migatory
forest songbird habitat?

Ci Yes D No

 

How concerned are you about migratory forest songbird diversity in the Western U.P.
Study Forest?

0 Not at all concerned

Please provide your opinion by indicating your agreement or disagreement with the
following statements Mark one response for each statemen

In my opinion, the Western U.P. Study Forest
should be managed to:

 

Inaease migratory forest songbird diversity even if there
are economic losses to forest-based industries.

 

 

 

Achieve a balance between sustaining forest-based
industries and migratory forest songbird diversity.

 

 

 

 

175

 

 

 

 

 

 

10.

MW Forest Songbird- 0192mm W

Of the 65 migratory forest songbird species in the Western U.P. Study Forest,
scientists have identified 19 as species of conservation concern due to factors
like the importance of habitat, threats to breeding success and evidence of
population declines

“renumberolbhdsnecessaryforeadispedestothrivelswnendyabovethe
scientificallydefinedtargetlevel for60fthesel9speciesofconsewation
concern.

Sdenfisudrlnkfliatmereasonswhysomeoftheseforestsongbhdspedssue
below the target population level are related to forest management
practices in the Western U.P. Study Forest

Forest management practices can be changed to Improve habitat for migratory
forest songbirds that are species of conservation concern.

 

/

WereyouawareoftheimportanceoftheWestemUP Studymeormlgratory
forestsongbird speciesofconservationconcem?

DYes 0N0

Howimportantislttoyou toprotecthabitatformigratoryforestsongbirdweciesof
conservation concern in the Western U.P. Study Forest?

Ci Very important

0 Important

U Somewhat important
U Not at all important

Please provide your opinion by indicating your agreement or disagreement with
the following statements. Mark one response for each statement.

In my opinion. the Western U.P. Study Forest
should be managed to:

Disagree

Agree

 

Protect habitat for migratory forest songbird species of
conservation concern. even if it results in economic _ ~ ‘
losses to forest-based industries. :35 _.

 

D
D

 

 

Achieve a balance between sustaining forest-based _ 52'“ ..
industries and protecting habitat for migratory forest I, . D U
songbird species of conservation concern. " 77...;

 

 

 

176

 

 

 

 

Deer

Deerieedontwigsandsmallbranchesoi’treesinthewinter. Thisactivttylscalled
deerbrowse.

Sdendficstudieshmshuundrathigbdurbrowsemybeslauingorinm
masneaiyeflnflnathgfinngeneraflmofcertahtypesotmmmthe
WesternU.P.Studyl-'orest.

Cmrendy,69%ofd1eWesternU.P.SmdyForesthasdeerbmwsekvekthatare
highenoughtod’lecttreeregeneration.

Lo'DeerBrowse HighDeerBrovae

   

Forestmanagementpractmcanbechangedtoreducetheeffectsotdeer
browseontheiorest.

Bycbanglngfliemmmmdmotfimberhawestsmkpouibleto
mdeauoundfliehndscapeneducmgdoerbmewhllemmthe
mnmrberofdeerinthearea.

 

 

 

 

 

 

 

 

 

 

 

 

ll. PleaseindicateyouragnanentmdhagreemeththdielofloMngstatanents.
Markoneresponseloreachstatement
EE E. E EE
50 < Elia}
a. lwasawareotdeerbrowsebetorereadingthis. D f C] C] D Q
h. lhaveseentheeffectsolhighdeerbrowseonlorests. u 0 u" o u
c. lamconcernedabouttheeffectsotdeerbrowseonthe TO“ U '0 Cl PU
appearanceottheWestern U.P.StudyForest. ,_ ~.
d. Forestmanagementshouldreducedeerbrowseeuenilit K ~§ E“
meamteuaecdezinfllema- :0‘“ 0 6.: 0 Ci
e. Forestmanagernentshouldmovedeeraroundthe i“
landscapetoreducedeerbrwseandmm :0} D ‘0‘; CI
deeerinthearea. ff 1..

 

 

 

 

177

 

 

 

 

 

 

 

 

Forest Management Programs

- Management decisions of private forest landowners can affect forest services that \

provide benefits to the landowner and others. and government programs can
encourage forest management practices that provide certain benefits.

' Sometimes. laws and regulations can be used to require private forest hndowners
to implement specific forest management practices.

0 Inodmmvoluntuypmgmmscanencwmgeprhmteforesthndommto
meetcertahforedmanagemntgoak;landounmarenotrequiudtoparfldpah.

- Atypeofvoluntaryprognmisfliejorestmt.whichprwidesfinandal
incenfivesfaprivatefuedbndmtomumgednbforeststomeetoertahgoak.
Voiuntoryforest easemerrts....
' Have been used in other states such as Maine and New Hampshire.
' Ensure sustained timber harvests and permanent protection of forested areas.

0 Canenstueflieadiievementofpanicularforestmanagementgoals.srxhas
hnprovhghabitatformigatoryforedsongbhdsofconservafionconcernor
reducingtheeffectsofdeerbrowseontreeregeneration.

 

12. Please indicate your agreement or disagreement with the following statements. Mark

one response for each statement.

 

 

 

 

 

 

 

 

 

a. l was already familiar with voluntary forest easements. {j ’11 -
b. i am familiar with laws or regulations for private forest E j '0 ‘ ‘.‘
landowners. . _ L t I. r;
c. I am familiar with other voluntary forest management Ci .1. ~ '4 U ‘;
programs for private forest landowners. __ .
,_, J a
d. Some forest management decisions made by private D r‘ "‘ Cl - ,
forest landowners should be regulated by laws. ‘
2. All private forest landowners should make their own D .1 ~ 7‘ Ci L *
management decisions without government programs. 3%,, " ””5 g- ,4"
f. All private forest landowners should be able to enroll in 5'»; CI fair“ C] ‘ 4
voluntary forest management assistance programs. '“ y .5 21.1.1

 

 

 

178

 

 

 

 

 

A Proposed Voluntary Forest Easement Program

Avoiuntsrytoresteasementprogamisaarenflybehgdevelopedforpflvate
forest landcvmers in theWestern U.P. Study Forest.

Private forest landowners intheWestem U.P.StudyForestthatoolantarfly enroll
receivefinancialtnceaflvutodevelopandfdbwanewfaedmanagemsntplan.

flubredmnagemmtplamdifierloreachlandmhatenrdhbutdwal
mnaibutebmeprogam'sobjedivaandmeetmeneedsdhdivldualhndownm.

Program Implementation

1hepmposedprogamwouidbehnptementedifitbapprovedbyMichiganvotenln
aspectaireferendum.

Thepmgrunwouldnotbehnplementedflitisnotapprovedbywdiiganvoters.

ProgramCost
lhevolunhryforesteasementprogameostemoneytoimplement.

Themoneytohmdtheprogamwmndbemedtozpayforesterstodevelopplam
(10%ofcosts),payadrnhbaafiwoosui5%ofcwb).andpaylandowneutoenrol
in theprogam (85% of costs).

Stateandlocaltaxeswlllbeusedtofundtheforesteasementprogram.

Choostng among Forest Easement Programs

Because there are many different forest easement programs that could be adopted
in the Western U.P. Study Forest. forestmanagersand poliqrmalrersneedtolmow
whichaspectsofforestsandwildiifeinthisaeaofMidiiganamlmpmtanttoyou.

lnflmefolbudngpagesyouwiflbepresentedwifliaseriesofpoulblepmgreau
thatshowwhatwouldchange intheWestem U.P. Study Forest ifthatvoluntary

forest easement program is implemented.
lheestimatedcosttoyomhouseholdwifldependonmeprogramfeatures.

Youwillbeaslsedtovote‘yes‘or‘no‘foreachscenario.

 

179

 

 

 

 

 

Thistabhrhmaposdblevduntaryforeueasanentprog'amfordeestemUP.
StudyForestltheareaonthebacltcover). Thetableslmwhatwwlddiangeinthat
areaiftheprogamlsirnplemented. Pleaseusethetabletomthequesflorubelow.

