232‘. 7. .. .
if: , ‘2»?

1.!

i1
. , E i.

am. at»?
.0. .l 5119...:
flvl - 1'13

:. .2:
. . .i 4!.
I... .F..\ ‘
:3. :IIMEI Ii.

» 1‘.
Eat. .

. ‘r : z: I.
:.r .153 .«x
.... .. i...

 

 

. a .
.1... 5.

. .11.. in... I»
:I vhlq 15.251.) .
013:2...‘4L
It: 25338..

(i J
sknki
a..- [.33

55:9!!!
cl $-
. fl

.. ld‘~
.0533;th .-

t tycituaru.
’t.‘ It: «.5,
.lrtzlgrn 4'...

FEE?
.2... a 1::

tfiflmflmflmmhzn

am .Lnr 4.3.53 . «mu. 1“an
is 2%, wig” v

‘ .uu : , 333:7..4, .5... V
s?:%.. a. . . kw... .rwwfi. .
B ,3 . . Inst: r3..v|\* n .n i “2%.. ‘
, \l . . .. I. 31»... . in. 3.3M

 

mm

2
Ki,- 5’; LIBRARY
Michigan State
University

This is to certify that the
thesis entitled

A Tool for Assessing Citizen Deliberative Decisions about
Contaminated Sites

presented by

Monica Day

has been accepted towards fulfillment
of the requirements for the

MS. degree in Resource Development

 

 

jaw m

 

Major Professor’s Signature

July 29, 2008

 

Date

MSU is an affirmative-action, equal-opportunity employer

-.—-—-—c—-—

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K'lProilAcc8Pres/ClRC/DateDue.indd

 

 

A TOOL FOR ASSESSING CITIZEN DELIBERATIVE DECISIONS
ABOUT CONTAMINATED SITES

By

Monica Day

A THESIS

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

MASTER OF SCIENCE
Resource Development

2008

ABSTRACT

A TOOL FOR ASSESSING CITIZEN DELIBERATIVE DECISIONS
ABOUT CONTAMINATED SITES

By

Monica Day

Despite acceptance of the benefits of public participation for clean up of
contaminated sites, agencies often struggle with executing public involvement methods.
The literature provides guidance on evaluating methods, but tools for assessing group
decisions were difiicult to find. This study tested an application of the National Issues
Forums (N IF) method for clean-up of dioxins on the Tittabwassee River. The research
recruited people to attend a forum in their neighborhood, discuss clean up options, and
develop consensus about the preferred solution. This thesis assessed the group decisions
according to six dimensions. Forum participants differed in some ways from the
neighborhoods at large and the quality of the group decisions varied. The study concludes
that while the subjectivity of any assessment method makes evaluation of outcomes
difficult, based on the tool developed for this study, the NIF method can produce good

decisions in neighborhoods impacted by contaminated sites.

 

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... iv
CHAPTER 1: INTRODUCTION ................................................................... 1
Research question ................................................................................................................ 3
Setting and approach ............................................................................................................ 4
Organization of thesis .......................................................................................................... 5
CHAPTER 2: LITERATURE REVIEW ............................................................................. 6
Scholarship on assessing group decision-making .............................................................. 15
CHAPTER 3: INTRODUCTION TO THE CASE STUDY ............................................. 19
Case: Tittabawassee River and Floodplain Dioxin Contamination ................................... 19
Study design and data used ................................................................................................ 20
Neighborhoods surveyed ................................................................................................... 26
CHAPTER 4: COMPARISON OF NEIGHBORHOODS TO FORUM
PARTICIPANTS .................................................................................... 29
CHAPTER 5: ASSESSING GROUP DECISIONS .......................................................... 47
Data used in assessing decision quality ............................................................................. 47
Assessment factors ............................................................................................................. 47
Calculation of the six factor scores .................................................................................... 51
Process Fairness ............................................................................................................. 51
Consensus ...................................................................................................................... 51
Decision Stability ........................................................................................................... 52
Preference Change ......................................................................................................... 52
Resolution of Tradeoffs ................................................................................................. 54
Knowledge Gain ............................................................................................................ 55
Converting factor scores into an individual decision assessment ...................................... 57
Constructing a group assessment ....................................................................................... 62
Other data related to factors ............................................................................................... 64
Group decision assessments ............................................................................................... 74
Good group decisions .................................................................................................... 74
Partially developed group decisions .............................................................................. 79
Forums with too little information to assess .................................................................. 84
Summary of assessments and reasons for differences observed ........................................ 84
CHAPTER 6: CONCLUSION .......................................................................................... 90
APPENDIX: RESPONDENT QUOTATION TABLES BY FORUM ....................... 97
BIBLIOGRAPHY. . . . . .. ................................................................................................ .105

iii

LIST OF TABLES

Table 3.1 Recruitment - participation at each stage of inquiry ................................ 26
Table 4.1 Comparison of income, education and racial composition of forum participants

to neighborhoods ..................................................................................... 32
Table 4.2 Reasons for not participating in a forum .............................................. 34

Table 4.3 Comparison of rates of forum participation by gender as reported in baseline
survey .................................................................................................. 35

Table 4.4 Comparison of sources of knowledge of the issue between forum participants
and non participants as reported in the baseline survey ........................................ 37

Table 4.5 Illustration of key knowledge responses across neighborhoods ................... 39

Table 4.6 Comparison of interests of forum participants to non-participants as reported in
the baseline survey .................................................................................. 41

Table 4.7 Comparison of preferred solution of forum participants to non participants as

reported in the baseline survey .................................................................... 43
Table 4.8 Frequency of blocking statements among forum participants ..................... 46
Table 5.1 Consensus positions by forum ........................................................ 66
Table 5.2 Change in mean level of acceptance of clean up scenarios ....................... 67
Table 5.3 Examples of tradeoffs and resolutions ............................................... 69
Table 5.4 Change in percentage of correct quiz answers ...................................... 70
Table 5.5 Change in knowledge self-assessment ............................................... 71
Table 5.6 Process fairness ......................................................................... 73
Table 5.7 Decision stability ....................................................................... 74
Table 5.8 Maple street group three (MSG3) .................................................... 77
Table 5.9 McCliggott Road (MCG) ............................................................. 79
Table 5.10 Gratiot and River Roads Group two (GRR2) ...................................... 80

iv

Table 5.11 Falcon View Estates (F VE) .......................................................... 82

Table 5.12 North Saginaw (NSA) ................................................................. 83
Table 5.13 Gratiot and River roads group one (GRRI) ........................................ 85
Table 5.14 Quality of group decisions based on proximity to river ......................... 86
Table A] MSG3 Preferred Solution Quotes .................................................... 97
Table A2 MCG Preferred Solution Quotes ...................................................... 98
Table A3 GRR2 Preferred Solution Quotes .................................................... 100
Table A4 F VE Preferred Solution Quotes ...................................................... 101
Table A5 F VE Preferred Solution Quotes ...................................................... 102
Table A6 NSA Preferred Solution Quotes ..................................................... 103
Table A7 GRRl Preferred Solution Quotes .................................................... 104

CHAPTER 1: INTRODUCTION

Citizen participation began to be institutionalized in the United States during the
rise of regulatory rule making in the first half of the twentieth century. Participation was
included in rule making to protect individual interests and facilitate public scrutiny of
government (Renn, 1995). During the 19503 and 19603 the value of participation to
improve public policy to improve policy, plans and projects was starting to be recognized
(Renn, 1995). Since then, nearly every federal agency has institutionalized some form of
citizen participation in their regulatory decision making procedures (Renn, 1995).

Citizen participation has been touted for reducing costs and leading to better projects by
reducing the likelihood of litigation and decreasing planning time and changes later in the
process (Abel & Stephan, 2000; Woolcock & Narayan, 2000).

In addition to easing policy making, administrators who make a good faith effort
to engage the public meaningfully can also increase citizen competence and desire to
participate, and engender trust between citizens and government paving the way to work
collaboratively to solve public problems. Benefits of quality participation efforts include
stronger agency—citizen relationships, improved agency credibility, better policy, and a
stronger democracy (Abel & Stephan, 2000).

Despite evidence that citizen participation yields high quality outcomes (Beierle,
2002), citizen participation continues to be a source of frustration for agencies and
continues to be underfunded. Making policy about the clean up of contaminated sites has
been particularly challenging for public health agencies because of the costs involved in
action and the risks involved in inaction. Contaminated sites posing health risks to

wildlife and humans exposed to the contaminants, including reproductive problems, birth

defects, cancer, tumors, neurological defects, immune dysfunction, and liver and kidney
ailments. Strategies to reduce exposure to the contaminant may require land use changes
that impinge further on communities and individuals.

The United States Environmental Protection Agency (EPA), the agency
responsible for oversight of the clean up of contaminated sediments, recognizes benefits
of early, active, and continuous community involvement in risk management strategies
(Council, 1996). In fact, citizen participation programs have become a standard
component of decision making processes for contaminated site remediation. Despite
these efforts, risks persist, the costs of clean-up may seem overwhelming, or the
remediation efforts may seem unacceptably long. Because of the serious health
implications, and because clean—up involves disruption of communities, contaminated site
remediations are contentious public issues that call for sophisticated citizen participation
efforts.

Agencies and other organizations leading and undertaking environmental clean—
ups are challenged with how to most productively engage the affected public in clean-up
decision making. This study suggests an approach for assessing citizen preferences
produced during a deliberative public involvement process known as the National Issues
Forums. A contaminated site remediation case in central Michigan provides the context

for the research.

Citizen Participation
The present study draws upon literature on citizen participation in public decision
making. Institutional barriers to meaningfully engaging affected citizens in public

decision-making recognized by this research include representative democracy and a

technocratic approach to risk management. The proposed solution is for public agencies
to build social capital with affected citizens by creating new opportunities for citizen
involvement, especially deliberative opportunities. The approach seeks to engage affected
citizens meaningfully in the ethical decisions inherent in contaminated sediments clean-
up decisions by using a structured group decision-making process that emphasizes
working-through to resolve tradeoffs in clean-up choice making.

Cleanup of environmental contaminants is ofien approached by the responsible
institutions as largely a technical problem (Council, 1996). While science and
mathematics play a critical role in hazard characterization and risk quantification, risk
communication, including decision making, requires normative judgments regarding
acceptance of risks. The technical experts who are able to quantify the risk provide
critical information for making risk management decisions. However, the people who are
exposed to the contaminants are in the rightful position to determine what risk is
acceptable. Supporting productive communication between these groups, experts and
citizens, is the contemporary challenge of risk communication (McComas, 1998).
Normative risk decisions that will affect many people cannot be made if people are
uninformed or have not thought through the consequences of their decisions. A process of

learning about and defining the options, and thinking through the tradeoffs is needed.

Research question
This research seeks to find out whether, when provided with objective technical
background information on the problem and potential solutions, affected citizens can

attain high quality clean-up decisions using a facilitated, structured decision-making

approach. To answer this question, the research first assesses to what degree forum

participants are representative of the area residents.

Setting and approach

To answer this question 378 residents were surveyed of whom 51 (15%) then
attended nine structured decision making discussions held in neighborhoods along the
Tittabawassee River in the Saginaw River basin in central Michigan. At the time of this
study, the Tittabawassee River downstream of Midland was in the risk assessment phase
of a clean-up of high levels of dioxin in the river sediments and floodplain. The source of
the dioxin was from a chemical production facility that disposed of wastes directly into
the Tittabawassee River before environmental regulations were established in the 19703.

Citizen involvement in cleanup decision making is challenged by scientific
uncertainty and low engagement of affected citizens. Scientific uncertainty around the
health effects from dioxin exposure at levels that are in the Tittabawassee floodplain has
been a vulnerability of the issue that has been exploited by powerful corporate actors.
Their rhetoric has been effective at polarizing the views of the citizenry, and interest
groups have formed for and against cleanup of the dioxin to try to influence the decision-
making. Responsible for leading the cleanup, decision makers such as Michigan DEQ
and USEPA are interested in knowing more about the interests of the “silent majority”
who are not engaged with the polarized interest groups. However, those not involved in
the debate are unlikely to participate in standard public involvement opportunities.

This research was part of an effort to create an institutionally sanctioned

opportunity for the “silent majority,” the affected but unaffiliated citizens, to discuss

several cleanup scenarios together with others in their neighborhood and come to

consensus about their preferred choice.

Organization of the thesis

Following this introduction, chapter 2 presents a survey of relevant literature.
Chapter 3 presents the case study and the larger research project of which this thesis is a
part. Chapter 4 presents demographic data from the study area and compares forum
participants to the larger population. Chapter 5 develops a method for determining the
ability of forum participants to reach about how to address the dioxin issue and displays

an assessment of the group decisions. Chapter 6 concludes the thesis.

CHAPTER 2: LITERATURE REVIEW

Public participation research is motivated by normative and practical goals.
People want a living democracy and effective public policy. Despite decades of research
touting the benefits, in practice, public participation activities often flounder. This
research is motivated by a real-world problem of making more just and satisfying
environmental clean-up decisions. An exploration of some of the scholarship on the topic

contextualizes the study.

Problem

Contaminated sites are complex environmental problems that have historically

been addressed in the United States with a technocratic approach (Fischer, 2000).
The technocratic, synoptic, or rational-comprehensive model is a scientifically based,
expert-driven approach that is characterized by a high degree of specialization and is
“rule bound,” hierarchical, and closed off to feedback (Kelly, 2004; Lachapelle et al.,
2003).

Risk assessment is an example of a technocratic approach to addressing uncertain
dangers. A typical risk assessment for addressing a contaminated sediments problem
includes:

0 Hazard identification: characterization of the harm associated with the
contaminant;
0 Risk measurement or quantification: researching the probability of illness or harm

due to exposure to the contaminant;

0 Risk communication: examining possible corrective actions including technical
aspects of different approaches to conducting a clean-up, deciding which clean-up
approach to take; and,

0 Risk management: conducting a clean-up, and evaluating the effectiveness
(Thompson, 2004).

Science can and should be used to describe and understand the technical aspects
of the risk; however, it cannot determine what the target for protection or repair should be
or resolve other normative issues (J asanoff, 1991). Normative problems, or situations
requiring discerning what’s important and deciding “how things ought to be,” require
more than facts and information; they require judgment. The individuals exposed to the
hazard should be part of deciding what level of risk is acceptable. However, engaging
citizens in decision making around risk management situations is challenging for
agencies.

The emotionally charged atmosphere of contaminated sites requires excellent
public involvement methods. If citizens are from the normative decision making process,
agencies risk degrading relationships with the public and their credibility. Furthermore,
where special interest groups dominate, as is often the case with contaminated sites,
citizens have an increased threat of vulnerability to manipulation. Excluding the public
from the decision making process, in the zero—sum atmosphere of representative
government is at best going to leave a minority of the public dissatisfied with the solution
chosen and at worst satisfy none. Decision makers attempt to limit the influence of
special interest groups, falsely assuming that if they “stick to the science” they can

maintain a scientific-rational basis for decision making. However, science cannot solve

the normative components of the decision making. What is lost in the technocratic
approach to risk analysis is insight into the decision at hand about what is acceptable to
the people who are most affected by the problem (Gregory, 2002).

Recognizing the risks of excluding and alienating the public, public institutions
have expanded risk analysis to include consideration of public interests and values and a
public involvement component (Council, 1996). It is now widely accepted that
contaminated site decisions require an understanding of science, engineering issues,
regulatory requirements, economic impacts and community values. The US.
Environmental Protection Agency (EPA), the agency responsible for oversight of the
clean-up of contaminated sediments, recognizes benefits of early, active, and continuous
community involvement in risk management strategies. Thus, public involvement
programs have become a standard component of decision making processes for
contaminated site remediation.

Although newer programs offer citizens a range of ways they can become
involved, most opportunities are found on the lower rungs of the ladder of public
involvement: information and education. Amstein (1968) developed the “ladder of citizen
participation” to delineate the gradations of participation based on degree of influence
citizens have over the outcome. The first two rungs: manipulation and therapy, Amstein
(1968) describes as non-participation because they are used primarily to change citizen’s
views to match the interests of the decision makers. Public involvement activities at the
manipulation rung are really public relations tools intended to persuade the public. The
middle set of rungs: informing, consultation, and placation involve openly engaging the

public, but fall short of ensuring that the public’s interests will affect the outcome.