 
 
   
 
 
 
    
    
    

Things in the Western U.P. Study Forest that
would be affected by the voluntary forest
easement pros-m

 

 

' Numberofforestindustryjobsinthearea
' Forat~basedrocreationandtourismjobslnth¢area ’

 

0 Percent of area with high migratory forest songbird
speciesdiversity

° Numberofmigratoryforestsongbirdspeciesoi
conservation concern that are d or above their
targa population (outw of 19 possible species) .

° Peroentofareawithdeerbrowsehighenoughto
affect tree regeneration

- Numberofdeerintheareaunore, same less)

  

 

 

18a. FuminamedSZOpuwuinmhmhofl’sMwouldyouvotsforfltts
program?
DYes 0N0

13b. meareyouaboutyourmsponsstoqusstlonna?
OVerysure DSure DSornewhatsure DSomewhatunsure DVeryunsure

13c. lfyouuumd‘yes‘tolhpleasegotothenextpage.
ifyouarmed 'm’tol3awhicirorteoftlrefollorulrigbeadmilaesyotiranawer?
D lwouldvote‘yes‘ifthecostunsmudilower.

D imtghtvote'yes'ifthecostwasmuchlower.
Cl lwouldnevervoteforthlsprogammomatterwhatthecostis.

 

180

 

 

 

 

ThbtabkshounaposslblevduntaryfaedeasemntprogranforflieWedanUP.
ShtdyForestftheueaonthebadtcover). 'l‘hetableshmwhatwouldchangeinthat
miftheproglamlsinplemented. Pleaseusethetablebanswerthequestiombelow.

Things In the Western U.P. Study Forest that
would be affected by the voluntary forest
easement program

FE“ mi» - 72.; 1:;
1 . l .: .33." -,

a...

' Numberofforsstindmtryjobsinthearea
° Forest-basedrecreationandtourismiobsmthearea

 

 

 

' Percent of area with high migratory forest songbird
species diversity

0 Number of migratory forest songbird species of
conservation ooncemthataredorabouethelrtarget

 

population (out of 19 possible species)
’99.“ . fggtw"

0 Percentofareawithdeerbrmvse highenoughto
affecttreeregeneration

- Number of deer in the area (more. same. less)

 

 

14a. ForanhmaseofSZOperyearinyomhousehold‘staxes.woddyouvoteforthis
program?

DYes 0N0

14b. Howarreareyouaboutyourresponsetoquestionllla?
DVerysure DSure USomewhatsure DSomewhatunsure DVeryunsure

14c. lfyouansvered‘yes'tol4a.pleassgotothenextpage.
lfyouanswered ‘no'tol4a.whidioneofthefolmringbestdssaibesyouramer?
Dlwouldvote‘yes' ifthecodwasnnidilower.

D lmightvote‘ yes' ifthecostmmuchlower.
Dlwouldnevervoteforthisprogram. nomatterwhatthecostls

 

181

 

 

 

 

This tableshounaposdble voluntary forest easernentprogram lortheWestem U.P.
StudyForescheareaonthebackcwer). Thetableshowswhatwouldchangeinthat
arealltheprogamlslmplemented. Pleaseusethetabletoamrerthequesttombelmu.

Tblngs In the Western U.P. Study Forest that
would be aflectsd by the voluntary forest
easement program

 

.., ,1 »-.'- v _ )7;
, " ..aukgrt
- 3"..2

‘ s.

t- H“
‘fi';,:‘

 

. _-.‘,_ gm, 3 , F: _
' Numberofforut industryjobs In the area

0 Forest-based recreation and tourism jobs in the area

high migratory forest songbird

  

 

° Percent of area with
species diversity

0 Number of migratory forest songbird species of
conservation concern that are at or above their target
populatton (out of 19 posible spedes)

t. ',; r I: :{T\ ..

 

   

..x

 

- Percent of area with deer browse high enough ‘0
affect tree regeneration
- umber oldeer

° to your household in increased annual taxes

   

In the area (more, same, less)

 
 

 

 

 

 

15s. ForanincreaseolSZOperyearinyourhwsehold'staxes,wouldyouvotelorthls
program?

DYes 0N0

15b. l-lowsureareyouaboutyourrewonsetoquestion 15a?
DVerym DSure USomewhatsure USomewhatunsure OVeryunsure
15c. llyouama'edyes'tolfipleasegotothenextpage.
llyouanswered ‘no'tolSa,whldron¢otthelollowlngbestducrlbesyourm?
Dlumldvote‘yes‘llthecostmmuchlorm.

D lmightvote‘yes‘ilthecostwasnmchlower.
D lwouldnevervotelorthlsprogrammomatterwhatthecostis.

 

182

 

 

 

 

mbhbhdrmapossibkvohmhwforedeasemntprogamforflreWestanUP.
StudyFmedlmeamaonhbadrm).1htaHeshmwhatwwlddrangehmat
areaifthsprogarnisirnplernsntsd. Plsaseussthstablstoarmthequsstionsbelaw.

Things in the Western U.P. Study Forest that
would be affected by the voluntary forest
easement program

 

[Economy
' Number of forest industryjobs in the area
0 Forest-basedreaeationandtourismiobsinthearea
rmsmbua- "1 . [3, a"
0 Percent of area with high migratory forest songbird
species diversity
' Number of migratory forest songbird species of

conservation concern that are at or above that target
population (out of 19 possible species)

 

' Percent ofareawith deerbrowsehigh enough to
affect tree regeneration

° Number of deer in the area (more same. lea)
m cost per year _ 23;; ‘ 'fl
Cost to your household' In increased annual taxes

 

 

 

16a. ForanInaeaseof's20peryearinyourhwsehold'staxes.wouldyouwteforflrb
program?

DYes 0N0

16b. i-lowmarsyouabwtyourresponssioqusstionléa?
DVerysure DSure DSomewhatsure DSomewhatunsure DVeryunsure
16c. lfyouanmedya'blfipleasegotothenextpage.
lfyouammd‘m'mlfiwhldrmddnfoloudngbeddeulbayourm?

Cl lwouidvote‘yes'iitheoostwasmudilmuer.
O Imightvote‘yes'ifthecostwasnmchlower.
D lwouidnevervoteforthisprogrammomatterwhatthecostis.

 

183

 

 

 

 

 

Thistabledrausapossibhvdunuryforedeasamntprog'amfordieWedemUP.
SiudyForestltheareaonthebadtcover). 'i'hetableshowswhatwouldchangeinthat
areaiftheprograrnisirnplemented. Pleaseussthetabletoamwerthequssflombslav.

 
 
 
   
 
 
  
  
 
 
 
  

Things in the Western U.P. Study Forest that
would be affected by the voluntary forest
easement program

 

' Numberofforestindustlyiobsinthearea
0 Forest-basedreaeationandtourismjobsinthearea
WWW

- Percent of area with high migratory forest songbird
species diversity

' Number of migratory forest songbird speciesof
conservationcomemthatareatorabooethelr
mapopulatton (outof 19 possible species)

. w Y

0 Percentolareawithdeerbrowsehigh enoudrto

 

 

aliecttree regeneration
' Numberofdeerinthearea(more sarne.less)
:4, or?“

Cost to your household' in increased annual taxes

 

17a. Foraninaeaseof$20peryearinyomhouehold'staxes.wouldyouvotelorthis
program?

CIYes CINo

17b. stureareyouaboutyomresponsetoquestionfla?
DVerysure DSure DSomewhatsure DSomewhatunsure CIVeryunwre

17c. ifyouansvered ‘yes’tol7a.pleasegotothenextpage.
lfyouanswered ‘no'tol7a.whichoneofthefollowingbestdescribesyouramuer?
D lwouldvote‘yes'iitheoostwasmudrluwer.

D lmightvote‘yes'lftheoostwasmudiiower.
Dlwouldnevervoisforthisprogrammomatterwhatthscostis.

 

184

 

 

 

 

18.

19.