Examples of these middle rung public involvement activities include distributing
information by the media, public meetings, and advisory committees. The top three
rungs: partnership, delegated power, and citizen control are rungs where citizens have
gradations of influence over the decision outcome (Amstein, 1968). Examples include
citizen juries, planning committees, and negotiation.

The lower levels of public involvement are ofien insufficient for blending expert
opinion and citizen interests into satisfying contaminated site clean-up decisions
(Gregory, 2000). Traditional avenues for public participation such as public hearings are
designed for the purpose of providing citizens with an opportunity to voice their opinions
or views of proposed government policies, plans or projects. Such a process includes
providing the public access to the proposal; a comment period is offered for public input
including a public hearing to listen to public reactions; and decision-makers then usually
have discretion in how they use the public input in their decision (McComas, 1998).
Where emotions run high and the problem is technically complex, as is usually the case
with contaminated sites, these relatively unstructured avenues for input are not likely to
yield useful information. There may be too little participation, or the public may be
organized into special interest groups where decisions may yield clear winners or losers.
In fact, these venues for public input may even exacerbate conflict between citizens and
experts leading to “decision paralysis” and public dissatisfaction (Fischer, 2000;
Lachapelle et al., 2003).

With inadequate public involvement processes, instead of informing the decision
making process, decision makers operating from a technocratic approach may view the

public’s expectations and desires as unrealistic and ill-informed. Despite believing in

theory in the value of public involvement, in practice, the public is viewed in large part as
a problem to be managed. The agency often responds by emphasizing the need for more
information (Yankelovich, 1991). It is reasoned that if the public were informed enough,
or understood the science better, they would adopt the same or similar solution that the
experts select, thus resolving the conflict. However, simply adding to the information
base rarely results in great success. While education is important, alone it is most likely
inadequate for arriving at solutions that are satisfying to the public and the agency (Arvai
& Gregory, 2003). For conflict resolution it is important for experts and the public to
have two-way discourse about the technical aspects of the problem, but also about other
aspects of concern to the public such as potential economic repercussions, aesthetic
implications, tourism effects, health consequences, or impacts on cultural traditions
(Arvai & Gregory, 2003; Slovic, 1997). Agencies may dismiss such concerns fi‘om the
process because they fall outside of their authority. Due to the specialized nature of their
responsibilities, agencies are typically constrained to focus their attention on technical
solutions within their domain.

By just having an open, non-directed public involvement process such as public
hearings or some other open public comment period, the agencies will probably not be
able to get the kind of information about public interests that can be useful for making
satisfying normative judgments in risk management (Gregory, 2000). With open input
many of the affected people may not participate, skewing the kind of information
gathered. Also, those who do participate may not have worked through the issues to

provide fully developed public opinion (McComas, 1998).

10

Public Involvement Methods - Deliberation and resolution of tradeoffs

Aware of the need for participatory practices for complex problems, practitioners
and scholars have been pursuing alternative public involvement methods since the early
19903 (Abel & Stephan, 2000). Conflict resolution, negotiated rulemaking, mediation,
bargaining, citizen juries, citizen panels, citizens” advisory committees, citizen initiatives,
study circles, planning cells, and consensus groups are some of the methods that have
been tried (Renn, 1995). Many such methods emphasize the role of discourse between
citizens and agencies for their effectiveness (Lachapelle et al., 2003; Slovic, 1997).

Using dialog and deliberation is one useful and practical way to engage citizens in
decision making. The aim is to facilitate the discussion to help participants to resolve
disagreements and differences in a non-adversarial manner, and steer them toward win-
win outcomes (Spano, 2001 ). Facilitation is also important to ensure that power
differences are held in check. Deliberative public involvement methods are credited with
their legitimizing effects for the agency, increasing sustainability of the decision,
improving policy relevance, contributing to citizen empowerment and capacity building,
and resolving difficult conflicts (Kelly, 2004; Kothari, 2001; Pellizzoni, 2003).

Contrasting the technocratic approach, Gregory (2002) articulates the importance
of wrestling with tradeoffs in risk management decision making:

With so much attention given to community participation in risk-based

deliberation, an obvious question is: Why do there remain so many

stalled, unpopular, or otherwise problematic risk-management initiatives?

We believe that a primary reason is the inability of most community-based

risk initiatives to deal effectively with the difficult value trade-offs that
emerge over the course of nearly all risk policy consultations.

11

Yankelovich (1991) also emphasizes the importance of resolving tradeoffs for
reaching public judgment. The term is used for a type of public opinion “that exists once
people have engaged an issue, considered it from all sides, understood the choices it leads
to, and accepted the full consequences of the choices they make.” He describes how
public judgment is cultivated and identifies its characteristics. He proposes that “the
quality of public opinion be considered good when the public accepts responsibility for
the consequences of its views.” According to Yankelovich, if the consequences of a
particular choice are acceptable and they are willing to be responsible for those
consequences, it is a good choice. Two additional characteristics of quality public opinion
include the firmness with which an opinion is held and “consistency,” the extent to which
their thinking is integrated rather than compartmentalized.

Most contaminated sites do not have a clear technical solution. The public may
not accept an agency proposal if there are substantial unwanted outcomes or if the public
has not worked through the tradeoffs between non-action and potential solutions
(Jasanoff, 1991). By not incorporating public values (Smith, 2003), or dealing with
tradeoffs of critical concerns (Gregory, 2002), traditional public involvement methods
often fail to yield satisfactory outcomes, may undermine the legitimacy of the agency,
and may poorly address the problem.

Public involvement theorists emphasize dialogue between decision-makers,
agency representatives, other stakeholders and the public (Slovic, 1997) to incorporate
values into the decision making process (Gregory, 2002). Giving attention to values in
risk management decision making is consistent with the conflict resolution practitioner‘s

recommendation to focus on interests, not positions in order to find mutually beneficial

l2

solutions (Gregory, 2002). However, for as much as discursive methods are
recommended, they are not widely used effectively.

Theorists propose various explanations for this shortcoming. One that has already
been mentioned is the technocratic approach of many public agencies. According to
Yankelovich (1991), this approach makes poor use of values because expert knowledge is
culturally revered while public or lay knowledge and feelings are culturally devalued.
According to Yankelovich (1991), there is a cultural bias against the emotive in favor of
expert knowledge. Similarly, the “objective” conception of risk: the probability of harm
based on mathematics and scientific knowledge, is not how the lay public usually views
risks in their lives. Instead the lay public’s risk perceptions are largely based on personal
experiences with risk. The social construction of risk is important to contaminated sites
decision making because how individuals conceive of the risk will impact the decision
making process. This difference between how the experts and policy makers conceive of
risk and the people who are living with the risk creates a communication challenge for
engaging the public in dialogue.

A second explanation for why pubic public policy makers do not deal with public
values well is for cognitive and psychological reasons: they are unpleasant and difficult
issues emotionally to deal with (Gregory, 2002). More specifically, drawing on the
decision sciences and behavioral psychology literature, Gregory (2002) identifies six
reasons that value tradeoffs are difficult: 1) they are multi-dimensional, for example there
may be human health concerns, economic concerns, social impact concerns or moral
concerns; 2) there is often outcome uncertainty making it difficult to come to a decision;

3) citizens are usually inexperienced and unfamiliar with making public policy decisions;

13

4) thinking and feeling carefully is difficult but necessary for disentangling values
tradeoffs; 5) incorporating emotional elements into the decision is difficult for the
administrators; and 6) such decisions often require learning over time to be made well
(2002)
To address these barriers, proponents recommend structured, values-centered
decision making (Gregory, 2002; Slovic, 1997; Yankelovich, 1991):
In contrast to the conventional, science-based process of risk analysis, a
values-based approach starts with the interests of citizens and attempts to
structure alternatives based on their concerns and expressed trade-offs
(Gregory, 2002).
Values are defined as:
an expression of concerns or interests about what matters to individuals, in
terms of both their preferences for different goods or activities and the
underlying moral or ethical beliefs that give rise to these choices (Gregory,
2002)
Tradeoffs occur whenever getting more of one thing of value requires giving up
something else that is also desirable (Gregory, 2002).
Why do the hard work of sorting through values tradeoffs? Gregory (2002)
explains that the resulting benefit is more satisfying decisions:
Addressing trade-offs is rarely an end in itself: in most cases, a focus on
trade-offs is recommended because it leads to a better definition of the
risk-management problem a more open discussion and better
understanding of the fundamental values of participants, and an improved
and more broadly accepted set of recommended risk-management
alternatives.
The theory suggests that one path to reaching sustainable risk management

decisions is to engage the affected public meaningfiilly in a deliberative decision making

process about the values tradeoffs inherent in each option. A call for this approach is not

14

new (Council, 1996); however, empirical studies of the feasibility of such an approach

are rare.

Scholarship on assessing group decision-making

Although important, collective decision-making about contentious environmental
problems is difficult work. Furthermore, how can practitioners know if group decisions
satisfy the affected public? Survey of the literature resulted in an abundance of theory
about what constitutes a good public involvement process, but little concrete guidance for
practitioner for evaluating public involvement outcomes. An assessment tool, which will
be described in chapter 5 was developed based on scholarship on evaluating public
involvement methods. Public judgment, process fairness, and resolution of tradeoffs are
core concepts introduced here. ‘

Reaching “public judgment” is a term made popular by Daniel Yankelovich
(1991), to refer to decisions where people have settled on and accept a position. It
indicates where struggling has ceased and acceptance can leave people feeling good
about taking action. Several factors from Yankelovich’s theory were modified for the
Tittabawassee River Contaminated sediments Case.

Yankelovich (1991) emphasizes three characteristics of public judgment:

1. Responsibility for the consequences of views.

2. Consistent, dominated by holistic thinking vs. inconsistent or compartmentalized
thinking.

3. Stability of views.
The characteristics are compared to public opinion as determined from polls.

Yankelovich (1991) provides guidance for how over time on various major national or

15

international issues, the public incorporates information with their values and accepts or
rejects policy choices. Yankelovich (1991) defines public judgment as carefully thought-
through opinion.

When making a decision, agencies and policy makers value knowledge about
what the public wants and feels about the issues. However decision makers need quality
public input about the problem, not simply public opinion. Public judgment, if captured
can provide guidance for policy makers about what will and what will not be acceptable
to the public. It incorporates people’s interests, values, feelings and ethics about an issue.
Public judgment, more than public opinion, can help to develop more lasting decisions
that serve the interests of more people and are therefore more just.

Deliberative risk communication theory (Beierle 2002; Arvai and Gregory 2003;
Habermas 2006) is another literature that provides guidance for how to assess decision
making. Much of the deliberative risk communication theory suggests that practitioners
apply structure to public involvement processes, rather than just accepting undirected
open comment such as may be provided at a public meeting. Process guidance
emphasizes issues of fairness (Habermas, 2006), resolution of tradeoffs (Arvai &
Gregory, 2003), level of knowledge (Arvai & Gregory, 2003), comfort with decisions
(Arvai & Gregory, 2003), how well choices reflect their concerns (Arvai & Gregory,
2003), and achieving consensus (Webler & Tuler, 2000) as indications of quality
decisions.

Webler and Tuler et al. (2001) identified five distinct ways that the public decides
whether the deliberative process was a good one. These include 1) the process “acquires

and maintains popular legitimacy,” 2) the process “facilitates an ideological discussion,”

16

3) the process is fair, 4) the process minimizes power differences, 5) the process results in
compromise (Webler, Tuler, & Krueger, 2001). The public decision reached during the
process must satisfy the population who is affected by the decision in addition to the
regulatory requirements.

Renn, Webler, and Wiedemann (1995) present suggestions for how to evaluate
various public involvement methods: citizen advisory committees, citizen panels, citizens
juries, citizen initiatives, negotiated rule making, mediation, compensation and benefit

sharing, and Dutch study groups. They suggest several factors that should be present if

9, 6‘ 9, ‘6

the process is good: “agreement on values,” “fairness, competence, efficacy,” and
“applicability.”

Process fairness, or procedural justice, is commonly regarded as a basic
prerequisite for effective communication. Jurgen Habermas is a leader in the theory.
Others such as Suskin & Kruschank have made the theory more accessible for
practitioners. Arvai and Gregory have published extensively on the nuances of resolving
tradeoffs and balancing values in environmental decision making (2000, 2002, 2003).

Level of knowledge and achieving consensus are two common criteria of quality
decisions. However both of these criteria have been critiqued and are not always included
in evaluations of decisions or processes. For example, many practitioners emphasize the
importance of clarifying values over gains in knowledge. Furthermore, according to
many practitioners, quality public deliberations may not result in consensus. Gregory
(2000) finds that if public deliberations result in a clarification of what is important to the

public and thus provides substantive input for decision makers, then the deliberation was

effective.

17

Consistent (holistic thinking), comfort with decisions, and how well choices
reflect participants’ concerns were not incorporated into the evaluation framework.
Coming together to support a chosen solution requires that some people (those who enter

with disparate preferences from the group position) change their perspective.

18

CHAPTER 3: INTRODUCTION TO THE CASE STUDY

This research examined the prospects for using deliberative methods to help
engage a broad spectrum of people and to help them make better decisions about
environmental management. For the case of decision-making about cleaning up the
Tittabawassee River and floodplain the research asks if participants in the forums reached
informed and lasting decisions. This chapter introduces the case and the larger study of
which this research is a part, the sources of data used in the research, and the analysis
undertaken.

This study is one part of a larger Michigan State University research study
sponsored by the United States Environmental Protection Agency. The larger study
assessed, from a variety of perspectives, the efficacy of public issue forums as a method
for achieving meaningful engagement of unaffiliated citizens in contaminated site
communities. The community living in proximity to the dioxin contaminated sediments
and floodplain of the Tittabawassee River in central Michigan was chosen as the test
case. Case studies allow the researcher to illuminate a particular problem by studying a

case in-depth (Patton, 2002).

Case: T ittabawassee River and Floodplain Dioxin Contamination

Higher than normal levels of dioxin have been found in the sediments and
floodplain soils of the Tittabawassee River. Dioxins, as a group of more than 210
chlorine-based chemicals, tend to be persistent in the environment and many are toxic
even in very small amounts. Testing has revealed dioxins levels in excess of the state’s
dioxin standard of 90 parts per trillion (ppt) and a number of areas in and around the river

(EPA, 2003; Diebel, 2005).

19

The source of the dioxin has been traced to production practices in the first half of
the twentieth century conducted at Dow Chemical Company’s Midland facility. In 2003
the Michigan Department of Environmental Quality issued the Dow Midland Plant a 10-
year Resource Conservation and Recovery Act Operating License that governs hazardous
waste management activities at the Midland facility. It includes requirements for
corrective action for dioxin contamination along the Tittabawassee River and floodplain.
The Michigan Department of Community Health issued a fish consumption advisory for
the Tittabawassee River downstream from Midland.

Like many other contaminated sediments cases, the dioxin issue is contentious
within the local community. The main point of the debate is the level of risk posed by the
dioxin contaminated sediments and soils. Dow Chemical, the responsible party, has
argued that the 90 ppt dioxin standard is unnecessarily low because there is no evidence
that people and wildlife are being hurt by the dioxin. The public are aligned on different
sides of the debate, some siding with Dow while others supporting the MDEQ. One of
the stakeholder groups, the Tittabawassee River Watch, is establishing a class-action
lawsuit against Dow to obtain medical monitoring of citizens living on the floodplain.
Among stakeholders, the issue has become polarized and the level of distrust amongst the

parties is high. There has been extensive media coverage of the issue within the region.

Study design and data used

The National Issues Forums (N IF) designed by the Kettering Foundation provide
a deliberative format for decision making about complex and contentious social problems
through structured choice-making (Mathews & McAfee, 2000). NIF facilitate decision-

making by first naming and framing the issue for participants in a compact and objective

20

document: the “Issue Guide.” The Issue Guide then serves as the basis for a discussion
of the tradeoffs between the example solutions to the problem.

Deliberative discourse requires that citizens have knowledge of the basic elements
of the public policy problem, the relationships among those elements, and about the
consequences and tradeoffs associated with alternative policies. To achieve this, the study
team prepared an Issue Guide on the Tittabawassee contaminated sediments issue. The
document provided the study participants with a framework for sorting through the
contaminated soils and sediments issue and the tradeoffs associated with alternative
courses of action.