Whenyoumadeymwdroicesofwhetherornottovotefortheforesteasement
prograrnsin Questiors 13 through 17, how importantwere the following attributes in
yourdecisions?

Number of forest industry jobs in the area

Number of forest—based recreation and tourism
in the area

Percent of area with high migratory forest songbird

Number of migratory forest songbird species of
conservation concern at or above their target
of

Percent of area with deer browse high enough to
affect tree

Number of deer in the area (more. same. less)
Cost to your household in increased annual taxes

Cl 0
D D

 

Pleaseusefltespacebdowtofistafewmasmsfaymudedsiomtovoteeifiier
'yes'or‘no'fortheforestemmentprogamsinQuestions 13ihrough l7.

 

 

 

 

 

 

 

 

 

185

 

 

 

 

How you use the Wutem U.P. Study Forest
Pbmerelawdiemapmmebxkoldmmnnywhenmngmefdmmgqum

20. Pleaseindieateflyouhmeverpmficipatedinanyofttnfoflwhgmeafioml
actMtiesinthelollmwinglocations.

   

l have participated In:

Benypiddng

  
   
 

 

 
 

Hunting (Deer)
Hunting (Game birds)
Hunting (Other, pleme specify

 
   
   
 
 
  

  
  

Mushroom gathering
. Other wildlife viewing
Sightseeing/Touring
Snowmobiling
Snowshoeing
Trail horseback riding
Other activities (Please specify

  
  
   
 

  

DDDDDDDDODDDDDDDD
UDDDDDUDDDDDDDDDDD

      

 

186

 

 

 

_

 

Your Opinions

. Wemuldliketolumhowywieeiabwtlorestsandtheenvironment. Pleasemark

oneresponseforeachitem.

ornotl . Forest.
forestisthereis tome.

in management.
to participate in community
its management
local communities.
my input matters to state

1 think that state land is
. private managed.
. 1 think private

to
certified wood
resources ensure
make the earth unlivablei
Humans are abusing the environment.

to modify the natural
environment to suit their needs.

. Humans were meant to nile over nature.

. will leam enough nature
works to be able to control it

nature easily upset.
“environmental crisis" has been greatly
are to number people
this earth can
humans with nature. it

D
0
Cl
0
0
Cl
D
0
CI
Cl
D
O
. D
D
D
0
Cl
C)
D

UDDUDDDDDDDDDDDDOUD

 

 

187

 

 

 

 

 

QuectionoAboutYou
Yarronswersbelowhelpusenswethatowinfanndion repraentsMichigan’sgener-nl
populdion. Pleaseremanbathdallammyouprouidewiflbehelddfldlyconfidenfld.

22. Doyouownpropertywlthtatheohadodanaofthomaponthebadrpage?

DYes
D No ==sp Pleasegotoquesuon25.

23. Howmanyacresdoyouown?
24. Wtuetheprimarymesofyqirpmpaty? Pleasedieckallthdapply.

DPrimayresidence ClRecreationaluse
DSecondhomeorcabin Ol-luntingforyourselforfamily
OForestproducts DHuntinglorothers
Olgiculture DOther

 

25. Areyouamemberofanytorestryorganizatiom? OYes 0N0
26. Areyouamemberolanyhuntingclubsororganhatiom? DY” 0N0
27. Areywanmnberofanyemiromnentaiorganizations? 0Y5 0N0

28. Whatisthehlghestlevelofeducationthatyouhavecompleted? Pleaudieckme.
DSomehighschool DSomecollege
DHighschoolg’aduateor 0Associatesdegree(2year)

equivalent DCoIegegraduateMyeardegee)
DTradeorvocationalschool OGraduateorprofeaslonaldegee

29. lnwhatyearwereyouborn? 19

30. Whatisyourgender? Cl Male 0 Female

31. WhatwasyourgrosshouseholdincomeinZOOll? dollars

32. lsanyoneinyourhouseholdempbyeddhecdyinanyoffiiefdangindusuies?
Pleasedieckallthatapply.

O Forestry 0 Hotels or otheraccommodation services
CILogging DOutdoorrecreation

D Woodproductmanufacturing D Naturalresourcemanagementagency
DTourism UNoneoftheabove

33. Howmanyyearshaveyoulivodin Michigan? years

 

188

 

 

 

34.

35.

36.

37.

38.

Howwouldyoudescribetheareainwhichyoulive?
Ll Urban OSuburban Cl Rural

Frornwhatcountriesorpartsoltheworlddidyouranwstorscome?

 

What is your ethnic backg'ound?

0 Hispanic. LatinoorSpanish origin 0 HawaiianorPadfic Islander
DWhite CIAsian
DAfricanAmericanorBlack DOther

 

D NativeAmericanorAlaskanative

What is your religious preference?

OProtestant OMuslim

DCatholic DNone

DJewish DOther
Howwouldyoudescribeyourpoliticalviews? ,

DExtremelyliberal OSomewhatcorusrvative

DSornewhatltberal DEadremelyconservative

DModerate.rniddleoftheroad

Thank You! This completes the questionnaire.

Weappreciateyourhelpandfeedbadc

lfyouwouldlilretoshareyourideasoropinlonswithus.pleaseusetherematnderofthis

page,orleelfreetoattachadditionalsheetstothissurvey.

 

189

 

 

 

 

 

 

Map of Michigan’s
Western Upper Peninsula
Study Forest

 

 

 

 

 

 

 

 

 

 

 

 

Pleasepbcsyourcanpldedquadonndnindnencbsedenvebpeandrdrmlto:

UP Forest 3: Wlldllfs Project. Attention: Lalla A Racevskis, Department of Agrl
Economics, Michigan State University, East Lansing. MI 48824-1039

 

190

 

 

APPENDIX 4:

SURVEY CORRESPONDENCE

191

 

-‘ “I

l) Prenoticc Letter - Printed on 8 1/2 x 11 watermark paper and hand signed

 

MICHIGAN STATE
U N l V E R S I T Y
Name Surname

Address

City. State Zip

Dear Name Surname:

You have been selected to participate in a study of forests and wildlife in
Michigan’s Upper Peninsula. The study is part of research at Michigan State
University. "lire project will provide needed information to local, regional, and
state agencies about residents’ opinions and concerns about forests and wildlife in
the Upper Peninsula.

All that we ask is that you complete a brief survey booklet that you will receive in
the mail in about a week. We are writing to you now since many people like to
receive advance notice of the survey booklet.

5 Thank you very much.

Sincerely.

seem 0?
AGRICULTURAL
ECONOMICS Laila Racevskis

Project Coordinator

WWW
Elam.“
40824-1030

aims-ms
FAX' 5173324“

Histamine-sum

1430an
“mm

192

2) Wave 1 Cover Letter - Printed on 8 1/2 x 11 watermark paper and hand signed

ospranem or
AGRICULTURAL
ECONOMICS

museum
moonlit-l

East we. is
48824-1039

517mm
FAX: 517M324“

mot “twain

“30th
“who“.

MICHIGANSTAIE
UNIVERSITY

Name Surname
Address
City. State Zip

Dear Name Surname:

You have been selected to participate in a study about forest and wildlife resources
in Michigan‘s Upper Peninsula. You may recall receiving a letter about this study
about a week ago. The study is part of an effort by Michigan State University to
Ieamaboutcitizens’ opinionsandconcernsregardingthcmanagernentofforests
and wildlife in an area of Michigan’s Upper Peninsula.

Your input is important because managing Michigan’s forest and wildlife resources
involves trade-offs that affect you. Results of the questionnaire will provide needed
guidance to local, regional, and state agencies about rcsidents’ opinions and
concerns about forested areas of Michigan’s Upper Peninsula

You have been selected as part of a scientific sample of Michigan residents. That is
why the survey asks a few questions about you and your household-so we can
make sure that we get a scientific cross-section of Michigan residents. Your
participation is vital to make sure that the information collected represents

everyone.

We realize that it takes time out of your day to fill out this survey and have enclosed
three first class stamps as a way of saying thank you for your help.

By completing and returning this survey, you indicate your voluntary consent to
participate in this study and have your answers included in the project data set. We
will keep your individual views entircIy confidential. Rest assured, your privacy
will be protected to the maximum extent allowable by law.

If you have any questions or comments about this study feel free to call me by
phone at: (517) 353-7898, fax at: (517) 432-1800, or e-mail at: racevskl@msu.edu.
If you have questions concerning your rights as a survey participant, please contact
Dr. Peter Vasilcnko. Chair of the MSU Committee on Research Involving Human
Subjects. by phone at (S I 7) 355-2180. fax at: ($17) 432-4503, or e-mail at:
ucrihs@msu.edu.

Thanks for participating in this study.

Sincerely,

Laila Racevskis
Project Coordinator

193

 

3) Reminder Post Card - Printed on 3 1/2 x 5 white card stock and hand signed

 

Dear Sir or Madam:

We recently sent you a booklet and request to participate in a study of '
Forests, People and Wildlife in Michigan’s Upper Peninsula. If you
returned the booklet, thank you very much. If you have not yet
completed the booklet, please take some time to do so now. Your input
is important to make sure that policy decisions reflect the views of
Michigan citizens. Thank you very much.

 

 

Sincerely,

Laila Racevskis

Project Coordinator

Michigan State University

Mascaras?

 

 

 

194

 

 

4) Wave 2 Cover Letter, Printed on 8 1/2 x 11 watermark paper and hand signed

oemmemor
AGRICULTURAL
ECONOMICS

WWW
12560de

East Lanna. ts
tenuous

5111353408

Fm 517K324”

Mist Minoan

man-W
“mm.

MICHIGAN STAIE

UNIVERSITY

Name Surname
Address
City, State Zip

Dear Name Surname:

We recently sent you a survey about forest and wildlife resources in Michigan’s
Upper Peninsula. Although we have received completed surveys from many
Michigan residents, to date we have not heard fi'om you.

lamwritingtoyouagainbecauseyourinputisvitall Youarepartofasrnall
sample of citizens that was selected to represent Michigan. We need you to
participate in this study even if you do not reside in or are not familiar with
the Upper Peninsula, because we need to represent the views of all Michigan
citizens.

Your input is important because managing Michigan’s forest and wildlife resources
involves trade-offs that affect you. Results of the questionnaire will provide needed
guidance to local, regional, and state agencies about residents’ opinions and
concerns about forested areas of Michigan’s Upper Peninsula.

Please take a few minutes to share your viewpoint by filling out this survey. By
completing and returning this survey, you indicate your voluntary consent to
participate in this study and have your answers included in the project data set. We
remind you that we will keep your individual views entirely confidential. Rest
assured, your privacy will be protected to the maximum extent allowable by law.

If you have any questions or comments about this study feel free to call me by
phone at: (517) 353-7898, fax at: (517) 432-1800, or e-mail at: racevskl@msu.edu.
If you have questions concerning your rights as a survey participant, please contact
Dr. Peter Vasilenko, Chair of the MSU Committee on Research Involving Human
Subjects, by phone at (517) 355-2180, fax at: (517) 432-4503, or e-mail at:
ucrihs@msu.edu.

Thanks for participating in this study.

Sincerely,

Laila Racevskis
Project Coordinator

195

 

APPENDIX 5:

 

SURVEY IMPLEMENTATION

196

 

This appendix describes the details of survey implementation. All survey
correspondence as well as the survey instrument can be found in Appendices 3 and 4 .

The first contact, the prenotice letter, was mailed to 2,000 Michigan households
via first class mail on Friday, April 8, 2005. The letter was printed on white watermark
paper with MSU letterhead, personalized for each respondent, hand signed and mailed in
#10 white watermark envelopes.

The second contact, the first survey mailing, was sent on Friday, April 15, 2005.
It was mailed in a white 10x13 envelope and included a personalized, hand signed cover
letter printed on white watermark paper with MSU letterhead with three first class stamps
affixed with a paper clip, a postage-paid white 9x10 business reply envelope and the
questionnaire.

Due to an error in the mailing process, a private company to whom the mass
mailing was contracted by the university’s mail processing system mailed the survey
using Pre-sort Standard mail rather than First Class mail, as had been requested. This
meant that the survey took longer to reach some areas of Michigan than it would have
had it been sent via First Class mail, and packages mailed to bad addresses were not
returned to sender.

The third contact, a personalized, hand signed reminder post card, was mailed via
first class mail on Friday, April 22, 2005. The fourth contact, a second wave mailing of
the survey, was sent to 1,057 Michigan households on Friday, June 3", 2005. This
mailing occurred 7 weeks after the initial prenotice mailing, which is longer than the time
interval between first and second mailings recommended by Dillman (2000). However,
because of the timing delay in the first mailing, returned surveys continued to come in for

several weeks, and therefore the second mailing was delayed until first wave returns

197

err

 

 

 

decreased substantially.

The second mailing consisted of a personalized, hand signed, region-specific
cover letter, a postage-paid business reply envelope, and the questionnaire. After the first
survey mailing, several people indicated that they did not believe they were qualified to
complete the survey because of a lack of familiarity with the region under study.
Therefore, the cover letters of the second mailing were revised to encourage those who
are not residents of the study area to respond to the survey. The letters were tailored to
residents of the Upper Peninsula who do not reside in the study area and those who are
not Upper Peninsula residents.

Codes from the American Association for Public Opinion Research (AAPOR) are
used to classify survey responses and to calculate survey response rates (AAPOR 2004).
These codes are provided in detail for all waves and regions in Table 3D.3. The first
mailing resulted in 90 bad addresses, 54 non-respondents, and 824 returned, usable
surveys, for a total first wave response rate of 43% (ibid). The second mailing resulted in
11 bad addresses, 15 non-respondents, and 130 returned, usable surveys, for a total
second wave response rate of 12% (ibid). In the first and second mailings, there were a
total of 101 bad addresses, 69 non-respondents, and 954 returned, usable surveys, for an
overall response rate of 50% for the entire survey (ibid).

Each of the eleven versions of the survey were randomly assigned to survey
recipients in each region. Versions 1 through 6 were each sent to 46 households in each

region, and versions 7 through 11 were each sent to 45 households in each region.

198

 

Table A5.1. Disposition of Survey Mailings

 

Completed

 

 

 

Number Non- Non- Response Cumulative

Contact Mailed Returned responses’ deliverablesJ Rate‘ Responses
Surveys

”it. 2,000 824 54 9o 43% NA
mailing
S°°.°.“d 1,057 130 15 11 12% 50%
mailing
Total 3,057 954 69 101 NA 50%

 

lAAPOR Disposition Code I : Returned Questionnaire

2AAPOR Disposition Code 2: Eligible "Non-Interview"; Includes individuals who
indicate that they are physically unable to complete the survey, individuals who have
died and returned surveys that are blank or too incomplete to process

3AAPOR Disposition Code 3: Unknown eligibility, "Non-interview ",' Includes incorrect
addresses, individuals who are no longer at the address

‘ Response Rate = Completed returned surveys/(N - Non-deliverables)

’ Cumulative response = Completed returned surveys/(2000 - Non-deliverables)

Table A5.2. Regional Response Rates

 

 

Study Area (SA) Rest of Upper Northern Lower Southern Lower
Peninsula (RUP) Peninsula (N LP) Peninsula (SLP)
62% 51.9% 45.6% 39.6%

 

199

 

Table A5.3. Detailed AAPOR Disposition Codes for Each Survey Mailing

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

axon Code Definition First Mailing 13%;:
1.1 Returned questionnaire - complete 824 130
2.1 Refusal and break-off 4 3
2.11 Refusal 15 4
2.12 Break-off questionnaire too 3 0
incomplete to process
2.26 Other notification that respondent l 0
was unavailable during field period
2.31 Death 17 4
2.32 Physically or mentally 14 4
unable/incompetent
3.23 Refused by addressee 0 2
3.2520 Insufficient address 5 2
3.253 No mail receptacle 0 l
3.31 Cannot be delivered as addressed 14 3
3.311 Attempted - Addressee not known 14 3
3.313 1 No such number 1 0
3.3134 Vacant l 1
3.314 Not delivered as addressed 6 0
3.3141 Unable to forward 38 O
3.32 Moved, left no address 10 l
3.9 Other 1 0

 

200

 

 

REFERENCES

AAPOR. 2004. Standard Definitions: Final Dispositions of Case Codes and Outcome
Rates for Surveys. Edited by American Association for Public Opinion Research.
3rd ed. Lenexa, KS.