Prior to writing the issue guide, the study team conducted an intelligence
gathering exercise to gain an understanding of the range of perspectives on the
Tittabawassee dioxin contamination and potential solutions. Face-to-face, in-depth
interviews were conducted with representatives of stakeholders associated with the dioxin
issue including representatives of the MDEQ, Dow, local governments, state and local
environmental organizations and citizen groups that had formed because of the dioxin
issue.

The issue guide was developed based on information learned during the
intelligence gathering efforts. The document attempted to set forth in a reader-friendly
manner the Tittabawassee contamination problem, how the risk management process is
used to make decisions, what possible courses of action might be taken, and what
tradeoffs are inherent in each of the choices. Concerns beyond the level of risk from
exposure to dioxin were compared for each outcome option including the ability to

reduce dioxin exposure, role of govemment, financial cost, impact on floodplain

21

residents, impact on property values, and key challenges. Every effort was made to
present the information in an unbiased manner so as to not lead the reader to any one of
the possible solutions. The reader was encouraged to form his or her own solution to the
problem.

Issue forum participants were recruited from selected neighborhoods situated
along the 22-mile stretch of the Tittabawassee River affected by contaminated sediments.
Half of the selected neighborhoods were lower socio-economic while the remainder were
middle class. Local residents who were not members of any recognized stakeholder
groups associated with the Tittabawassee case were the target population because one of
the goals of the research is to assess the NIF approach for engaging populations who are
affected by the decision but who are frequently underrepresented in public decision
making. A screen was developed to determine whether an individual met our definition of
an “unaffiliated resident.”

The target issue forum size was 6-10 to ensure enough people to generate a
deliberative conversation but not too many that some individuals would stop engaging.
The issue forums were organized by neighborhood with each participant attending one
forum. Neighborhoods were determined by a minimum of 50 residences, aerial
photographs to identify neighborhoods on or near the floodplain, census units for
delineating boundaries, and windshield neighborhood visits to ground truth. The forums
were conducted at neighborhood venues such as schools, church halls, and YMCAs. At
each forum, a trained neutral moderator facilitated the discussion. The forums lasted

approximately two hours.

22

Recruitment to the forums was done by knocking on doors in the study area.
Recruitment by door knock is only practical for problems that are confined to a relatively
small area. However, it was conducted both to improve the response rate and accuracy of
the survey and also to encourage higher attendance at the forums. This recruitment
method, while more intensive than mailed invitations, was employed to have a larger
sample size and more representative sample than a mail survey. For these reasons it is a
practical method for real world public decision making.

An effort was made to reduce sampling error in canvassing neighborhoods. Every
house in a neighborhood was approached. In an effort to reach people at home who may
be working different shift schedules, each neighborhood was canvassed twice; doors
unanswered during the first visit were knocked again later that day to attempt to reach
those who may have been out during the first round of canvassing.

To provide a basis for comparing forum participants to the larger population in
the study area, a baseline survey was conducted in the area with as wide a sample as
possible. The team of four interviewers canvassed the neighborhoods by knocking on
each door and asking adult residents to answer a brief questionnaire. The survey
instruments were pre-tested with a group of residents from a community outside of the
study area. To increase consistency in data collection, the four interviewers who
conducted the door-to-door surveys underwent a training workshop prior to working in
the field and daily debriefings were conducted.

To measure the effect of the issue forums on preferences and knowledge, the
study participants were surveyed at three points in the study process. First, if respondents

to the baseline survey expressed interest in participating in a forum, they were offered a

23

$25 gift certificate as incentive and as gesture of appreciation for their participation.
Those who expressed an interest in participating were given an informed consent
document to read and sign and the voluntary terms of their participation were explained
verbally.

Individuals who signed the consent document were asked a set of 14
supplementary questions; henceforth this is referred to as the pro-forum survey. After
completing the pro-forum survey, each respondent was given a copy of the Issue Guide
and was asked to read it prior to attending the issue forum.

The second survey came at the end of each issue forum. The participants were
given a brief questionnaire to complete before leaving the venue. For comparison
purposes, many of the same knowledge and preference questions from the Neighborhood
Survey were repeated. The self-administered questionnaire also sought feedback on their
experience with the issue guide and forum.

The third and final survey instrument was mailed one month after the respondent
attended an issue forum. Once again, many of the same knowledge and preference
questions were asked as in the earlier questionnaires. Participants self-administered the
questionnaire and mailed it back to the researchers. The purpose of the third survey was
to determine the stability of participants’ views, i.e. whether their views remained
consistent immediately after the forum and a month later.

My role in the larger study involved preparing for data collection, data collection,
and data management. I helped to test the research instruments, make modifications to
the surveys, provide training to other surveyors, prepare surveys and answer booklets,

administer the surveys, mail the follow up surveys, and analyze the data. I assisted with

24

the issue forums by setting up the rooms, greeting participants, recording and displaying
participant comments, and distributing self-administered surveys at the end of issue
forum discussions.

To recap, the data used in this thesis includes a baseline survey covering a wider
sample within the study area; a three-step survey of the forum participants, and data
collected during the forum itself on the process that each group took to arrive at a
decision.

Table 3.1 shows the sample sizes and the proportion of canvassed households
participated in different stages of the study. Of the doors knocked, 46% did not come to
the door, and 21% said “no” when asked if they would like to answer questions. The
remaining 33% of the households in the neighborhoods answered the baseline survey.
14% of the doors knocked the resident agreed to participate in the forum and completed a
pro-forum survey. This group excludes 3% of total households that were screened from
the study because they were affiliated with groups that had vested interests in the clean-
up decision. Many of the pre-forum survey respondents dropped out and did not actually

attend the forum; a total of 5% of initial doors knocked resulted in a forum participant.

25

Table 3.1 Recruitment - participation at each stage of inquiry

 

 

 

 

 

 

 

 

 

 

% of % of pre
% of % of % of remaining forum
baseline eligible to survey
N doors doors 3 rv It r nd nts
knocked Answered u ey comp e e espo e
respondents pre forum
suwgy

Doors

Knocked 1042 100

Doors
Answered 550 53 100

Baseline

Survey 336 32 61 100
Completed
Screened* 30 3 5 9 100
Pre forum

Survey 142 I4 29 42 46 100
Completed
Attended a 51 5 9 15 17 36

Forum

 

 

 

 

 

 

 

* Stakeholder group afliliation screen Each baseline respondent was asked if they, or anyone in their
household is affiliated with a series of nine different stakeholder groups Michigan Department of
Environmental Quality, an elected government official, the Lone Tree Council, Michigan Department of
Community Health, the Dow Chemical Company (current or retired employee), the Tittabawassee River
Watch, a member of a lawsuit against either Dow or the Department of Environmental Quality, Michigan
Department of Agriculture, or the Tittabawassee Voice. Of the 37 8 asked, 19 indicated that they are
involved in a lawsuit against Dow or the DEQ, 14 with Dow Chemical Company, and one respondent
indicated involvement with Michigan Dept of Community Health. These individuals were not asked to
participate in the research (“screened” from participation) because the research was targeting participation
fi'om non-affiliated citizens.

The purpose of the research was to examine a different type of engagement of citizens than is
typically done. Usually stakeholder groups are assembled and target the most active citizens. However this
research sought to understand the average citizen, sometimes referred to as the “silent majority.” To target
these individuals the screen was developed and those already involved in the issue in some way were
excluded.

Neighborhoods surveyed

Nine neighborhoods in the affected area were delineated for canvassing. Freeland
Mobile Home (FMH) is the furthest upstream and close to Midland and not in a
floodplain. Freeland Town (FLT) is also located in a non-flooding area in Titabawassee

Township. F alconview Estates (F VB) is a modular home park with more elderly residents

26

located outside of the flooded areas. Two neighborhoods, McGliggot Road (MCG) and
Gratiot and River Roads (GRR) flood frequently and have confirmed elevated levels of
dioxins on their property. During the study many households in these neighborhoods had
been contacted by Dow or the DEQ and were involved in a lawsuit against Dow or the
DEQ. Four neighborhoods in the City of Saginaw North Saginaw (N SA), East Saginaw
(ESA), Maple Street East (MSE), and Maple Street West (MSW), were canvassed. These
neighborhoods are located along the Saginaw River, do not regularly flood, and are
racially diverse.

The river and floodplain downstream of a manufacturing source in Midland, on
the Tittabawassee River is known to be contaminated with dioxins. Although not densely
populated, there are residential neighborhoods throughout the floodplain area. Individual
homes located along the river that were not clustered were not sampled because many of
them were known to be involved in a lawsuit against the manufacturer or against the
DEQ (stakeholder groups were excluded) and because of difficulty in sampling widely
spaced houses.

Nine neighborhoods more densely populated and in proximity to the contaminated
site were chosen for the study. Each residence was canvassed and if they were not in a
“stakeholder group” they were asked to participate in the study. The data were grouped
by region based on proximity and demographic similarities (see table 4.1). The groups
are: “Frequently Flooded” (FF), “Adjacent to Floodplain (AFP)” and “City of Saginaw”
(SAG).

Nine forums were held throughout the canvassed neighborhoods, some without

any forums, and some with more than one. Two neighborhoods, East Saginaw (BSA) and

27

Freeland Town (FLT) did not have a forum because no residents showed up to the
meetings. One neighborhood, Gratiot and River Roads (GRR) had two forums (GRRI
and GRR2) because so many people from that neighborhood agreed to participate in a
forum. The surveyed respondents from the two Maple Street neighborhoods (Maple
Street East and Maple Street West) attended one of three Maple Street forums named
Maple Street One (MSGl), Maple Street Two (MSG2), and Maple Street Group Three

(MSG3). Combined, nine forums were held across seven of the surveyed neighborhoods.

28

CHAPTER 4: COMPARISON OF NEIGHBORHOODS TO FORUM
PARTICIPANTS

This chapter demonstrates how the forum participants are similar or dissimilar to
non-participants and the general neighborhood population. Income, education, race,
gender, knowledge, interests (reported as “other impacts”) and preferred solution are
detailed in tables 4.], 4.2, 4.3, 4.4, 4.6 and 4.7. In addition, tables 4.5 and 4.8 illustrate
the key knowledge responses and frequency of blocking statements within forums. Table
4.2 lists the reasons people cited when they declined to participate in a forum. The
information is provided to illustrate some differences between participants and non-
participants to understand to what degree forum participants were representative of the
target population. Demographics of forum participants are compared to census data about
the neighborhoods’ characteristics in Table 4.1. Income, percent earning a high school
diploma, and percent white are listed for the different neighborhoods sampled where a
forum was held.

Median household incomes ranged from $13,250 to $65,200. Incomes of forum
participants were on average somewhat lower than the median incomes of their respective
neighborhoods, according to the census. MCG, GRRl, FVE, MSG2, and NSA forum
participants had lower incomes than their respective neighborhoods and GRR2, MSGI ,
and MSG3 participants had higher incomes than their respective neighborhoods at large.
The lower incomes of forum participants may be due to the fact that individual homes on
large lots in some census tracts were not canvassed because they were not within the

neighborhood. For example, the census tract that included Freeland mobile home park

29

had homes located along the river on large lots outside of the mobile home park that were
not canvassed.

In all of the neighborhoods, the percentage of adults with a high school diploma
or higher was at least 77%. Some of the forums appear to have slightly higher graduation
rates because no one was in attendance who did not have a high school diploma (MCG,
FMH, and MSW). Likewise, according to the census, none of the neighborhoods were
100% white, although some of the forums were 100% white (MCG, FMH). Maple Street

East, Maple Street West had the highest percentage of non-white participants.

30

Table 4.1 Comparison of income, education and racial composition of forum
Jarticipants to neighborhoods

Pre-forum survey

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Neighborhood Focus Group Geographic Demographic Census figures for forum
location variable figures . .
participants only
M cGliggot McGliggot Frequently median income $59,896 Qi=8) $35,000
Road (MCG) Road (MCG) flooded % h-S- diploma 36 100
% white 96 100
Gratiot and median income $51,] 1 1 (n=7) $30,000
Riv(e(lj[l:|:£j()is l % h.s. diploma 87 86
Glam" and Frequently % white 96 86
River Roads flooded
(GRR) Gratiot and median income 351,11 1 (:14) $62,500
River Roads 2 % h.s. diploma 87 100
(GRR2)
% white 96 100
Median income $38,929 (n=l 1) $25,000
Falcon View Falcon View Adjacent to _
Estates (FVE) Estates (FVE) floodplain % 115- diploma 92 82
% white 95 91
Freeland Freeland Median income $65,200 (n=4) $20,000
M b'l H M b'l H Adjacent to .
o ' e ome o ' e ome floodplain % h.s. diploma 90 100
(FM H) (F MH)
% white 99 100
median income $53,173
Freeland No forum Adjacent to . No forum held
Town (FLT) held floodplain %h-S- diploma 97
%white 97
. Median income $32,400 (n=4) $40,000
MSG] City of '
Saginaw % h.s. diploma 78 100
% white 64 50
Maple Streets Median income $32,400 (n=3) $15,000
East and MSGZ City of _
West (MSE) Saginaw % h.s. dlploma 78 66
(MSW) % white 64 67
f Median income $32,400 (n=4) $41,250
City 0
MSG3 Saginaw % h.s. diploma 78 100
% white 64 75

 

31

Table 4.1 Continued

 

 

 

 

 

 

 

Pre-forum
. . survey figures
Neighborhood Focus Geographic Demographic Census for forum
Group location variable figures . .
participants
only
North North . median income $25,167 (n=6) $15,000
S , S . City of _
(:grfw (afilsrfiy Saginaw 0/0 h.S. diploma 77 83
% white 67 83
. , median income $13,250
East Saginaw No forum City of o , No forum held
(ESA) held Saginaw A’h's' diploma 57
%white 8

 

Table 3.2 showed participation at each stage of the study. There were 378 people
who started a baseline survey. Of those, thirty people were screened from participating
because they or someone in their household reported that they were affiliated with one of
the stakeholder groups identified on the baseline survey. The remaining reasons
volunteered why they could not participate are listed in Table 4.2. The responses were
open ended and were not solicited. Thirty people said that they could not attend because
of an accessibility issue (they had to stay home because of being someone’s primary
caregiver, they couldn’t physically get to the forum, etc). Twenty-eight people said they
didn’t have time. Not everyone declining to participate in the study offered a reason, and

some respondents offered more than one reason.

32

Table 4.2 Reasons for not participating in a forum

 

 

 

 

 

 

 

 

 

Reason for not participating (30ml % and
Screened for Affiliation 23(30)
Caregiver, mobility issue, health, literacy or other

disability. 23(30)
Lack of time 22(28)
“I’m not interested” & “moving” 13(17)
Prefer to not get involved in things 9(12)
Work conflict 6(8)

“1 don’t have anything to offer” 2(3)
Total * 1 00(128)

 

* Total number of responses does not equal the number of respondents because some provided an answer
and some did not. The surveyor did not ask for a reason. These were volunteered reasons after being asked
if they would like to participate in a forum.

Table 4.3 shows the participation rates of women compared to men. The third
column, “Rate of participation in forums within groups” shows what percentage of men
and women participated in their respective groups from those who were surveyed. For
example, 17% of the men surveyed attended a forum while only 10% of women surveyed
attended a forum. Of the 51 people attending a forum 30 of them were men and 21 were

women.

33

Haw—a ab Goa—Elmo: o». 3:8 on moi—B 6:35:32. 5392. am 26842— ? ammo—mun 2:43.

 

 

.x. 35 e\ BE wow—53o 38 om 5.8 e\.. om
U 3. A5 Hon.— AS on 53-33:: Amy o». moi—B wanna—33m um wanna—come:
«Eon—fie 3 2.3.3.2. non—iguana a @2638? o». no: mo moi—E
<2..qu . 33:: . . . .
can—:38 2.33 25.. 534.2qu newton: 1.95
manna—33 u a 38s mud—Gm»...
Zorn A33 ficfiv $08 5s. 3
3:53 Gowv mafimwv 333 Ho. 3
H053. A331... SowaMV :56:

 

_.. Emfiom 5 En 33.880 88 o». moan. pedomnmam mm m uoaogfimo om so: 68.39va ammvonmo 38: mm cane—Boa 3. ESQEM 90 R82: 0». mow—5D 333358
3. En page: om so: most: 3383:? m3 oxmBEo m2. Ea :55 3S. GA e\e mm 85$an 3. E<EEm mofie 3. 3.x:

n... Emfiom E 9o :38 om 330632. E 328 £51: magma m8 ounce—Boa 3. amfiambm So 5559. om moans 633685.. B m Samoa. Q95 3. Ba 3532. on.
2:43. 438338 mmibm 93 8632.. me. @8829 o». So So 2Au+u3 8:0 $63 Ban. 3.x: on we umnmomumfloa E m mosh: Go: 31.15. Condom—now En
ugowummos 3:8 395 m. Em" “cameo—ma macaw mm o<o~ 9. 559 830833 5 90 338m. 2 8QO 9...: 3X. om :5 Bo: wagon 3933 m 338. 0: En 092

ESP 3.x. o». $583 320on unwound m meg.
inmosaoq fimm 553853 «8 23 c.8333?