Dillman, D.A. 2000. Mail and Internet Surveys: The Tailored Design Method, 2nd
edition. 2nd ed. New York: John Wiley and Sons, Inc.

201

 

 

 

 

APPENDIX 6:

SURVEY RESULTS - DESCRIPTIVE STATISTICS

202

......M_ k

Table A6.1. Socioeconomic Characteristics of Survey Respondents

 

 

 

 

Characteristic Survey MichiEL
N 953 9,938,444 '
Own property in study area 18%

Average # of acres owned 39

 

Primary property uses of property owned in the

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

study area:
Primary residence 79%
Second home or cabin 9%
Forest products 1%
Agriculture 0
Recreational use i 0
Hunting for yourself or family 8%
Hunting for others 0
Other 3%
Member of forestry organization 3%
Member of hunting club 20%
Member of environmental organization 7%
Gender
% Male 77% 49% ‘
% Female 23% 51% ‘
Median Age 55 35.5 '
Education
% Some high school 5% 11.9% '
% High school or equiv 26% 31.3% '
% Trade or vocational school 9% N/A
% Some college 21% 23.3% '
% Associate’s degree 10% 7% '

 

203

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Characteristic Survey Michi&_
% College graduate 17% 13.7% ‘
% Graduate or professional degree 12% 8.1% "
Median household reported income $50,000 $44,667 ‘
% employed in resource-based industry 13%
Average # years lived in Michigan 47
% residing in:
Urban area 9% 75% ”
Suburban area 20% N/A
Rural area 71% 25% ”
Ethnic background:
% Hispanic, Latino or Spanish origin 0.7% 3.3% "
% White 94.5% 80.2% ‘
% African American or Black 1% 14.2% '
% Native American or Alaska native 1.9% 0.6% '
% Hawaiian or Pacific Islander 0.2% <l% '
% Asian 0.4% 1.8% '
% Other 1.3% 1.3% "
Religious preference:
% Protestant 39% 53.7% °
% Catholic 36.1% 29.5% °
% Jewish 0.8% 0.9% c
% Muslim 0.2% 0.4% °
%None 11.3% 11.1%“
% Other 12.6% 4.4% °
Political views:
% Extremely liberal 2% 12.6%°
% Somewhat liberal 15% 12.6% °

 

204

 

 

 

Characteristic Survey Michigan

 

 

% Moderate 45% 40% °
% Somewhat conservative 32% 30.9% °
% Extremely conservative 6% 2.6% °

 

‘ tip://guickfacts.censt§.gov/gfdl, Profile of Demographic and Social Characteristics 2000;
Accessed 6/27/05

" http://www.michngng/ggcuments/urbg rural 42192 ijdf, Census 2000 Urban and Rural
Population for Michigan, Counties and County Subdivisions; Accessed 9/02/05
6

. . anda. o / dalit .' ‘7' = General Social Survey
2002, Quick Analysis for Survey Results, Accessed 9/02/05; Reported statistics represent
East North Central Region (WI, IL, IN, MI, OH)

205

 

Table A6.2. Response Frequencies for Forest Management Environmental Attitude
Statements

 

Survey Statement (Question number) Percent Response N

 

In my opinion, the Western U.P. Study

Forest should be managed to: SA A N D SD

 

Meet the needs of people. (2a) 19.5 49.9 19.5 9 2.2 915

 

Meet the needs of wildlife. (2b) 38.5 49 9.5 2.5 0.5 920

 

Meet the needs of communities that are
economically dependent on forests, no
matter what effects this has on the
environment. (2c)

4.8 13.4 15.8 41.2 24.8 916

 

Balance environmental needs with the needs
of communities that are economically 31.5 50.1 13.3 4.4 0.8 931
dependent on forests. (2d)

 

Meet the needs of future generations. (2e) 40.1 47.1 9.8 2.5 0.5 923

 

Maintain forest industry jobs. (30) 18 52.1 20.8 7 2 884

 

Maintain forest-based recreationjobs. (3b) 13.2 58.5 21.2 5.9 1.2 887

 

Protect forest and wildlife resources. (3c) 39.7 51.1 7.1 1.3 0.8 891

 

Achieve a balance between maintaining
forest-related jobs and protecting forest and 46 43.9 8.4 1.1 0.6 906
wildlife resources. (3d)

 

Increase migratory forest songbird diversity
even if there are economic losses to forest- 6.8 15.4 30.3 35.5 11.9 881
based industries. (7a)

 

Achieve a balance between sustaining

forest-based industries and migratory forest 26.5 48.8 15.7 6.3 2.7 905
songbird diversity. (7b)

 

Protect habitat for migratory forest songbird
species of conservation concern, even if it
results in economic losses to forest-based
industries. (10a)

7.7 20.1 25.6 34.7 11.9 896

 

Achieve a balance between sustaining
forest-based industries and protecting
habitat for migratory forest songbird species
of conservation concern. (10b)

28.1 49.2 14 6.7 2 908

 

206

 

 

Survey Statement (Question number)

Percent Response

 

1 was aware of deer browse before reading
this. (11a)

22.5

48.6

9.9

13.4

5.5

901

 

I have seen the effects of high deer browse
on forests. (11b)

17

42

18.6

17.1

5.3

902

 

I am concerned about the effects of deer
browse on the appearance of the Western
U.P. Study Forest. (11c)

11.5

42.2

30.5

12.9

899

 

Forest management should reduce deer
browse even if it means tewer deer in the
area. (11d)

9.5

24.4

19.6

33.8

12.7

907

 

Forest management should move deer
around the landscape to reduce deer browse

and key the same number of deer in the
area. (I I e)

20

41.6

22

12.1

4.3

909

 

I was already familiar with voluntary forest
easements. (12a)

3.8

20.4

31.2

35.4

9.1

886

 

I am familiar with laws or regulations for
private forest landowners. (12b)

3.1

18.6

32.4

35.7

10.2

885

 

I am familiar with other voluntary forest
management programs for private forest
landowners. (12c)

2.8

21.7

30.5

35.6

9.4

883

 

Some forest management decisions made by
private forest landowners should be
regulated by laws. (12d)

34.4

21.3

24

12.4

889

 

All private forest landowners should make
their own management decisions without
government programs. (12c)

14.1

24.4

22.8

32.7

5.9

893

 

All private forest landowners should be able
to enroll in voluntary forest management
assistance programs. (12f)

27

57.1

12.3

0.6

896

 

Whether or not I visit the Western U.P.
Study Forest, just knowing that the forest is
there is important to me. (21a)

38

39.1

13.8

6.2

2.9

894

 

I have participated in community forest
management. (21b)

207

5.4

13.9

39.7

26.7

14.3

875

 

 

 

Survey Statement (Question number) Percent Response N

 

I would like to participate in community

forest managfiement. (210) 6.5 16.7 49.8 19.2 7.8 872

 

I think the state communicates its forest
management goals clearly to local 2.9 17.5 42.2 27.7 9.7 893
communities. (21d)

 

I think my input matters to the decisions of
state forest managers. (21c) 6 30.8 38.1 17.4 7.6 895

 

I think that state forest land is well

managed. (21]) 1.7 26.6 51.8 15.1 4.8 898

 

I think industrial private forest land is well
managed. (218) 1.8 21.4 55 17.2 4.6 889

 