34

Table 4.4 compares forum participants and those surveyed who did not attend a
forum along three knowledge variables: heard of the issue, source of information, and
self assessment of knowledge level. Not surprisingly, people who had not heard of the
issue participated at lower rates than people who said they had heard of the issue. The
sources of information most frequently reported were from the media, broken up into two
groups, one, newspaper (26% of responses), or two, television or radio (37% of
responses). The sources of information that yielded the highest participation rates were
public meetings (70% of the people who had heard about the issue from a public meeting
attended a forum) and direct mail (20% of those who heard about the issue from direct
mail). Not very many individuals said that they heard about the issue from a public
meeting (ten total) but seven of those ten people attended a forum. The only information
about the public meeting was that it was a public meeting, we don’t know if it was on the
issue or something else. The very high rate of participation within that group shows that
people who have attend public meetings are likely to attend another such activity. The
table shows different totals because some of the responses were open ended and could
have multiple responses (e.g. source of information). Several of the variables had low

numbers.

35

Haw—n Ah 0252—18: a». .3563 a». 5.233% a. :3 $28 3228: mow—:3 33532.3 2:. no:

Uuaobwgnm um nag—.39 E :5 Han—mun 3.4%

”832.8 38 on

83:: 33233». 5.8.x. 3. 6932353
2. .— c2638? 3. E 82:: i=5:

:2. 31:22:: wag—.1:

23638 38:

.x. 3 .3358 E:— .x. 32858.
A5 Hen: :5 on 53......35 2:. A5 3.
«548.2. umaauga 32::

Qua—.58 83253 3.33258

Knot—am.“ «$192»

m~<o :83 3.90 mmmco

 

 

 

 

 

 

 

 

 

 

 

.48.. ©an wmfimd 338 .oo 3
:20: 38 538 #8 mm _
HOH»? ©qu 523$ :53 C

moEdm ow 533530: mac—Mu? ummbosmo £5305
Zoémwmwoa A33 N2 53 $63 _ 5 3
Hgoimmon 2. 3&0 A83 “3203 ~2me «N _ H
€oa a». 80:9 88 :33 3.2.8 3 G
932 8m: 33 533 32$ EH No
35:0 Boaabm :3 .23 Ed Emu go
King—2.88 ma
099. 85.08 38 flmfid QGV mo 5
HOH>F ammvosmom GmmV 8%Wwwom GS Rmvoamam

 

Ma: 88382: cm 5838mm Eli A223 aomvonaga $50 $323 .529 awed o». Ba 5%.. 8633 $63 so” ~5on 305 99.,
rues—mama .95 39: :5 $28.

 

 

 

 

Em: Ed :38 3A3 _ K 3
32:55 2va ANCNS 303V _ C a
hos 33v #253 mu: 8 ma 5
HOH>H... QumV Sofimov H838

 

5.08; <5 3850 Emma“: :55ch cm 303a muméanaa 959.2: 95303 Ea moan @3303 8:5 :98 808 Sun 050 Rmvonma. Eases—m $30 £53
83033 93 SQ runs; :85 om Em mmmco 226 no” gram 8 88mm 90:. _n<o_ cm 5.933%. .75 @830 £30 uwaomufiau E m moans 88:8 98 90% can no

rues—mama 0», En mmmco Ba 90333 2.0 so” Saga 5 :5 88..

34.558 5 9a .4333» 38 om moan: nfimomugfim mm m vanoonfimo om no: “539va .3338 33.. mm 85533 6% E595 En U383 om moan: 33068.8

3 2.5 93.83 3, now 328 3393:? m9. 8883? won 28 23:0 883m 3%. Emu .x. mm 8.0538 g Esnmum m..\.. cm .oo\o.

36

Haw—a ah 85...

iiflmfiom 5 Ba :38 0». 330630: E 338 393 mac—gm: m3 cans—$8 cw amiagm 9a 2552 o». H.028 333356 E m oaomoQ A33 3 :8 9562 o».
9.50% Homvonaga masm Eu.” ”59:2. ms. 38.829 o». 60 AN «Sow namuonaga $o+~v «$8 3638 9m” 93‘ Rab.» :83 0», 9o mmmco. 1x: 6 Ramona
Baavfioa E m 355. Awkwufixo. 00vaan En 339330: 38m 39%. m. 92 twang—ma macaw mm o<a~ 9. ESQ. nounomgaa 3 :8 8255. r 8QO 93 fix. ow
Ea Ramona 3238 £5 :3 n33 :83 0». En mmmca ”:2.an m 83.8. 0: So 092 ESQ. 3.x. om @283 .8238 8:0 35 98 Saw :3 38.3 cm Ea mmmca
magma m 328. 5 E53.» 632 308:8 o». ooBumnmoa 98. BS” 505% 8 Raga—mam doom—5o So 88 25:: mac? mm 853a.

37

Table 4.5 Illustration of key knowledge responses across neighborhoods

% and (n) of survey
respondents who

% and (n) of survey
respondents who

 

 

 

 

 

 

 

 

 

 

 

Neighborhood Focus Group answered “no” to “have answered that they heard
. about the issue from a
you heard of the Issue?” . .
public meeting
Saginaw Maple Street East (MSG 1 , 2, 3) 27(7)* (0)
East Saginaw (ESA) 26(1 l)* (0)
Saginaw Maple Street West (MSG 1 , 2, 3) 2K9)“ 20(2)
North Saginaw (NSA) 12(5) 10(1)
Freeland Mobile Home Park (FMH) 7(3) 10(1)
McGliggott Road (MCG) 7(3) 20(2)
Gratiot and River Roads (GRRI) 5(2)* 10(1)
Gratiot and River Roads (GRR2) (0) 10(1)
Freeland Town (FLT) 2(1) (0)
Falconview Estates (FVE) 2(1) 20(2)
TOTAL 1 00(42) 1 00( l O)

 

‘Chi square Sig p<.003

Table 4.5 shows the breakdown of responses to some key results from table 4.4.
Although the most individuals reporting they had never heard of the issue were from the
Saginaw Neighborhoods, at least one survey respondent from each of the nine
neighborhoods reported that they had never heard of the issue. It is not shown in the
table, but the two people who participated in a forum and responded that they had never
heard of the issue attended the NSA and MSGZ forums. Similarly, those reporting that
they heard about the issue from a public meeting came from several neighborhoods. It is
not shown in the table, but the people who responded that they heard of the issue from a
public meeting but did not attend a forum were the two individuals from Maple Street
West and one person from MCG.

Table 4.6 compares the interests of forum participants to non-forum participants
as reported in the baseline survey in response to the question, “Are there other ways that
the dioxin contaminated sediments and soils issue could impact your community?” This
was an open ended question and so respondents may have given more than one answer.

The responses were grouped into the categories listed in table 4.6. The interests that were

38

expressed at higher rates by forum participants than non forum participants were that the
dioxin issue reduces property values; people won’t want to move here or will move away;
it will hurt the economy; it increases stress; and that they don’t think that it causes harm.
People with concerns about the impacts on the environment, land use, health, and future

generations were not as likely to participate in a forum.

39

000.0 0.0 008000.000 0.. 0.8.00.0 0.. .03.... 0030.00.50 .0 000.002.0083 00 00000.00 .0 .00
0000000 00300.

 

 

.x. 000 .x. 0.. 0 ”0000.00 8.0 on ”0.0 .x. 0..
A... 0.. 000000000 A... 0.. 00.01. 00.00.0000:
C .00.. 00000.0 A... .00.... 0.0.... 400-003.... 000000000 00.00.0008 00 0 .0 .02....
00.43.00 00.43.0008 .02.... 000000800 0.. 000 $0.0...
9.0.0000 . . 0030.00... .23.
00.00000. 0030.000... 00000000 8.0.... «.0000
.30...0.0 8000000 <0...00.0
.0903 0.0003. <0.000 3... N03... 0 .A. d :0 No
0.0500 .00 00380.00... 30. N083 .28 a. .N
0000.0 £00.. £00. .0 .000 G8 .030. .23. .00 .0
00.0 0.. <0... .090 05.0%. 00...
.00 00000.0...
0.000 000 0000m00 A00. 000.. 03 30.0. 0%.. w. m
000 0.0.00. 000.. @0000.
0.00.0 0000.0 38 58 03. 0. .
.0080000 0.0000 3.. 0A.... .03. ~00 00
U9... 0.0.. .. 000000 00.... A. 3 0:8 03. .80 u...
7.00. 00.8 00.0.0 m0008..000 :0. 2.8 03. d .0
H0500. 000000000 .0000 00000 EN... * .023... .000...
E00000.

 

.. 00.0. 0 .0 00. 2.0.8.00. .0 .00 00.000. 0.. 003.0% 0000000006 0000000 00.00 0000.0 m0<0 .008 .000 000 0030.. 70000.0 00000000 0000000.

00.050000 .0 .00 .40000000 8.0 0., .088 00.00.0008 00 0 008008.00 0.. 000 00....0.000. 800000 8.0.. .0 00.00.0.00 0v. 0.<.0.0m .00 00.000. 0.. 00.00. 00.00.0030
0v. .00 00.000. 0.. 000 .088 00.00.0000. 00. 0.8.00.0. 000 .00 00000 80.00000 0.0000 .00.. N00 00 .0 00.00.0.00 0v. 0.<.0.0m .000 00. 000.

3:. .u.m0.00 .0 .00 :88 0.. 00.00.000.00 .0 00.0.0 $00.0 90000.. 0.0 00.000000 0.. 0.<.0.0m .00 00.000. 0.. 00.0.0 03.0.0008 .0 0 00.0m0J. 9.00... 00. .00 .08.
000.00. 0.. 003.0% 0000000000 m.<.0m .00. 0000.0... .00.. 0.8.00.0. 0.. .00 .0 00.00... 0000000008 A. o+mv 0.00 00000.00 .00. .03. 000.. 0.0.0 0.00.0 000000 00...... .300.
G 0000000. 0030.00.00 .0 0 .08... G: muuwfiov. 000.00...0m .00 00....0.00..00 8.00 0009.0 ... .00. 0.5.00.0... @000 .0 00.0. 0.. 0000.. 808000.00 .0 .00 00.0000. ..
.0000 .00. Box. 0.. .00 00000.0 0020000 200 0.00... 0.0.00 00000.. 00000 00.... 0000000 0 00.0.0. 0.. .00 0.00. 0000. 00... .00. 0., 00.0000 00.4300 €00 8000.00
.00. .0000 0.0 090000 000.0 0000.0 0.0.0 0.00... 0000000 0 00.0.0. .. m.<00 0 0000.. 0.00050 00008001000 .000 .00. .000.0m 0. 0000008000 0000000 .00 8.0
«1.0.0 «.0000 .0 00.0.20.

40

Table 4.7 compares the preferred solution of forum participants to non
participants as reported in the baseline survey. Because it is an open ended question,
some people reported more than one preferred solution. The third column shows the rate
of forum participation within groups. For example, 14% of the people who said they
don’t know participated in a forum. Comparing the percentages in the first two columns
shows the relative frequency of the response. For example, 16% of the responses among
non-participants were to regulate or litigate. And this group was slightly under
represented in the forums because only 9% of people suggesting regulation or litigation
attended a forum. Dredging, increased citizen involvement, moving people out of the
contaminated areas were among the least popular ideas among participants. The preferred
solutions that were most represented were research, monitoring, or further investigation;
respondent eliminates a solution (such as not dredging); nothing can be done; high tech

solution; I’m not convinced there is a problem; and I trust the experts to take care of it.

41

H96? 94 0:86:18: 3. 6338.2— 8832— cm 8.88 wanna—85¢ 8 :2. 3188.88 3 3.832— 8 =8

"8.8—8." 2:43.

 

”8.88.3 38 S.

 

 

 

 

 

 

 

 

 

 

 

 

e\.. 28 e\.. 2:— n . . 5.8 e\ on.
H. — Aflv On. Haw—vOH—mflm 3% ABV On. 02:- wa—J—O—‘N—ug Um . mo . .
$382.3 88:2. A5 c8 82783.8 now—888 a wanna—“Ema a. no: .835 8:8: 8
, mczavda . . 2.32 an» fig 25¢ 33:: «in—:5
335888 38 3.88 v 3:... u I:
gun—.83 8.62.2: 1938888 38 mac—Em
78:85 8.8288 8850
U084 58$. GMV $88 328 3 E
WEE"? :fimflo 33 38 3 Ed 8 _ o
:98: x 8.. 88 338 03 3 8.
H3886? 8088150 8 G C 808 328 3o NA
8.93 853888 88
8... 83208
3.0.38 908 ~38 80:88? 38 008 5:3 TE 8
888 3 flow 8:88 om
90 988.
Gamma Ed 83$ 28 3 .\.
”98289: «.8883 w C3 .2 5V 2d woo 3
8888...
898,68 0888 85.889: £8 mCNV 2: mm x
28 88.30838: 8 8a
$.80
W88 3883mm. 28 No: 5 N28 co 3
§Qo8§a8m ow :8 $89
8.08% 288808 8 3%
882m 888 968:8
283 @820 08 03.5 :3 Nb: 3 52V 3 m
8888888 mag
:00 m08088m: ow 082 C 3 Nfiov ~28 Sm ;

mason: $8882:

42

H88 94 803.9

 

 

 

 

 

 

 

 

..\o 8.— .x. 8.. ”88088 88 0». ”98 e\ 0n
A5 0.. 8200888 5. A5 0». H088 828888 3 . .o . .
. A5 H08. 2:82.820: 8
88888.. .8820: 808-888 8.8088 a 88888 0.. =0: . .
85.88.. . . . . 83.8 838
338888 :68 888 2.3.2.8.: 88088 3: mi:
@8888 88283 8.18888 88...»... m w
.388 mm 80 88:08. 8088m _ o _ o q A N: mo
28 8 88 A V . A V 8
U0 80888 884 8888. §
888 88 88 0». w 08—. A8 8.28 8.25 am 5.
888
m880v. 8 82888 80:80: A9 :38 ~28 88 3
20" 003808 888 mm m
30888 38 .28 £8 moo mo
:8 8m” 88 369.3 8 8—8
88 0». 8.. on .988 w 8 A3 :38 ~28 88 um.
88 $628..
00<2 8 A3 83 088 o o
HOHZ. 88088 A088 283. A381. 68 38

88:88 88088 @8803:

_. Mon «8888. 8 880888 83. 8% J 884 888. 8x88 Qaam8m mm 808 88 88:08...

.1. w088 8088 mm8 808 88 08 89.88. 8888 8 £8 8 088 889 88:88 88088 £8,808.
..:.:._Em=8m 8 88 .8888 88 0m 888 858888 8 8 88888 0m 808 33888: 88088 88.. 8 888888 3. E<88m 88 08888 0», 838 8880888
3. 88 8888 0m 80: ”.088 8.8.8888. m8. 98.888. 8. 88 880888 828888 8 8:80: ~08. woo o\o 8 888888 3. 8<88m 9X. 3. 88.

Iimmmfidm 8 88 :88 0% 8388888 8 828 £88 «388.. m8 388888 3. &<88m 88 888.. 0». 88 88.20888 8 w 8880. @038 3.. 88 88—
:8an 0w "8.82 8808888 mm<8m 88 8322. won 38888. 0m 88 5 88— .823. 8808888 2~+C :80 8808.88 88 83. :88 88 80888 0888
8878888 88 8588888 8 88 88 mm 8889. m..\.. 2 8803 3388888 8 8 8:8 A: 8 mnmfov. 00888.8m 88 3308888 828 898 R 88 88.883
macaw mm 0<Q ow 882 8888888 8 88 8.088? = 8QO 88 mo\o 0m 88 8808. mE.88¢ 2:0 88w 8 8.888 8 88 8.0288 mm 8.883 0888 88488808
8888 m 888. 08 88 082 :88. 08v. 3.x. 0m 8808 88.888 <80 88088 88 J 8.8» 88 968.8 8 88 88 on 3.. 01.888 x 8 88 8688.. 8883 m
888. r 888 m 832 88.88 0m 0088808 88 8% 828m 8 888888 8088 88 88 £888 «388 mm 8838.