I think non-industrial private forest land is

well managed. (21h) 1.8 19.8 56.9 17.2 4.3 888

 

I have purchased environmentally certified
wood products. (2117 4.1 21.8 45.5 18.7 9.9 861

 

I would like to be able to purchase ‘
environmentally certified wood products. 9.3 34.6 41.9 9.1 5.2 872

 

 

(211')

Human skill and resources will ensure that

we do not make the earth unlivable. (21k) 20'7 46'1 ”'3 l 1'2 4'6 889
Humans are severely abusing the 19 35.2 21.6 1.7.] 7.1 893

environment. (211)

 

Humans have the right to modify the natural

environment to suit their needs. (21m) 3'4 16 25:5 36-4 18-3 394

 

 

Humans were meant to rule over nature. 43 15.2 21.8 35.3 23.4 89 4
(21n)

Humans will eventually learn enough about

how nature works to be able to control it. 7 .7 26.7 26.2 27.4 12 888
(210)

 

The balance of nature is delicate and easily
upset. (21p)

The so-called “environmental crisis” has
been greatly exaggerated. (21q)

26.9 42.1 15.9 10.9 4.3 893

7.4 20 31.4 27 14.2 886

 

We are approaching the limit to the number

ofpeople this earth can support. (21,.) 13.6 29.4 31.2 21.2 4.6 888

When humans interfere with nature, it ofien

Jiroduces disastrous consequences. (21s) 23.5 44 19:9 10:7 1-9 732

208

Table A6.3. Response Frequencies for Recreational Activities

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Recreational Activity (Question number) Percent Response N
tllrlel 3153:; In 3;“; Never in
U.P. Study “game; Michigan
Forest
Berry picking (20a) 44.5 44.5 11 836
Birdwatching (20b) 42.1 37.9 20 805
Boating (20c) 49.9 42.9 7.2 853
Camping (20d) 50.7 36.5 12.8 814
Cross-country skiing (20c) 29.7 32.8 37.5 738
Firewood gathering (20]) 46.9 38.2 14.8 803
Fishing (20g) 55.5 36.9 7.6 841
Hiking (2011) 50.2 ‘ 32 17.8 799
Hunting (Deer) (201') 43.9 29.7 26.5 789
Hunting (Game birds) (201) 36.8 27.8 35.4 709
Hunting (Other, please specrjy ) (201;) 26.4 2 4. 8 48.8 5 64
Mushroom gathering (201) 31 34.8 34.2 761
Other wildlife viewing (20m) 55 32.4 12.6 793
Sightseeing/Touring (20n) 60.5 28.9 10.5 798
Snowmobiling (200) 36.1 28.9 35 726
Snowshoeing (20p) 27.5 18.9 53.6 692
Trail horseback riding (20q) 12.9 23.4 63.7 564
Other activities (Please specify (20’) 63.2 18.8 18 345

 

209

Table A6.4. Other Response Frequencies

 

 

 

 

 

 

Survey Question (Question number) Percent Response N
Yes No
Have you ever vrsrted the Western U.P. Study 722 27.8 694
Forest? (1)
Were you aware of the importance of the Western
U.P. Study Forest for migratory forest songbird 30.2 69.8 908
habitat? (5)
Were you aware of the importance of the Western
U.P. Study Forest for migratory forest songbird 16.5 83.5 904

species of conservation concern? (8)

210

APPENDIX 7:

DESCRIPTIVE STATISTICS BY REGION

211

 

Table A7.1 Regional Response Frequencies for Forest Management Environmental
Attitude Statements (Pearson Chi-Square < 0.10 = ‘; < 0.05 = I""; < 0.001 = “*)

 

Survey Statement (Question number)

Percent Response Within Region N

 

In my opinion, the Western U.P. Study

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Forest should be managed to: SA A N D SD N

Meet the needs of people. (2a) 907
Study Forest 23 50.5 15.8 9.3 1.4 291
Rest of UP 22.6 49.8 19.2 6.3 2.1 239
Northern LP 16.6 48.2 22.6 10.1 2.5 199
Southern LP 13.5 50 23 10.1 3.4 178

Meet the needs of wildlife. (2b) 912
Study Forest 40.7 49 7.2 2.4 0.7 290
Rest of UP 34.3 52.7 10.9 1.7 0.4 239
Northern LP 42.4 45.3 8.4 3.4 0.5 203
Southern LP 36.1 48.3 12.2 2.8 0.6 180

Meet the needs of communities that are

economically dependent on forests, no 908

matter what effects this has on the

environment. (2c) "*
Study Forest 5.9 12.1 11.4 43.3 27.3 289
Rest of UP 6.3 19.7 17.6 35.7 20.6 238
Northern LP 5.4 6.9 19.8 43.6 24.3 202
Southern LP 0.6 12.8 16.2 43.6 26.8 179

Balance environmental needs with the needs

of communities that are economically 923

dependent on forests. (2d)
Study Forest 31.1 50.9 10.9 7.2 0 293
Rest of UP 33.1 49 15.1 1.7 1.3 239
Northern LP 30.3 50 14.9 3.8 1 208
Southern LP 32.8 48.6 13.1 4.4 1.1 183

 

212

 

Survey Statement (Question number)

Percent Response Within Region N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Meet the needs of future generations. (2e) 915
Study Forest 40.3 47.1 9.2 2.7 0.7 293
Rest of UP 41.3 48.3 6.3 3.3 0.8 240
Northern LP 38.6 47.5 11.9 1.5 0.5 202
Southern LP 41.7 43.3 12.8 2.2 0 180

Maintain forest industry jobs. (30)“ 877
Study Forest 21.3 51.8 18.8 7.4 0.7 282
Rest of UP 21.5 52.8 16.7 6.9 2.1 233
Northern LP 17.6 48.2 23.3 7.8 3.1 193
Southern LP 8.9 54.4 27.8 5.9 3 169

Maintain forest-based recreation jobs. (3b) 880
Study Forest 12.7 56.2 24.4 5.7 1.1 283
Rest of UP 12.9 61.6 16.4 7.3 1.7 232
Northern LP 18.5 50.8 23.6 6.2 1 195
Southern LP 8.8 65.9 20 4.1 1.2 170

Protect forest and wildlife resources. (3c) 884
Study Forest 36.6 53.9 6 2.5 1.1 284
Rest of UP 39.5 51.1 7.7 0.9 0.9 233
Northern LP 45.9 45.4 7.1 1 0.5 196
Southern LP 39.2 52 7.6 0.6 0.6 171

Achieve a balance between maintaining

forest-related jobs and protecting forest and 899

wildlife resources. (3d)

Study Forest 44.3 46.7 8 0.7 0.3 287
Rest of UP 45 45 8.8 0.8 0.4 238
Northern LP 47.5 41.4 8.6 2.5 0 198
Southern LP 46.3 43.6 8.5 1.1 1.7 176

 

213

 

 

Survey Statement (Question number)

Percent Response Within Region N

 

Increase migratory forest songbird diversity

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

even if there are economic losses to forest- 874
based industries. (7a)*"”"
Study Forest 6.7 10.6 24 42.8 15.9 283
Rest of UP 7 17.2 30 32.6 13.2 227
Northern LP 7.7 16 35.1 34.5 6.7 194
Southern LP 5.3 20 36.5 28.2 10 ‘ 170
Achieve a balance between sustaining
forest-based industries and migratory forest 898
songbird diversity. ( 7b)
Study Forest 26.3 47.4 16.6 6.9 2.8 289
Rest of UP 24.9 50.6 16.9 5.5 2.1 237
Northern LP 26.9 49.2 ' 14.7 6.1 3 197
Southern LP 29.1 47.4 13.7 6.9 2.9 175
Protect habitat for migratory forest songbird
species of conservation concern, even if it 888
results in economic losses to forest-based
industries. (10a)***
Study Forest 6.7 14.4 22.1 40 16.8 285
Rest of UP 7.5 21.2 27 30.5 13.7 226
Northern LP 8.1 22.8 28.4 33 7.6 197
Southern LP 8.9 24.4 26.1 33.3 7.2 180
Achieve a balance between sustaining
forest-based industries and protecting 900
habitat for migratory forest songbird species
of conservation concern. (10b)
Study Forest 26 50.3 14.2 6.9 2.4 288
Rest of UP 27.8 48.7 15 6 2.6 234
Northern LP 31.7 47.7 11.1 8 1.5 199
Southern LP 28.5 49.7 14.5 6.1 1.1 179

 

214

 

Survey Statement (Question number)

Percent Response Within Region N

 

1 was aware of deer browse before reading

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

this. (11a "* 893
Study Forest 31.6 54.4 5.6 7 1.4 285
Rest of UP 25.8 49.8 9 12.9 2.6 233
Northern LP 17.2 50 12.6 13.1 7.1 198
Southern LP 9 36.7 14.7 24.9 14.7 177

I have seen the effects of high deer browse 89 4

on forests. (11b)***

Study Forest 22.8 51.6 13 10.5 2.1 285
Rest of UP 19.7 42.5 18 16.7 3 233
Northern LP 16.7 40.4 19.2 19.2 4.5 198
Southern LP 4.5 27.5 ~ 28.1 25.3 14.6 178

I am concerned about the effects of deer

browse on the appearance of the Western 891

U.P. Study Forest. (11c)"

Study Forest 15.1 38 28.9 14.1 3.9 284
Rest of UP 12.6 40.9 27.4 16.5 2.6 230
Northern LP 7.1 44.4 33.8 12.1 2.5 198
Southern LP 9.5 48.6 32.4 7.3 2.2 179

Forest management should reduce deer

browse even if it means fewer deer in the 899

area. (11d)"‘

Study Forest 10.5 23 16.4 36.6 13.6 287
Rest of UP 13.3 21 17.6 34.3 13.7 233
Northern LP 5.1 25.8 25.8 30.8 12.6 198
Southern LP 8.3 27.6 21 32.6 10.5 181

 

215

 

 

Survey Statement (Question number)

Percent Response Within Region N

 

Forest management should move deer
around the landscape to reduce deer browse

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

and keep the same number of deer in the 902
area. (11e)"‘**
Study Forest 19.7 40.5 21.5 12.8 5.5 289
Rest of UP 18 45.9 16.3 14.2 5.6 233
Northern LP 17.1 39.2 26.6 13.6 3.5 199
Southern LP 27.1 40.3 24.9 6.1 1.7 181
1 was already familiar with voluntary forest 878
easements. (12a)"*
Study Forest 3.9 21.1 33 35.1 6.8 279
Rest of UP 5.3 25 29.8 31.6 8.3 228
Northern LP 4.1 20.9 30.6 36.7 7.7 196
Southern LP 1.7 12 30.9 39.4 16 175
1 am familiar with laws or regulations for 877
private forest landowners. (12b)"‘ * *
Study Forest 2.9 19.7 31.5 37.3 8.6 279
Rest of UP 3.9 23.6 34.5 30.1 7.9 229
Northern LP 3.6 19 32.3 35.4 9.7 195
Southern LP 1.7 9.8 31 40.8 16.7 174
I am familiar with other voluntary forest
management programs for private forest 875
landowners. (12c)**
Study Forest 3.2 23.5 27.4 37.5 8.3 277
Rest of UP 2.6 22.3 34.9 32.8 7.4 229
Northern LP 4.1 24.6 31.3 32.3 7.7 195
Southern LP 1.1 13.8 29.3 39.7 16.1 174

 

216

 

Survey Statement (Question number)

Percent Response Within Region N

 

Some forest management decisions made by

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

private forest landowners should be 88]
regulated by laws. (12d ***
Study Forest 5.7 28.7 18.3 26.9 20.4 279
Rest of UP 8.7 29.9 21.2 29.9 10.4 231
Northern LP 8.7 32.8 26.7 22.1 9.7 195
Southern LP 9.1 52.3 19.9 14.2 4.5 176
All private forest landowners should make
their own management decisions without 885
government programs. (12e)** *
Study Forest 19.6 29.2 19.2 26.7 5.3 281
Rest of UP 13.9 26 24.2 30.7 5.2 231
Northern LP 11.2 23.9 25.4 35 4.6 197
Southern LP 8 15.3 23.9 43.2 9.7 176
All private forest landowners should be able
to enroll in voluntary forest management 888
assistance programs. (12f)
Study Forest 24.5 57.4 13.1 3.9 1.1 282
Rest of UP 27 .7 60.2 10.4 1.7 0 231
Northern LP 31.6 54.6 10.2 3.1 0.5 196
Southern LP 25.7 55.9 14.5 3.4 0.6 179
Whether or not I visit the Western U.P.
Study Forest, just knowing that the forest is 887
there is important to me. (21a)
Study Forest 38.9 40 12.5 6.8 1.8 280
Rest of UP 36.5 41 13.1 6.8 2.7 222
Northern LP 39.7 37.7 14.7 4.4 3.4 204
Southern LP 37.6 38.7 14.4 6.1 3.3 181

 

217

 

 

Survey Statement (Question number)

Percent Response Within Region N

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 have participated in community forest 868
Mment. (21 b)*"""
Study Forest 5 13.3 42.4 28.4 10.8 278
Rest of UP 4.6 15.1 44.3 25.6 10.5 219
Northern LP 8 16.1 40.7 21.1 14.1 199
Southern LP 4.1 10.5 28.5 32 25 172
I would like to participate in community 865
forest mananement. (21c)
Study Forest 5.8 20.3 48.6 19.2 6.2 276
Rest ofUP 6.4 15.1 51.1 21 6.4 219
Northern LP 8.7 15.3 54.6 13.8 7.7 196
Southern LP 5.7 14.9 44.8 22.4 12.1 174
I think the state communicates its forest
management goals clearly to local 886
communities. (21 d)*"”"
Study Forest 1.8 14.2 36.3 34.5 13.2 281
Rest of UP 3.6 19.6 43.3 24.6 8.9 224
Northern LP 4.9 16.7 46.8 24.6 6.9 203
Southern LP 1.7 21.3 45.5 23 8.4 178
I think my input matters to the decisions of 888
state forest managers. (21c) "*
Study Forest 4.3 26.3 32.7 25.6 11 281
Rest of UP 4 37.5 37.9 13.8 6.7 224
Northern LP 8.8 28.9 40.7 14.2 7.4 204
Southern LP 8.4 32.4 43 12.8 3.4 179

 

218

 

Survey Statement (Question number)

Percent Response Within Region N

 

I think that state forest land is well
managed. (21])*"”"

891

 

Study Forest

0.7 19.8 52.3 19.8 7.4 283

 

Rest of UP

2.2 27.8 51.1 16.3 2.6 227

 

Northern LP

2.5 33.5 48.3 9.9 5.9 203

 

Southern LP

1.7 28.7 55.6 11.8 2.2 178

 

I think industrial private forest land is well
managed. (21g)

882

 

Study Forest

2.8 23.8 49.5 18.9 5 281

 

Rest of UP

2.2 24.2 55.2 14.8 3.6 223

 

Northern LP

1 21.6 57.3 15.1 5 199

 

Southern LP

0.6 14 59.8 20.7 5 179

 

I think non-industrial private forest land is
well managed. (21h)

881

 

Study Forest

1.4 19.2 54.4 19.6 5.3 281

 

Rest of UP

1.8 22.3 56.8 15.5 3.6 220

 

Northern LP

2 20.2 60.1 13.8 3.9 203

 

Southern LP

2.3 18.1 57.1 19.2 3.4 177

 

1 have purchased environmentally certified
wood products. (211)

854

 

Study Forest

2.9 21.1 48 17.8 10.2 275

 

Rest of UP

4.7 24.5 40.6 22.2 8 212

 

Northern LP

4.1 21.6 45.9 17 11.3 194

 

Southern LP

5.2 20.8 47.4 17.3 9.2 173

 

219

 