43

Table 4.8 displays the frequencies of selected variables from tables 4.6 and 4.7
among forum participants. They include 7 variables from the above tables, chosen
because they are “blocking” or solution-inhibiting expressions. The people expressing
these views participated in disproportionately higher rates than many of the other views

expressed in tables 4.6 and 4.7.

44

Haw—a A.« 5.388% 0m 285.5 888—888 BEEN 83:: 888888
ZOO OE 38w Oahu 29» a 3mm— HSmQN a 88—

 

Hwamomaor. 80888? 8.82. 88888: H H - u - u H H H 32H H8
H.858 808 H» 8.883 88 H - - w - N - - H 3.28
H 88; 888 888 883 88.8 m - N - - N - - - 3.28
”880888 $8888 a 8:808 88 8 888m - H - N H N - - H 3.28
.388 Hm :0 88308. 88.» no 888m - H - H - H - - - «.28
20" 89.888 888 Hm 8 8.0288 H - - H - H - - - H28
888 H” 8 88 962.8 H - H - 28

 

885.888.8853"... 3.3 8.3 m3 3:: 83 .32va 83 8.3 83 833

45

Summary

Forum participants differed from the larger population in some interesting ways.
Demographics, knowledge, reasons for participating, and interests were compared. The
low number of observations makes it difficult to determine if the differences observed
can be attributed to chance. More men participated in a forum than women (59% of
forum participants were male, 41% female).

Some participants who declined to participate gave reasons. People frequently
said that they didn’t have time, couldn’t get out, were moving or weren’t interested,
preferred to not get involved in things. Three people said they don’t have anything to
offer.

Although few people said they heard about the issue from a public meeting (ten
people), 70% of those people attended a forum. The interests that were most frequently
citied as concerns from the dioxin issue by participants included “reduces property
values” (31% of forum participants stated this concern), “people won’t want to move
here, will move away, or harms the economy” (19% of forum participants gave this
answer), and “harms the environment” (17%). These numbers were roughly similar to
those of the general population. Likewise, the leading responses by non-participants were
some of the leading responses by forum participants. Two exceptions include “regulate,
and litigate”) and “dredge.” People with these views participated at lower rates than the

people who stayed home.

46

CHAPTER 5: ASSESSING GROUP DECISIONS

This chapter develops and applies a method for assessing group decisions. The
first part of the chapter presents the approach for developing the method, after which it is
applied to the case. The approach combines theory about deliberative risk communication

with public judgment.

Data used in assessing decision quality

Participants were asked a variety of open-ended, scale, and true/false survey
questions before and after their participation to observe the effect the issue forums and
issue guide had on group decisions. The discussions were recorded and transcribed and
process evaluation questions were asked following forum participation. A comparison of
responses before and afler the survey provides the basis for analysis. In addition,
differences between the responses immediately after the forum and a month later provide
data about the extent to which participants retained what they had learned at the forum,

and about the stability of adherence to the group consensus.

Assessment factors

As introduced in chapter 2, collective decision-making about contentious
environmental problems is difficult work, and because of the inherent subjectivity, it is
difficult to assess the success of collective decision-making efforts. In accordance with
the discussion in chapter 2, this study combines methods from deliberative risk
communication theory (Arvai & Gregory, 2003; Beierle, 2002; Habermas, 2006) and
public judgment (Yankelovich 1991) to develop a set of aspects of the forum decisions
that could be measured with the data collected. The criteria for this research are

creatively developed by the author but are based on the literature. The criteria are

47

referred to as “factors” and they characterize the decisions made. The following factors
are the components of the assessment tool:
0 Consensus: Participants are in agreement with the group position.
0 Decision stability: Preferences of the individual expressed immediately following
the deliberation were held over time.
0 Preference change: Participant’s favored solution before the deliberation changed
after the deliberation.
0 Resolution of tradeoffs: Mitigation of downsides of a chosen action were
incorporated the proposed action.
0 Knowledge gain: Participants were able to more accurately articulate the problem
and or solution.
0 Fairness: The process, facilitator, and information provided were fair.

As discussed in the literature review, many authors have written about what
constitutes a good public involvement process; however I was unable to find in the
literature a suitable assessment tool for assessing the group outcomes. Therefore, a
decision assessment tool was developed based on factors that emerged from the literature
cited above. The assessment tool assigns scores for each of the six factors to individuals
and groups involved in the decision making. The possible scores for each factor are: 1 =
yes, the factor was found (i.e. the criterion was met); 0 = no, the factor was not found (i.e.
the criterion was not met); and n.a. = unknown (because relevant information was not
available.) For each case, the scores for the six factors are summed and an overall score
is assigned to the case. The score is equal to the sum of the individual scores divided by

the number of criteria scored (i.e. those with an unknown score are excluded).

48

The raw overall score is then translated into an assessment of the quality of the
decision. The highest overall assessment score is 1 or 100% (e.g. a score of 1 is achieved
for each of the six criteria, such that 6 / 6 = 1 or 100%). The minimum possible score is 0
or 0% (e. g. the score for each criterion is 0; 0 / 6 = 0 or 0%). Given that for some criteria
the score is unknown, the denominator when calculating the overall score might be less
than 6. In other words, the quality assessment is derived from the number of Observances
of each factor divided by the possible Observances for that factor.

Assessments are based on a combination of data collected at the group and
individual levels. Focus groups at the forums produced decisions about the preferred
action regarding dioxin clean up. The positions reflected in these group decisions were
reached through consensus. Sometimes they contained multiple components. The
consensus positions were recorded as a visual tool on flip chart paper during the forums.

The scores for each factor are developed from survey responses. A group’s score
is the aggregate of the scores of individuals within the group. Individual responses to the
open-ended survey question, “What is your preferred solution to the Tittabawassee
contaminated sediments and soils issue?” answered at three points during the process
over time (before the forum, immediately after the forum, and a month afier the forum),
provided the data used to determine factor assessments. Individual group members
received scores for each of the following factors if information is available:

0 Consensus: was the individual participant’s decision consistent with the group
decision?
0 Decision stability: Did the participant’s preferred solution immediately after the

forum remain the same one month later?

49

Preference change: Did the participant’s favored solution before the forum change
following participation in the forum?

Resolution of tradeoffs: Was there evidence that the participant recognized
downsides of a chosen action and incorporated that recognition into their
decision?

Knowledge gain: Did the participant learn about the issue over the course of their
participation in the study? The primary indicator of knowledge gain was an
observable increase in participant’s ability to accurately articulate the problem
and/or solution.

Fairness: Did individuals express that they thought the process, facilitator, and
information provided were fair? Ability to participate, incorporation of their
views, unbiased information, and unbiased facilitator are components of fairness
assessed.

Then, with individual assessments completed, the individual assessments were

tallied within each group to come up with a group assessment (1, O, .5, unknown). The

numerical scores were checked for meaning and validity by considering the responses in

context (compared to the other participant responses in that group) and they were

rescored. Thus the assessments were assigned iteratively: before rules for the factors were

firmly developed, during development of the factor rules, and following group

assessments.

Several factors were considered for the assessment tool that were not ultimately

included. For example, one of the key components of public judgment as defined by

Yankelovich (1991) is “consistency,” to what degree the person’s values conflict with

50

their decision. This factor was not included in the tool because it was too difficult to

measure.

Calculation of the six factor scores

This section shows how each of the six factor scores was calculated.
Process Fairness

The process fairness score reflects the participant’s response to the set of survey
questions on the post forum survey asking about process fairness. If a participant
indicated that the process was not fair on two or more out of 6 process fairness questions,
they were assigned a O for process fairness. One or fewer responses indicating that the
process was not fair resulted in a score of 1 for process fairness observed.

Consensus

The consensus score reflected the degree to which the individual participant’s
preferred solution was reflected by the group position. A score of 1 (yes) was assigned
when the individual preferred solution after the forum, immediately after or in the final
questionnaire, was part of the group-identified decision. A score of 0 (no) for consensus
was assigned when the individual response did not match the group’s preferred solution,
and an “n.a.” was assigned when there was not enough information.

For example, the group position for Freeland Mobile Home Park (F MH) was
“dredge hot spots, impound and treat.” Participant #300’s individual response, “Put in a
wastewater treatment plant by Saginaw Bay and dredge the river,” received a 1 because it
reflects the group position. Participant #285 of that same group received a 0 for their

response, “Just leave it alone.”

51

Often an individual position would not represent the full group position because
the group decision included multiple solutions. When the individual position reflected
only part of the group decision, a score of 1 was assigned. For example, the group Falcon
View Estates (F VE) included a detailed step by step process: “Reduce exposure, flood
control, open space, dredge hot spots and wait for bioremediation for the floodplain when
available.” Case #27 was scored 1 for stating “land management and open space
corridor,” which captured part of the group decision.

Decision Stability

A decision stability rating of l was assigned for participants where their preferred
choice following the forum was the same choice one month later. A score of O was
assigned where the choice changed substantially, and an “n.a.” was assigned when the
final questionnaire was not returned. For example, in the Maple Street Group 1 group,
case #58 demonstrated a stable decision when stating immediately following the
discussion “dry dredging or the use of new technology to break down contamination” and
one month later “dredging.”

Preference Change

The score for preference change is set equal to 1 for an individual 1) whose
preferred solution as expressed in first post forum questionnaire or the final questionnaire
differs from their preferred solution as expressed in the baseline survey, or 2) if the
position was not developed in the baseline survey, the first post forum questionnaire or
the final questionnaire response provides evidence that they formed a position. The

participant may maintain some of their original views, but their views have been

52

modified to include the views of other participants or other elements of the group
decision.

The preference score is 0 if the individual preferred solution as expressed in the
final questionnaire is the same solution expressed in the baseline survey and does not
include an addition of a new position or new elements of the group position. The first
post forum questionnaire may have indicated a new position that has changed from the
baseline survey; however, the final questionnaire does not include the new position. This
is an indication of temporary preference change. The position may demonstrate learning
new information but the new information appears to fall short of affecting their preferred
solution. Zero indicates either no preference change, or no long term preference change
such as in cases where positions may have changed at first post forum questionnaire but
reverted back to original position on the final questionnaire.

A score of n.a. or unknown indicates that change in views is indeterminate due to
missing or incomprehensible information. Either the baseline survey response is left
blank or it is incomprehensible, or both the final questionnaire and the first post forum
questionnaire were left blank or the responses that were given were incomprehensible.

Preference change was also determined by comparing responses before
participation with responses following participation. Preferences are distinguished from
knowledge because preferences captured a person’s position where knowledge measured
the accuracy and detail of the case. The accuracy of positions was not assessed as a part
of this study. For example, if they selected “open space corridor” that was not considered
more or less correct than “dredging.” A score of 1 was assigned for preference change

when the preferred solution before participation was substantially different from the

53

preferred solution following participation. A score of O was assigned when consistency in
preferences before participation and afler participation were observed. “Na.” was
assigned when not enough information was provided to make an assessment.

As an example, case #4 of the Falcon View Estates group demonstrated a change
in preferences. Before the forum he or she stated, “Studies to see if there is a health
problem, until that’s done, then act.” Following the forum he or she stated, “Identify
where they exist and make efforts to avoid exposure to them. Continue to focus
technological-biological research methods to negate dioxin contamination.” The initial
position was to not act until it is determined if dioxin poses a health risk. Following the
forum his or her position has changed to avoid exposure and treat the contamination.
Resolution of Tradeoflfs

When participants raised concerns about a clean-up approach, and their concerns
were incorporated into the group position, “resolution of tradeoffs” is exhibited and the
score is set equal to 1. The final questionnaire response includes some mention of
concerns and addresses those concerns in the solution. Also, to receive a 1, the outcome
must be that the desired solution is the same as the group position.

An individual received a 0 for resolution of tradeoffs when a concern was raised
and impeded their ability to agree with the group position. Or, 0 is assigned if the
participant expresses that a certain solution is not possible because of a tradeoff. The
group position is not acceptable to the participant because of a particular unresolved
concern. The participant may reach a decision that incorporates his or her concern into
his or her own personal position, but the incorporated solution is not reflected in the

group decision.

54

A score of n.a. or unknown is assigned if the responses given in the first post
forum questionnaire and final questionnaire do not include mention of a concern, or there
is no response, or the response is incomprehensible.

An example of resolving a tradeoff was demonstrated by participant #335 in the
McCliggot Road group (Appendix), “A combination of open space corridor and
neutralizing dioxins via chemicals after the technology has been firrther developed and
more is known about the side effects.” The tradeoff of neutralizing dioxins with chemical
treatment is that the treatment may create harmful side effects. By combining moving
people out of the contaminated area, “0an space corridor,” with applying the
neutralizing treatment after the “technology has been further developed and more is
known about the side effects,” the participant demonstrates that they were able to
incorporate their concerns about the downsides of the solution into the decision.
Knowledge Gain

Participants were given a score of 1 for knowledge gain where there was evidence
that they learned from their participation in the forum and it appeared that the learning
persisted over time. A 1 is given if the response on the first post forum questionnaire and
final questionnaire indicated that the participant gained new information, more clear
understanding of the issue, or corrected for previously held false beliefs about the issue
when compared to the respondent’s baseline survey response.

Participants received a score of 0 if they demonstrated short term learning but
appeared to have lost that knowledge in the final questionnaire. The response on the first
post forum questionnaire may indicate gaining knowledge (such as new information,

more clear understanding of the issue or correction of previously held false beliefs) but

55

the final questionnaire response demonstrates a loss of that knowledge. A zero is assigned
when the participant maintained incorrect information over time. The lack of knowledge
gain may have been due to a failure of the method or the facilitator’s failure to provide
information, such as more definitively establishing the risk of dioxin. A score of O was
also assigned when at no point in their responses did a participant demonstrate having
gained knowledge from participating in the discussion.

A score of n.a. or unknown indicated that responses on one or both of the post-
forum questionnaires were too short or incomprehensible to make a determination of
whether the participant gained knowledge or not.

As an example, before participation in the forum, #221 ’5 preferred solution was,
“Stop dumping. Filter water. Clean up trash.” This response reflected a common
misconception that the pollutants are currently being discharged. The post forum
responses, “Dry dredging,” immediately following the discussion and, “Dry dredging if
the people involved in the dry dredging can be geared up in protective gear,” reflected
gaining knowledge about the problem and the risks associated with clean—up. Thus, #221
was scored a 1 for knowledge gain.

In contrast, participant #252’s response before the forum and after the forum did
not change. The misconception that the health risks are not well established and that the
best solution is to study the issue was maintained over the course of the treatment.
Participant #252’s initial response, “The way they are going is alright - - doing the study”
is essentially the same as after the forum, “Evaluate the health risks on people first. What

level can we live with? Wait and see,” and in the final questionnaire, “Need more

56

answers to how people and wildlife are affected.” Thus #252 was assigned a score of 0
for knowledge gain.

Similarly, participant #244’s responses, “Flood control and total clean up” and,
“Total clean up” did not provide enough information to demonstrate learning when
compared with their response before the forum, “Complete clean up. Stop Dow from
emitting anything. It doesn’t make sense that Dow can still dump.” #224 dropped the
misinformation portion of the response, “It doesn’t make sense that Dow can still dump,”
but the responses following the discussion were not precise enough to make a

determination of whether they learned very much during the forum. Thus #224 received

an “n.a.” for knowledge gain.

C on verting factor scores into an individual decision assessment

Each forum participant received an overall individual decision assessment score
based on their six factor scores. Individual decisions were scored according to the
following set of rules.

When individual positions were in agreement with the group position over time,
the individual was scored 1. A decision is categorized as a quality decision if “consensus
decision” and “decision stability” both hold, and at least half of all the remaining factors
also hold. The factors “consensus decision” and “decision stability” were determined to
be the most important factors; both had to hold for a decision to be considered a quality
decision. I determined the other four factors: preference change, resolution of tradeoffs,
knowledge gain, and process fairness to be of lesser importance for determining the

quality of a decision. To summarize:

57

0 (1) Quality decision was assigned to individuals who received an assessment
of 1 for both consensus decision a_nd decision stability and received a score of
1 on at least three (50%) of the six factors.