Survey Statement (Question number)

Percent Response Within Region N

 

I would like to be able to purchase

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

environmentally certified wood products. 865
(211)
Study Forest 8.3 31.8 43.7 9.4 6.9 277
Rest of UP 9.6 37.9 39.7 9.1 3.7 219
Northern LP 11.8 29.2 45.6 8.2 5.1 195
Southem LP 8 42.5 37.9 8 3.4 174
Human skill and resources will ensure that 882
we do not make the earth unlivable. (21 1:)
Study Forest 19.6 42.5 21.8 11.1 5 280
Rest of UP 18.8 50.4 15.2 12.1 3.6 224
Northern LP 21.5 47 18 9.5 4 200
Southern LP 24.7 46.1 12.4 12.4 4.5 178
Humans are severely abusing the 88 6
envrronment. (211 *"
Study Forest 17.9 34.3 19.6 19.6 8.6 280
Rest of UP 17.9 29 26.8 22.3 4 224
Northern LP 20.7 37.4 19.2 13.3 9.4 203
Southern LP 20.7 41.9 21.2 11.2 5 179
Humans have the right to modify the natural 887
envrronment to suIt their needs. (21 m **
Study Forest 2.8 20.2 23.8 31.6 21.6 282
Rest of UP 4.9 14.3 26.9 41.7 12.1 223
Northern LP 1.5 15.8 25.2 34.2 23.3 202
Southern LP 3.3 12.2 26.1 40.6 17.8 180

 

220

 

Survey Statement (Question number)

Percent Response Within Region N

 

Humans were meant IO rule over nature.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(21n) 887
Study Forest 3.9 16 20.9 33 26.2 282
Rest of UP 4.9 14.8 21.5 39.5 19.3 223
Northern LP 4 16.8 22.8 33.2 23.3 202
Southern LP 4.4 13.3 21.7 36.7 23.9 180

Humans will eventually learn enough about

how nature works to be able to control it. 881

(210)

Study Forest 8.2 22.8 26.7 28.8 13.5 281
Rest of UP 6.8 32 26.1 25.2 9.9 222
Northern LP 9.5 26.9 23.9 28.4 11.4 201
Southern LP 6.2 26.6 28.2 26 13 177
The balance of nature is delicate and easily 88 6
upset. (21p)
Study Forest 26.9 43.1 11.7 12 6.4 283
Rest of UP 26 39.9 20.2 10.8 3.1 223
Northern LP 28.4 42.3 13.4 12.4 3.5 201
Southern LP 26.3 42.5 20.7 7.3 3.4 179
The so-called “environmental crisis” has 879
been greatly exaggerated. (21q)
Study Forest 8.7 23.1 31 23.8 13.4 277
Rest of UP 6.3 20.8 29 28.1 15.8 221
Northern LP 9.9 18.3 32.7 27.7 11.4 202
Southern LP 4.5 15.6 33.5 30.7 15.6 179

 

221

 

Survey Statement (Question number) Percent Response Within Region N

 

We are approaching the limit to the number

 

 

 

 

 

 

 

 

 

of people this earth can support. (21r) 881
Study Forest 15.4 30.1 31.9 17.2 5.4 279
Rest of UP 12.9 29 29.9 25 3.1 224
Northern LP 13.5 27.5 29 26 4 200
Southern LP 11.2 31.5 34.3 17.4 5.6 178

When humans interfere with nature, it often 728

produces disastrous consequences. (21s)"

Study Forest 21.1 43.1 16.8 15.5 3.4 232
Rest of UP 19.8 44.8 25 9.9 0.6 172
Northern LP 26.2 45.9 18.6 7.6 1.7 172
Southern LP 28.3 42.8 20.4 7.2 1.3 152

 

Individuals from all regions tended to agree with statements about forest

management to protect wildlife. The statements for which UP respondents tend to differ

from LP respondents involve trade-offs between economic and ecological goals of forest

management. UP respondents tend to disagree with statements that support forest

management for ecological benefits that may come at a cost to forest based industries.

This result is sensible because the residents of the study forest and other areas of the UP

bear the costs of changes in jobs in the forest-based industries in the UP, whereas people

from the LP may not feel that they bear these costs and therefore express attitudes in

favor of forest management that protects forest ecosystem services even in the face of

economic losses to the region. Respondents from all regions, however, tended to agree

that forest management should attempt to achieve balance between economic and

ecological goals.

222

Table A7.2. Socioeconomic Characteristics of Survey Respondents by Region

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Characteristic Region
SA RUP NLP SLP
N N N
% that own property in 51% 142 9% 22 2% 3 1% 2
study area'***
Average # of acres owned2 40 132 40 17 32 3 65 1
% member of forestry 4% 12 4% 9 4% 8 2% 3
organization‘
% member of hunting 29% 85 16% 38 21% 43 12% i 23
club'*** ‘
% member of environmental 6% 17 5% 13 10% 20 9% 16
organization'
Education'**
% some high school 5% 15 4% 10 5% 11 5% 10
% high school or equiv 28% 81 28% 69 25% 50 20% 37
% trade or vocation school 12% 35 6% 15 8% 16 7% 13
% some college 21% 60 21% 51 20% 40 25% 46
% associate‘s degree 9% 25 10% 24 10% 20 14% 25
% college graduate 18% 53 18% 43 22% 44 11% 21
% graduate or professional 8% 23 12% 30 11% 23 18% 33
degree
Average age2 55 292 56 241 55 205 55 185
Genderl
% male 82% 239 75% 182 77% 157 74% 138
% female 18% 53 25% 61 23% 48 26% 48
Mean household income2 S60K 204 $56K 168 $60K 140 $67K 147

 

 

'Pearson chi square test of independence; "' = p < 0.10, " = p < 0.05, *“ = p < 0.001

2ANOVA test used to compare means

223

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Characteristic Refln
RUP NLP SLP
% employed in resource 19% 48 16% 33 7% l2 7% 11
dependent industry’***
Average # of years lived in 45 292 47 240 50 204 48 185
Michigan4
Urban/Rural/Suburb.’***
% living in urban area 2% 5 16% 38 8% 17 14% 25
% living in suburban area 7% 19 15% 35 18% 36 _ 50% 87
% living in rural area 91% 261 69% 164 74% ’ 148 36% 63
Ethnic background’***
% Hispanic, Latino or 0% 0 <1% 1 <1% 1 2% 3
Spanish origin
% White 96% 273 94% 221 98% 195 90% 162
% African American or 0% 0 0% 0 V 0% 0 5% 9
Black
% Native American or <1% 2 5% 11 1% 2 1% 2
Alaska native
% Hawaiian or Pacific <1% 1 0% 0 <1% 1 0% 0
Islander
% Asian <1% 1 <1% 2 0% 0 <1% 1
% Other ethnicity 2.5% 7 <1% 1 <1% 1 2% 3
Religious Preference“
% Protestant 36% 102 38% 86 39% 76 46% 83
% Catholic 39% 109 42% 96 32% 63 29% 52
% Jewish <1% 2 0% 0 1% 2 1% 2
% Muslim 0% 0 0% 0 0% 0 1% 2
% None 12% 33 10% 23 12% 23 12% 22
% Other religion 12% 37 10% 24 17% 33 11% 19

 

 

3Pearson chi square test of independence; "‘ = p < 0.10, " = p < 0.05, *** = p < 0.001

‘ANOVA test used to compare means

224

 

 

 

 

 

 

 

 

Characteristic Region
SA RUP NLP SLP
Political Views’
% Extremely liberal 1% 3 2% 5 2 4 4% 7
% Somewhat liberal 10% 29 16% 37 16 30 19% 35
% Moderate 47% 131 48% 109 42 79 42% 76
% Somewhat conservative 33% 92 29% 67 34 64 29% 53
% Extremely conservative 8% 23 4% 10 7 l3 5% 9

 

 

’Pearson chi square test of independence; "' = p < 0.10, " = p < 0.05, "* = p < 0.001

225

   

uu;WW"