0 (0) Low quality decision was assigned to individuals if they did not receive an
assessment for both consensus decision and decision stability, and they did not
receive a score of l on at least three factors. Alternatively, low quality
decisions were reached where not enough information was available (n.a.) on
one of the key factors (decision stability or consensus decision) and there was
a 0 for more than half of the other factors.

0 (0.5) Moderate quality decisions resulted from receiving a l for consensus
decision or decision stability but not both, and greater than or equal to 50%
for the total percentage.

0 (n.a.) Unknown was assigned when not enough information was available for
consensus decision and/or decision stability.

Although the lesser factors provide insight into the participant’s ability to adhere
to the group position over time, their absence does not create discordance with the group
decision. When an individual decision reflects consensus and maintains that general
position over time, the overall group position is stable because it is reflected in the
individual positions. If that individual maintains their position but falls short of explicitly
resolving tradeoffs, learns in the process, expresses concerns about fairness of the
process, or shifts their views in the process, the group position still maintains its integrity.

However, the other, “lesser factors” are usually present if the “greater factors” of stability

58

and consensus are demonstrated. The lesser factors then are indicators of quality
decisions, but not requisites.

Assessing quality is inherently subjective. To determine a scale and evaluation
criteria, I thought about, “What is the goal of the forum?” and, “What kinds of decisions
are needed in this type of scenario?” A more or less rigorous scale could be developed
depending on the needed outcomes. This scale was constructed to result in a wide spread
for the data so as to be able to discern findings that could illuminate how the forums
could be improved to serve this purpose. For example, in this exercise we needed groups
to come to a decision within two hours of discussing the topic. In the real world, they
may have more time to consider the issues but they may not. Given these time
constraints, what kind of decisions did we find? How much change in knowledge was
observed? How stable were the decisions over time? The answer to these questions is that
there was a range of decisions that were made, some better than others. Quality is a
relative term and is contextual. However, this scale could be adapted to many different
scenarios.

Examples

Case #313, McGliggot Road, demonstrates a quality individual decision
assessment as it received a score of 1 for every factor. The detailed survey responses are
shown in the appendix and a summary is in table 5.9 of this chapter. The requirements for
the first factor, fair process were met; the consensus position were met: Case 313’s
individual preferred solution “A combination of neutralizing the soil and river with new
technological chemicals and an open space corridor” was consistent with the consensus

decision of “open space corridor; neutralize hot spots (bioremediation); test on GM

59

property before using wide scale.” The decision was stable because he or she stated as
the preferred solution in the final questionnaire “control future land use to new
development (open space corridor) . . . coupled with new chemical treatment of the soil
and water which will be available in a few years . . . “ The two post-forum questionnaire
responses contrast with the baseline survey response of “educate the people about the
parts per trillion and the low percentage [of dioxin] we’re talking about here . . . “
demonstrating learning and preference change. Finally, tradeoffs of the possible negative
impacts of new technology were incorporated into the decision, declaring to test it first.

Some participants who were assigned a “good quality decision” only exhibited
some of the second tier factors. To receive a quality assessment, they had to have a
position that was consistent with the consensus, the decision had to be stable, and 50% of
all factors had to be exhibited. Several participants with quality decisions, did not
completely maintain preference change over time (0 for preference change), did not
exhibit resolution of tradeoffs (n.a. for resolution of tradeoffs), or did not exhibit learning
(0 for knowledge gain).

Poor quality individual decision assessments were assigned based on receiving no
more than one “yes” assessment for the core factors (consensus decision and decision
stability) and less than 50% of “yes” assessments for all factors.

For example, #285 neither agreed with the group position nor was his or her
position stable over time (received a “0” in both consensus decision and decision
stability.) #285 also demonstrated an unresolved cross pressure “is it even possible to

clean up dioxin?” The decision in the final questionnaire to “do nothing” was deemed a

60

poor quality decision because it did not exhibit any of the factors of quality decisions
including learning or preference change.

Decisions that fell in the moderate-quality range were characterized by clean-up
preferences that did not perfectly reflect the group position over time, but demonstrated
some components of the group position. Moderate quality decisions had 1 for either
consensus decision or stable decision and “yes” for greater than or equal to 50% of all
factors. Or, moderate quality decisions may have had both consensus decision and stable
decision at ls and less than 50% for total factor percentage. Participants who agreed with
the consensus decision immediately following the forum but their agreement was not
stable over time were assigned moderate-quality assessments.

Moderate quality decisions are characterized by beginning to form consensus, but
were yet unable to reach agreement and demonstrate all of the characteristics of quality
positions. Examination of moderate quality decisions is useful for practical application of
the issue forum methodology to identify what might aid a group in reaching consensus.
The methods applied in this study, in which unaffiliated citizens were asked to come to
agreement on a contentious and ofien emotional environmental health issue within two
hours, was purposefully brief. The goal was to determine what quality of group decisions
ordinary citizens might reasonably be able to achieve with the fewest resources from the
agency in regard to preparation and facilitation. The moderate-quality decisions provide
insight into the potential and also the weaknesses of the methodology for achieving

quality group decisions on a difficult problem and with few resources.

61

Not enough information to make a determination (n.a.), was assigned when the
factor data were not available. For example, this assessment was assigned if the survey

was not returned or if that question was lefi blank by the respondent.

Constructing a group assessment

Group scores were assigned based on the aggregates of the individual scores in
the group and aggregate factor scores. 4 values were possible: (1) yes, quality decision
reached, (2) poor quality group decision achieved, (3) moderate-quality decision
achieved; and (4) n.a., not enough to determine the quality of the group decision. I

assigned these scores as follows:

Yes, a good quality group decision was reached when at least 80% individual scores

“yes” and where the total factor proportion was at least 80%.

0 No, poor quality group decision assessment was made when greater than 50% of the
individual assessments were “no” or “n.a.” and the total factor proportion was less
than 50%.

0 Moderate quality group decision assessment was made when, of a sample size more
than 4, after n.a. values were removed and at least 50% but less than 80% individual
assessments were “yes” or “moderate-quality.” Also, the total factor proportion fell
between 80% and 50%.

0 Not enough information to make a determination (n.a.) was assigned when there was

a very small sample size (less than four individuals in the group had assessments

when n.a. values were removed) and valid cases had mixed individual assessments.
In addition to a score, each group’s decision was described, guided by the

following questions:

62

I.

Did the group arrive at consensus? What was the group decision? What percent of
participants agreed with the group decision?

What were the downsides to the position? To what degree were they identified
and incorporated into the position? To what degree were they unresolved?

How stable was the group position for participants? What percentage maintained
the group decision one month later? Of those who did not, did they revert to a
previously held position or did they seem to develop a new position, different yet
from the group decision? What evidence is there that learning occurred or did not
occur? What evidence is there that positions changed or did not change? What
connections are there between learning and position change and development of
stable group positions?

What is the quality of the group decision?

Analysis of the spread in quality of the decisions informed the formation of

assessment rules. For example, the four quality levels of individual decision quality
(quality, moderate, low, and n.a.) informed the assessments of the group decisions,
highlighted the various pathways of individual decisions within groups, and clarified
factor meaning. Variation was observed within each of the quality levels, but, an effort
was made to develop the most simple assessment scale possible and maintain meaningful
differentiation. For example, within the assessment of individual decision score, a 1 was
awarded to individuals who maintained their position throughout the levels of inquiry,
and to those who were slow to agree with the group position but eventually settled on it.
The assessment was made this way because long term adherence was needed for the

group decision to be of a high quality, and temporary changes for an individual, so long

63

as they maintained the group position over time, could be tolerated. Participants could
“sleep on it” and develop agreement with the group position over time, and contribute to
the quality of the group decision. Conversely, the way the data were analyzed, agreeing
with the group position immediately following the forum but then later changing position

would contribute to a group decision assessment of poor or moderate quality.
Other data related to factors

This section displays additional data about the decision factors. Residents were
asked questions about knowledge and their acceptance of clean-up options, however
those questions often did not produce the kinds of results that were useful for assessing
the quality of their decisions. The additional results are provided to illustrate what more
is known about knowledge, acceptance of clean-up scenarios, and process fairness and to
focus in on each of the factors. The results of 5.2, 5.4, and 5.5 were not included in the
final version of the assessment tool.

Table 5.1 shows the consensus positions that groups developed during focus
groups and were recorded on flip chart paper. For example, the Maple Street East (MSE)
focus group decided that the best solution would be to “Dredge the hot spots, and conduct
ongoing testing and re-dredging over time.” Two neighborhoods are listed, Freeland
Town (FLT) and East Saginaw (ESA) but no one showed up to their forums and therefore
there was no group decision. As shown in Table 5.1 all of the groups but two
incorporated dredging into their group clean-up decision. It is interesting to note that
before participating in a forum, only three respondents who attended forums identified
dredging as the best solutions (MCG case #331, GRR2 case #275, FVE Case #14)

(Appendix) yet 23 individuals, and seven group positions expressed preference with

64

dredging following the deliberation. Discussing the alternatives together had a dramatic

impact on the articulated preferred solution. Responses to questions in the surveys asking

respondents to rate their level of acceptance (Table 5.2) (Preforum survey question 4 and

the first post forum question 6) did not demonstrate preference change as clearly as

responses to the open ended survey questions (Baseline survey question 8) “What

actions, if any, should be taken to solve the Tittabawassee dioxin contamination issue,”

first post forum survey question 7, and final questionnaire survey question 8.

Table 5.1. Consensus positions by forum

 

Forum

Consensus position

 

McGliggot Road (MCG)

Gratiot and River Roads Group I (GRR)

Gratiot and River Road Group 2 (GRR2)

Freeland Town (FLT)

Falcon View Estates (F VE)

Freeland Mobile Home (F MH)

Maple Street Group I (MSG l )

Maple Street Group 2 (MSG2)

Maple Street Group 3 (MSG3)

North Saginaw (NSA)
East Saginaw (ESA)

Open space corridor. Neutralize hot spots with
bioremediation. Test on GM property

Dredge hot spots in river & floodplain. Flood control.
Insure residents against illness. Cap low concentration areas

Voluntary relocation in high contamination areas. Wildlife
area. Do nothing elsewhere but with health monitoring

No forum held

Reduce exposure, flood control, open space, dredge hot
spots and wait for bioremediation for the floodplain when
available

Dredge hot spots. Catch & treat downstream flow with a
chemical solvent in a pond

Dry dredge and treat. Control flooding. Dow & GM
together

Heat treat hot spots, buy out people who want to go, dredge
after the river is clean

Dredge hot spots, ongoing testing and re dredging over
time

Control flooding. Dredge — dry. Bioremediation floodplain
No forum held

 

Table 5.2 shows the change in mean acceptance level of different clean-up

scenarios. Dredging had an initial aggregate score of 2.8 (just below 3, “neither

65

acceptable nor unacceptable” on the five point Likert Scale used to measure level of
acceptance of the alternative clean-up solutions). Following participation Table 5.2 shows
that the dredging increased by 0.3 overall, just above 3 “neither acceptable nor
unacceptable.” Asking people to decide in a group results in different decisions than
asking people individually in a survey. This is the case even when they had just
participated in a discussion in which the group came up with a different solution.

Table 5.2 shows which solutions were overall more or less acceptable before and
after participation. The distribution was greater after participation, with leaving
contaminated soils and sediments in place the most unacceptable (1.9 mean each) and use
of new technologies to neutralize dioxins, dredging, and banning agriculture the most

acceptable (3.3, 3.1, and 3.0 respectively).

Table 5.2 Change in mean level of acceptance of clean-up scenarios*

 

 

 

 

 

 

 

Action Pre- First Post
forum forum
mean mean

Move residents off of properties with dioxin contamination 2.6 2.4

Dredge or dig up and remove the contaminated soils and river 2 8 3 1

sediments ' '

Leave the contaminated soils and sediments but cover them

. . . 2.2 1.9

With a layer of clean sods and sedlments.

Ban all agricultural activities on properties with contaminated 2 8 3 0

soils ' '

Take no action. Leave the contaminants in place and let nature

. 2.2 1.9
take care of them over time.

Use new technologies such as the use of bacteria or heat to 3 4 3 7

breakdown and neutralize the dioxins.

 

*Scale: I very unacceptable, 2 unacceptable, 3 neither acceptable nor unacceptable, 4 acceptable, and 5
very acceptable; n=51

Table 5.3 maps some common thought pathways that led to acceptance of various

clean-up options. For example, moving residents was a relatively unpopular solution,

66

receiving a 2.4 mean acceptance after forum participation (Table 5.2). When discussed in
the forum people expressed that it wasn’t right to force people to move because they live
on property with contaminated soils. People expressed concern that people wouldn’t be
able to sell for a fair price because who would want to buy their property when it is
contaminated? When proposed that moving could be voluntary, some people agreed that
moving people would be okay. This is what is meant by resolution of tradeoffs. Other
tradeoffs may not be resolved, such as if the expense is too high or if Dow might leave
town because of the expense. Where this was the concern, and lef’t unresolved, moving
residents did not become a more acceptable solution. Table 3.9 is not an exhaustive list of
potential clean-up actions, their tradeoffs and resolution, but it does present many of the
most common downsides discussed for each clean-up scenario and helps to explain why
some scenarios after the forums were accepted while others were not. For example,
leaving dioxins in place had the downside of not reducing the risk of living in a
contaminated environment. There is no resolution of the problem. The most common
psychological response is to minimize the potential harm that living in a dioxin
contaminated environment might pose. The issue guide and forum leaders maintained
that the scientific evidence demonstrates that a dioxin contaminated environment at this
level likely poses risks, may have resulted in few participants expressing that doing

nothing and leaving contaminants in place is an acceptable solution.

67

Table 5.3 Examples of tradeoffs and resolutions

 

Action

Cross Pressure

Cross pressure resolution

 

Move residents

Expense too high, Dow
may leave

 

Infringe on people’s rights
to live where they wish

Make it voluntary

 

Dredge

Expensive

Yes, but clean up is worth the
cost

 

Expensive, Dow will have
layoffs or leave

They’ll probably shut down
anyway

 

Downstream
contamination or
contamination of
communities that live near
the landfill

If it can be done safely, than
okay.

 

Leave dioxins in place &
cover & or take no action
at all

Doesn’t reduce exposure

 

Ban agricultural activities
on contaminated land

(not discussed)

 

Use new technologies to
neutralize dioxins

Unintended consequences

Test on a small scale first

 

Knowledge change is the fourth variable examined to determine quality of the

group decisions. Data from two survey questions are presented in Tables 5.4 and 5.5.

According to these measures, in aggregate form only a slight increase in knowledge

following participation in a forum was observed. For example, Table 5.4 shows the

change in percentage of correct quiz answers. The biggest increase in correct answers

was in response to the true false question “Due to elevated dioxin levels, the Michigan

Department of Community Health has issued a fish advisory for the Tittabawassee River

Downstream of Midland” (correct answer is true). While these questions were asked to

assess the general knowledge level of forum participants and to see if they learned during

the forums, they do not provide the whole picture about learning during the forums. As

68

demonstrated in the open ended response questions about preferred solution, participants
demonstrated a great improvement in their ability to articulate the problem and their
preferred solution. This learning was best demonstrated on a case-by-case basis and not
in aggregate. The survey questions are presented here to demonstrate that the survey
attempted to quantify the learning, but was inadequate for capturing the learning that was
achieved. Participants were also asked in the surveys to assess their own knowledge
about the dioxin issue (Table 5.5). Participation in the forum resulted in an overall shift in
self-reported gains in knowledge. For example before the forum, 37% of participants
expressed that they felt they had a “low level of knowledge or understanding of the
Tittabawassee dioxin contamination issue” which was reduced to 12% after the forum.

See Table 5.5 for the changes measured using two survey questions.

Table 5.4 Change in percentage of correct quiz answers

 

 

 

 

 

Quiz question Pre- Post-
forum forum

Whether or not a chemical will cause harm to a person’s health 80 82

depends in part upon the amount of the chemical that enters the body.

Toxic substances such as dioxins may pose less of a health risk to 84 86

young children because of their smaller body size.

All people living across the United States have at least a very small 65 80

amount of dioxin in their bodies.

Due to elevated dioxin levels, the Michigan Department of 77 88

Community Health has issued a fish advisory for the Tittabawassee
River Downstream of Midland.

 

If a toxic substance is stored in a way that no one is exposed to it, the 39 27
risk of injury or illness is zero (i.e. none)

 

In the case of the Tittabawassee dioxin contamination, which 55 56
government agency is responsible for approving any dioxin clean-up

plan? (Options were Michigan Department of Environmental Quality,

US Environmental Protection Agency, Michigan Department of

Community Health, Local municipal government, or don’t know.)

 

N=51

69

Table 5.5 Change in knowledge self-assessment

 

 

 

 

 

 

 

 

Questionnaire Level of knowledge Pre Post forum

question forum % (n=51)
% (n=49)

“What is your current low level of knowledge 37 12

level of knowledge or medium level of knowledge 47 65

understanding of the ...high level of knowledge 16 24

Tittabawassee dioxin

contamination issue?” 1 00 100

Response: “1 have a...”

“Mark the box that best ...no knowledge 24 4

reflects your ...some knowledge 71 86

knowledge of dioxins ...a lot of knowledge 6 10

 

and their effects”

Response: “1 have . . . “ Total 100 100

 

The survey asked six questions about participants’ failings about the fairness of
the process (Table 5.6). According to the survey responses (Table 5.6) and the open
ended comment section, participants expressed near unanimity that the research
experience was fair. 100% of participants reported that they found participating in the
forum to be a positive experience; 94% answered “false” to the statement that “The
facilitator seemed biased and favored some solutions over others;” and so on as shown in
the table.

The lack of spread to the survey data on process fairness limited the opportunity
to observe the relationship of its influence on the assessment. If process fairness were
observed as a problem, it is expected that it would affect individual’s ability to come to
quality decisions. Thus, it would be expected to have as high an order of importance in
the assessment as adherence to consensus position and decision stability. However,

because concerns about fairness were not observed process fairness was calculated as a

70

lesser factor along with perception change, resolution of cross pressures, and knowledge
increase.

Process fairness is an area that would benefit from future study. Public
participation efforts are usually sponsored by an agency, not by a neutral third party, and
the trust relationships may not always be present. I suspect that because the research was
conducted by a university and sponsored by the US. EPA, we were able to operate with a
solid level of basic trust. If the forum were conducted by one of the major stakeholders
on the issue, I suspect that trust would have been a more influential variable.

Having a neutral third party, and having the effort as a research exercise,
probably contributed significantly to the high level of trust exhibited in the surveys.
Practitioners should take care to develop processes that support to the best of their ability
a fair process: protecting individual identity, checking power in the discussions through
the use of a facilitator, through transparency and responsiveness to participant’s concerns.
Because all of the groups expressed satisfaction with the fairness it is treated as a
prerequisite for forming a good group decision. Non participation may be interpreted as

resulting from potential participants’ uncertainty of process fairness.

71

Table 5.6 Process fairness

 

Fairness Variable % and (n)

 

Satisfaction with forum experience:
“Did you find participating in the issue forum to be a positive experience?”

Yes 100(51)
Neither positive nor negative 0
No 0
No answer 0

Neutrality of the Facilitator:
“The facilitator seemed biased and favored some solutions over others”

True 6(3)

False 94(48)
Opportunity to express their views:

“The facilitator made certain that all participants were given the opportunity to
express their views”

True 98(50)

False 2(1)
Respect from other participants:

“During the issue forum, my views were respected by other participants”

True 98(50)
False 0
No answer 2(1)

Neutrality of the information in the issue guide:
“The issue guide provided an unbiased overview of the Tittabawassee
contaminated sediments and soils issue”

True 88(45)

False 8(4)

No answer 4(2)
“.I found the issue guide difficult to read or understand”

True 0

False 96(49)

No answer 4(2)

 

Finally, decision stability was assessed by comparing respondent’s post-forum
preferred solution to their response to the same question one month after the forum. A

decision was stable if the response was the same before and afier the forum. The

72

summary of the stability measures for each focus group is shown in table 5.7. 61%

percent of the participants had stable decisions, 18% did not have stable decisions, and

22% did not provide enough information to determine whether their position was stable.

Table 5.7 Decision Stability

 

 

 

 

 

 

 

 

 

 

 

Focus group Decision Decision is Not enough Total n
is stable % not stable % information
(n) (n) n.a.
McGliggot Road (MCG) 100(8) 0(0) 0(0) 100(8)
Gratiot and River Roads Group 1 57(4) 0(0) 43(3) 100(7)
(GRRl)
Gratiot and River Roads Group 2 75(3) 25(1) 0 100(4)
(GRR2)
Falcon View Estates (FVE) 45(5) 36(4) 18(2) 100(11)
Freeland Mobile Home (FMH) 50(2) 0(0) 50(2) 100(4L
Maple Street Group 1 (MSG1)* 50(2) 25(1) 25(1) 100(4)
Maple Street Group Two 33(1) 33(1) 33(1) 100(3)
(MSG2)“
Maple Street Group 3 (MSG3)* 100(4) 0(0) 0 100(4)
North Saginaw (N SA) 33(2) 33(2) 33(2) 100(6)
Total 61(31) 18(9) 22(11) 101(51)

 

*Maple street groups 1, 2, and 3 were made up of participants from neighborhoods “Maple Street East” and
“Maple Street West”

In aggregate form, some key findings from the data are as follows:

1) 95% of forum participants felt the process was fair.

2) Few participants demonstrated gains in knowledge as measured by survey quiz
questions.

3) The spread of mean levels of acceptance of different clean-up scenarios
increased with some scenarios becoming more acceptable such as new technology,
dredging, and banning agricultural activities, and other actions becoming less acceptable
including taking no action, and leaving contaminated sediments and soils in place.

4) 61% of participants demonstrated stable decisions.

73

Group decision assessments

Earlier in this chapter the scale for assessing decisions by using the six factors that
characterize decisions was introduced. Factor scores were derived from comparing
participant responses before forum participation to their responses to the same question
afier the forums. As described above, group assessments were determined by tallying the
assessments for each individual in that group. The individual assessments were developed
by tallying the assessments of each factor for that individual. The first part of this chapter
also presents the rules for group assessments. The following section presents a discussion

of the group assessments organized by level of quality.

Good group decisions

Three of the nine focus groups produced “good” decisions. This was determined
on the basis of demonstration of individual positions consistent with the group position,
expression of process fairness, the stability of the group participants’ views over time,
participants’ demonstration of change in preferences and lasting gains in knowledge
following group participation. As explained in chapter 4, to receive the assessment of
good group decision, the majority of the individuals had to have a score of 1. This trend
was measured and reflected in the total factor proportion. Factor proportions amounting
to 80% or greater received a “good” assessment. Maple Street Group Three (MSG3),
McGliggot Road (MCG), and Gratiot and River Roads group 2 (GRR2) all were
determined to have good group decisions. The results for these three groups are
summarized in Tables 5.8 through 5.10.

Maple Street Group Three (MSG3) demonstrated the clearest case of reaching a

good group decision with all of the factors examined for a good decision reflected by all

74

of the participants. The consensus position reached by MSG3 was to “dredge just the hot
spots with ongoing testing and re—dredging if necessary over time” (see appendix). This
decision was a quality decision in part because there was 100% agreement among
participants on the preferred solution. Also contributing to the quality decision was that it
was unanimous and stable for group participants. None of the participants came to the
discussion with the position yet all participants came out of the forum agreeing with the
position and it persisted.

Another factor that contributed to the group decision being a good one was that
the tradeoffs of dredging that were identified by this group were incorporated into the
decision. The tradeoffs of dredging identified by MSG3 included: the expense given the
uncertainty of the health risk posed by dioxin; the pollution could be so widespread that
dredging would be ineffective; and, that dredging might increase the risk of exposure by
spreading dioxin. To address the expense of dredging given the uncertainty of its
effectiveness, the group decided to dredge just the hot spots. To address the concern that
dredging may be ineffective, they incorporated testing into the solution. The third
downside was not widely incorporated into the group position. However, #129 expresses
understanding that containment would likely ameliorate the problem because it would
“help keep the stirred up dioxins from getting into the air or further flowing down river,”

one of the identified downsides of dredging.

75

Table 5.8 Maple street group three (MSG3)

 

 

Consist-
. cut with De- Prefe- Trade- Know-
Falr . . . N/ Assess-
Case Process con- crsron Is rence offs ledge total men t
sensus stable changed resolved increased
position
94 l l l l l 1 6/6 1, good
1 l8 1 l l 1 Na 1 5/5 1, good
122 I l l l l 1 6/6 1, good
129 l l l l l 1 6/6 1, good
Factor 4,4
Pro- 4/4 4/4 4/4 4/4 3/3 4/4 6/6 ’
. good
portion

 

l=yes, observed, 0=not observed, 1/2 = partially observed.

Unlike Maple Street Group 3, McGliggott Road Group (shown in table 5.9) did
not have 100% “Yes” for every factor analyzed. However, the position reached by the
McCliggott Road Group (MCG) was assessed as a quality group decision because the
participants in MCG were in 100% agreement with the consensus position, for 75% of
participants the decision was stable over time, and the factor proportion for the additional
variables was greater than 50%. The consensus position for MCG was to turn the
floodplain into an “open space corridor” and following successful pre-testing, neutralize
the dioxin with a chemical treatment. Most of the participants maintained this position
over time. The other factors of preference change and increase in knowledge were
demonstrated my most MCG participants.

Furthermore, the MCG group position incorporated tradeoffs. The downside of
the chemical treatment voiced by the group was the uncertainty of the safety of the
treatment. This downside was addressed by deciding that more testing should be
conducted before applying the chemical treatment on a wide scale. #335 states “a

combination of open space corridor and neutralizing dioxins via chemicals after the

76

technology has been further developed and more is known about the side eflect's”
(emphasis added).

A downside of the open space corridor is the challenge inherent in relocating
residents. Participants neither raised nor resolved this particular downside, a sign that the
full impact of the solution may have been overlooked. Although MCG reached general
agreement on open space corridor and chemical treatment and 75% of participants
maintained their positions, some people returned, in part, to their original preferences
prior to participating in the forum. #331 indicated a preference for dredging prior to
participating in the forum, did not mention dredging immediately after the forum, but
added dredging into their preferred solution one month later in the final questionnaire.
#337 wrote “?” in the final questionnaire and #327 wrote “nothing should be done,” an
indication that the forum left some individuals uncertain or feeling that no action should
be taken, deviating from the consensus to treat the contaminated soil.

MCG selected the high tech solution as their group position, a position most
widely accepted by all of the preforum survey respondents (Table 5.2). The neighborhood
is a frequently flooded neighborhood and the issues inherent in relocating residents were
not fully discussed nor resolved. However, the decision to develop an open space corridor
and neutralize the hot spots following a testing period was widely accepted by the

majority of the participants over time, indicating an effective-enough group discussion.

77

Table 5.9 McCliggott Road (MCG)

 

 

 

Consist-
. ent with De- Prefe- Trade- Know-
Farr . . . N/ Assess-
Case Process con- crsron rs rence offs ledge total men t
sensus stable changed resolved increased
position
3 I3 I l I I I I 6/6 I, good
330 I I I I n.a. I 5/5 I, good
33 I I l I I I I 6/6 1, good
335 I I I I I I 6/6 1, good
342 I I I l I l 6/6 1, good
345 I l I I I 1 6/6 1, good
327 I I 0 n.a. 0 I 3/5 1/2 mid
337 0 I 0 I n.a. 0 2/5 1/2 mid
Fac‘or . 7/8 8/8 6/8 7/3 5/6 7/3 7/8, good
Proportion

 

l=yes, observed, 0=not observed, I/2 = partially observed.

Gratiot and River Roads Group Two (GRR2) developed a group decision that fit
the criteria for a “good group decision” (Table 5.13). 100% of the GRR2 participants
mentioned some type of flood control to address the dioxin contamination problem,
demonstrating good agreement with the consensus position. It is important to note that
the GR neighborhood is a frequently flooded area. Flood control would immediately
address a serious contamination problem for this community and therefore locally is not a
contentious clean-up choice.

The respondents reported little discussion of tradeoffs.

The group position of flood control persisted for more than 50% of GR Group
Two participants (3/4). Those for whom the position did not persist reverted back to their

original positions of “undecided” (#274) and “Total clean up” (#244).

78

A non-controversial solution of flood control yielded an easily agreed upon
consensus position. The added benefit of reducing water damage made up for the
possibility that dioxin is not harmful and thus treatment an unnecessary use of resources.

Table 5.10 Gratiot and River Roads Group two (GRR2)

 

 

 

Consist-
. ent with De- Prefe- Trade- Know-
Farr . . . N/ Assess-
Case Process con- clsron IS rence offs ledge total men t
sensus stable changed resolved increased
position
269 I I I I n.a. I 5/5 1, good
275 I I l I I I 6/6 1, good
244 I I I 0 n.a. 0 3/5 1, good
274 I I 0 I n.a. I 4/5 I/2 , mid
Facmr . 4/4 4/4 3/4 3/4 1/1 3/4 3.5/4, good
Proportion

 

I=yes, observed, 0=not observed, I/2 = partially observed.

Partially developed group decisions

Three groups, Falcon View Estates (FVE), North Saginaw (N SA), and Gratiot and
River Roads Group One (GRRI), had moderate quality decisions. Falcon view Estates
(F VE) and Gratiot and River Roads Group One (GRRl) were moderate quality decisions
because they had difficulty forming a group decision. North Saginaw (N SA), on the other
band, formed a group position; however the position was not stable for 50% of those who
returned the surveys.

As shown in the appendix, FVE did not arrive at a single group decision. Instead,
they developed a series of solutions and individuals mentioned components of the group
position. However, most of the individual preferences did not include the entire sequence
of the clean-up scenario. The complete sequence developed during the meeting was to
reduce exposure to people by controlling flooding and creating an open space corridor.
Those changes would be followed by dredging just the areas with high concentrations of

dioxins and then in the future, if determined safe, the group suggested applying

79

bioremediation techniques to remediate floodplain contamination. 18% (two participants
#4, and #5) expressed this complete sequence in the post forum questionnaire. 72%
expressed some component(s) of the sequence, but not the entire sequence. 18% reported
an entirely different position and 9% did not provide a response. The individual who did
not agree with the group position (#11) stated that no action should be taken.

By including so many clean-up scenarios, FVE group members built in
contingencies for the individual scenarios (resolution of tradeoffs). For example, reducing
exposure and moving people out of the contaminated areas would not get the dioxin out
of the environment, but only address the problem for people in the immediate vicinity.
Limiting the dredging to just the hot spots addresses the expense and invasiveness of
dredging the entire contaminated site. The unresolved downside however seems to be
whether dioxin is actually harmful enough to warrant action. This uncertainty was
unresolved preventing 27% (#4, #5, and #11) from fiJlly supporting the group position.
Other unresolved tradeoffs included the potential spread of dioxins from dredging (#14).

The position was stable for 64% of the participants who provided responses. Case
#4 reverted to his or her original position one month following the forum. Case #30 was
undecided a month later. #43 and #23 expressed one element of the group position
immediately following and a different component one month later.

In sum, the FVE group wrestled with the clean-up scenarios and developed an
approach that most participants agreed with. The position was not fully developed
because some of the participants concerns were not incorporated. In the end, this resulted
in a moderate quality group decision because there were several individuals who were not

in support of the group position. Falcon View Estates findings are reported in table 5.11.

80

Table 5.11 Falcon View Estates (F VE)

 

 

 

Consist-
. ent with De- Prefe- Trade- Know-
Falr . . . N/ Assess-
Case Process con- crsron rs rence offs ledge total men t
sensus stable changed resolved increased
position
5 I I I n.a. I 1 5/5 1, good
27 I I I I n.a. I 5/5 1, good
34 I I I I n.a. 1 5/5 1, good
I4 I I I 0 I I 5/6 1, good
23 I I 0 n.a. I I 4/5 1/2, mid
30 I I 0 I n.a. I 4/5 1/2, mid
43 I I 0 I n.a. I 4/5 1/2, mid
4 I I 0 I 0 I 4/6 I/2, mid
I I I 0 I 0 0 0 2/6 0, low
9 I n.a. n.a 0 n.a. I 2/3 n.a.
I6 I n.a. n.a n.a. n.a. 0 I/2 n.a.
Pam" . 1 1/1 1 8/9 5/9 5/8 3/5 9/1 I 6/9, mid
Proportion

 

I=yes, observed, Oznot observed, 1/2 = partially observed.

The position developed by the North Saginaw (N SA) focus group (table 5.12) was
determined to be a moderate quality group decision. There was agreement about the
solution. The participants who completed a questionnaire agreed on dry dredging as the
preferred solution. Another positive aspect of the group decision was that downsides to
dredging were alluded to in responses by several participants. For instance, #180’s
response includes concerns about the destruction by dredging. Less invasive treatment,
bioremediation was suggested for floodplain areas to not disturb homes and trees.
Another downside alluded to by #180 is the uncertainty of effectiveness of dredging. The
solution incorporated testing to determine if save levels have been achieved before
applying the treatment throughout the basin. Participant #221 identified another downside
to dredging: the health concerns of those conducting the clean-up. Incorporated into

his/her solution was providing protective gear.

81

However, not all participants’ concerns were adequately addressed. Participant
#176’s concerns about the effectiveness of dredging to address the problem were not
adequately addressed during the forum and his or her position returned to their original
position. The other participants who returned the final questionnaire and who provided
answers about clean-up maintained the group position one month following the forum, an
indication of quality group decisions. However lasting consensus was only expressed by
two of the North Saginaw participants. To be a higher quality position the other
participants also need to demonstrate evidence of long term learning and/or long term
preference change.

Table 5.12 North Saginaw (N SA)

 

 

 

Consist-
. ent with De- Prefe- Trade- Know-
Farr . . . N/ Assess-
Case Process con- crsron rs rence offs ledge total men t
sensus stable changed resolved increased
position
22I I I I I I I 6/6 1, good
180 I I I 0 I 1 5/6 1, good
I76 I I 0 I 0 1 4/6 I/2, mid
I96 I I 0 I n.a. 0 3/5 0, low
I 78 I I n.a. n.a. n.a. I 3/3 n.a.
222 I n.a. n.a. n.a. n.a. n.a. I/ I n.a.
Proportion 6/6 5/5 2/4 3/4 2/3 4/5 2.5/4, mid

 

I=yes, observed, 0=not observed, 1/2 = partially observed.

The decision that was developed by Gratiot and River Roads Group One (GRR)
(Table 5.13) was of moderate quality. 57% of GRRI participants agreed with the group
consensus to study the health impact of dioxin. However 28% provided no answer.
Furthermore, participant #267’s views were not included at all in the group decision.
#267 said they prefer dredging in the frequently flooded areas and to cover the non-flood
areas. Though this participant preferred dredging, a position which was not agreed upon
by the group, they did mention the group position of gathering more information in

additional comments section of the final questionnaire, reflecting some consistency with

82

the group position. There was also disagreement about moving people and dredging.
#226 contrasts with #245: #226 wrote, “Further study of the health impact before
moving” (emphasis added), while #245 wrote, “Immediately move people out of area
highly contaminated.” More time and further discussion may have been needed to agree
upon a timeline.

The downside to continued study as the solution, referred to by #245, is that while
the study is going on people continue to be exposed to a potentially dangerous
contaminant. The solution incorporated was to assist people with moving out of the area.
The downsides of action (as opposed to just study) were not explicitly stated by those
opposing action (#226 and #252). Further discussion and an attempt to resolve the
downsides of action may have helped #226 and #252 move closer to agreement about
taking action.

Although they didn’t form agreement, the positions were stable among 100% of
those participating. 50% (#226 and #245) demonstrated changed preferences. #245
entered the group discussion preferring testing, but exited the forum with desire for action
first, testing later. #252 and #267 did not change their views during participation.

GRRI did not come to group consensus. They started to discuss the issues and
identify solutions, but did not agree as a group on a solution. It is possible that the
polarized views in the neighborhood, as evidenced by presence of a lawsuit may make it

identifying a common solution more challenging in this neighborhood.

83

Table 5.13 Gratiot and River roads group one LGRRI)

 

 

 

Consist-
. ent with De- Prefe- Trade- Know-
Farr . . . N/ Assess-
Case Process con- crsron rs rence offs ledge total men t
sensus stable changed resolved increased
position
245 I I I I n.a. I 5/5 1, good
226 I 0 l I 0 I 4/6 1/2 Mid
267 I 0 I 0 n.a. I 3/4 I/2 Mid
252 I 0 I 0 0 0 2/6 0 Low
224 I n.a. n.a. n.a. n.a. n.a. I/ I n.a.
250 I n.a. n.a. n.a. n.a. n.a. I/ I n.a.
264 I I n.a. n.a. n.a. I 3/3 n.a.
Proportion 7/7 2/5 4/4 2/4 0/2 4/5 2/3 Mid

 

I=yes, observed, 0=not observed, I/2 = partially observed.

Low quality group decisions

No groups had low quality decisions.

Forums with too little information to assess

Three groups, Maple Street Group One (MSGl), Maple Street Group Two
(MSG2), and Freeland Mobile Home (FMH) were assessed as n.a. because 50% or more
of the individual assessments were scored n.a. (tables not shown) or there were too few

group participants to confidently assess the quality of the decision.

Summary of assessments and reasons for differences observed

Nine forums were held across three groupings of neighborhoods: frequently
flooded, City of Saginaw and Adjacent to Floodplain. Three of them did not have enough
information to assess the quality of the group decision; three groups had partially
developed decisions; and three groups were assessed as good quality decisions. Table
5.14 shows the pattern of quality of the group decisions in relationship to proximity to the
river. Location appears to have been a significant factor in the quality of the decisions

made. None of the frequently flooded neighborhoods had “not enough information” or

84

“no forum held". In fact, one neighborhood, Gratiot and River Roads (GRR), had so
much interest that two forums were held. Alternatively, the neighborhoods adjacent to the
floodplain did not produce any good decisions. The quality of the decisions that came out

of the City of Saginaw forums were spread across the range from good to no forum held.

 

Table 5.14 Quality ofjroup decisions based on proximi to river

 

 

 

 

 

 

Frequently City of Adjacent to
Flooded Saginaw floodplain
Good MCG (n=8) MSGB (n=4)
decisions GRR2 (n=4)
Partially GRR1(n=7) NSA (n=6) FVE (n=11)
developed
decisions
Not enough MSG1(n=4) FMH (n=4)
information MSG2 (n=4)
No forum held ESA (n=0) FLT (n=0)

 

 

 

 

 

In addition to location, it appears that participation is also related to the quality of
the decisions: the lesser the participation, either by not showing up, or not completing
surveys, the lesser the quality of the decisions. Not enough information (n.a.) could result
either from not enough people showing up to attend the forum or for incomplete answers
to survey questions. Participation, as discussed above seems to have a geographic
relationship. Neighborhoods where there was either no forum held or there was not
enough information to assess the forums came from the City of Saginaw neighborhoods
and the adjacent to the floodplain neighborhoods. The Saginaw neighborhoods are not

located on the Tittabawassee River and at the time of the study there was little popular

85

knowledge about the high levels of dioxin contamination in the Saginaw River, which is
adjacent to the surveyed neighborhoods.

Location and participation are likely related to salience: the closer one is to the
issue, by means of having dioxin contaminated soils flooding their basements, the more
likely one is to participate in a discussion about the issue. Some survey questions offer
further insight into the salience of the issue for residents: shown in tables 4.4 and 4.5.
They asked asking about participant’s knowledge about the issue. The findings support
the observation that fewer people from the adjacent to floodplain neighborhoods and
Saginaw neighborhoods participated than the frequently flooded neighborhoods: 86% of
survey respondents who said they’ve never heard of the issue were from the Saginaw
neighborhoods. In the Saginaw neighborhoods, survey respondents were more likely to
have heard about the issue from television or radio than from other sources. Another
survey question that provided insight into salience was the screen question (for details,
see footnote of Table 3.1). The frequently flooded neighborhoods had the highest rates of
being screened from participation due to their close affiliation with the issue, such as
being part of a lawsuit related to the issue against Dow or the DEQ. Based on survey
responses for “reasons offered for not attending,” (table 4.2) people living in the Saginaw
neighborhoods had more difficulty attending a forum. They more frequently reported that
they were caregivers, had to work, and couldn’t get out because of a disability. Most
extreme barriers to participation were observed from the East Saginaw Neighborhood
where residents had very low average education (Table 4.1).

Education and income were considered as possible explanations for the quality of

decisions and there appears to be some association. The fifth column in Table 4.1 shows

86

the income and education of the forum participants. The frequently flooded
neighborhoods had the highest incomes compared to the City of Saginaw neighborhoods
and the Adjacent to floodplain neighborhoods. The good quality decisions were
developed in forums where participants had the highest median incomes (>$35K), and
100% of participants had a high school diploma. MSGl had similar demographics but a
not enough information decision.

Freeland Town (FLT) was an exception. Freeland Town had high levels of
education and income, but low levels of participation (no forum held). They are adjacent
to the floodplain, potentially explaining their low levels of participation. However
respondents from FLT expressed high levels of awareness of the issue as compared to the
City of Saginaw neighborhoods (Table 4.5), only one person surveyed answered “no” to
“have you heard of the issue?” Also, with higher education levels and higher incomes,
the barriers present in attending were less in Freeland Town than for neighborhoods
where people have fewer resources such as in Saginaw. When surveyed residents from
that neighborhood reported higher than average comments that they “don’t like to get
involved in things” and that talking about the issue causes people anxiety. Those
comments suggest perhaps a lack of desire to confront an emotionally difficult problem.
Freeland Town is also the neighborhood in closest proximity to the City of Midland,
where the company responsible for the dioxin contamination is located. The desire to
avoid the problem to avoid conflict may be stronger among residents in the Freeland
Town neighborhood, especially since they do not flood and are therefore not likely to

face direct exposure to the contaminants.

87

Education, income and salience (including location) do not provide adequate
explanation about the partially developed (mid quality) decisions. These decisions came
from across all three neighborhood groupings (frequently flooded, City of Saginaw, and
adjacent to floodplain). The partially developed decisions also had higher numbers of
participants (n) (Table 5.14). One explanation for lower quality decisions may be re the
numbers: perhaps it is more difficult to come to consensus the more people trying to
make a decision. However, upon investigation, a comparison of the factor assessments
across groups does not show a pattern that it is lack of agreement on a decision that
caused the poor quality decision, although that does seem to be somewhat of a factor for
all three partially developed decisions. A clearer pattern is observed when making
individual factor assessment comparisons across the groups: the presence of 0’s for
certain factors signal lower quality individual decisions. All the individuals with good
decisions had no unresolved tradeoffs (0’s) and all individuals expressing unresolved
tradeoffs (0’s) had partially developed decisions. Similarly, for most individuals scored
0’s for knowledge gain or preference change, the individual score was less than “good.”
The exception was preference change. Several individuals scored a 0 for preference
change, but had an overall individual assessment of good.

Also, the presence of “blocking statements” may explain why some groups were
unable to develop quality decisions. Table 4.8 shows that with the exception of NSA,
forums with partially developed decisions (GRRI, NSA, FVE) had higher frequencies of
“anti-solution attitudes” (blocking statements) expressed by their participants on the
baseline survey. Blocking statements include 7 categories of comments expressing

negative attitudes for reaching a solution such as “nothing can be done” and “talking

88

about it causes stress.” Two of the three forums with partially developed decisions
(GRRI and F VE) had 11 and 12 “anti-solution” attitudes respectively. Alternatively, the
three groups with good decisions (MCG, GRR2, and MSG3) had far fewer “anti-
solution” attitudes with 6, 3, and 2 respectively.

The exception to this trend was North Saginaw (NSA) where only one anti
solution comment was expressed. The individual assessment table (5.12) and associated
quotes (Appendix) provides insight into why NSA was unable to develop a good
decision. The reason for the lower decision score for NSA was that for one participant
(#176) the decision wasn’t stable. He or she went back to their original view that
dredging is too difficult to be successfirl. Another participant (#196) was given a low
individual score because his or her final questionnaire response was not specific. If the
decision were more firm for these participants then this would have been a good quality
decision.

In summary, proximity to known contaminated areas, income and education,
resolution of tradeoffs, saliency, barriers to participation, and presence of “anti-solution”
attitudes, are some observed associations with low rates of participation and lower quality
decisions. Higher levels of salience, higher levels of income and education, mobility,
resolution of cross pressures, and lower levels of anti-solution attitudes, are associated

with higher response rates and quality decisions.

89

CHAPTER 6: CONCLUSION

Democratic governance depends on effective public participation. Contaminated.
sites clean-ups, siting of landfills, and location of new highways are among the more
challenging public policy problems. Engagement is needed, but working with the public
is often emotionally challenging due to the possible health impacts on the people living
near the site. Although risk managers have been advised to incorporate public
participation early and often when developing risk management policy, too often the
public involvement efforts fail to result in satisfying policy outcomes (Gregory 2002).
Some of the problems with existing options include the perceived amount of time and
energy required to engage the public, loss of control of the decision, and difficulty in
engaging the silent majority. Recent literature on public involvement suggests
practitioners employ deliberative methods and actively seek participation of
underrepresented populations. The larger study tested the National Issues Forums for
developing decisions about a contaminated site. This thesis assessed the quality of the
group decisions that resulted from the deliberations.

The first part determined the extent to which forum participants were
representative of the sampled neighborhoods. From the data collected and the low
numbers it is difficult to assess definitively, but it appears that people with certain
characteristics participated at higher rates than others. Those living in the contaminated
floodplain and those with more knowledge about the issue participated at higher rates
than those living outside of the contaminated area or who have no knowledge about the
issue. People who reported that they heard about the issue from Dow or DEQ mailings, or

from a public meeting (survey did not differentiate if the public meeting was about this

90

issue) also participated at disproportionately higher rates than residents reporting gaining
knowledge of the issue from other sources of information such as newspaper and word of
mouth. Alternatively, people who reported that they heard about the issue from TV or
radio participated at lower rates.

Some barriers to participation reported by survey respondents included lack of
time, need to stay home to be a caretaker, lack of transportation or other mobility
problems. Conflict avoidance and concern about openly sharing their views about what
should be done was also expressed by some residents. Two of nine neighborhoods had no
forum because no showed up both outside of the flooded areas. One neighborhood in the
frequently flooded area had two forums because so many residents expressed interest.
Survey respondents expressing blocking views when asked about what should be done or
other impacts participated at higher rates than respondents expressing altruistic concerns,
and problem solving attitudes such as views that dioxin harms the environment or
impacts land uses, or that the solution is to raise awareness about the issue or regulate the
problem. Participants were more likely than non participants to be opposed to dredging.

To assess the quality of the group decisions, a decision tool was built with 6
factors identified in the literature from Yankelovich, Gregory, Jasanof, Slovic, Webler
and Tuler. Three focus groups developed good decisions, three came to partially
developed group decisions, and three groups did not have enough information for a
complete assessment. The decisions from the frequently flooded neighborhood forums
tended to be of good quality while decisions from the adjacent to floodplain
neighborhood forums tended to be of low quality or were not held. The decisions from

the Saginaw neighborhood forums were spread across the quality spectrum.

91

The assessment tool combined individual factor scores of 1 (yes), 0 (no) and n.a.
(not enough information). Consensus or personal agreement with the group decision, and
stability of the decision were the two major factors contributing to a decision assessment.
Despite a mathematical possibility in the assessment, certain lesser factors (consensus,
knowledge change, preference change, and resolution of cross pressures) signaled lack of
developed individual decisions when the factors were assessed 0 (for “no”). For example,
where tradeoffs are unresolved, individuals did not learn during the forum individuals do
not develop quality decisions. The expression of blocking views also signaled difficulty
reaching agreement.

One implication of this finding is that the decision assessment tool can be useful
for helping practitioners charged with helping groups make decisions, identify issues that
need further exploration. For example, where “0’s” were observed for resolution of
tradeoffs, practitioners can look back at the data and record the critical concerns that need
to be overcome before participants may be ready to come to a decision. In this case, the
forums raised a number of critical concerns regarding action to clean up the
Tittabawassee River and floodplain. For example participants were sometimes blocked
from coming to a lasting decision to dredge because of their concern about contaminating
downstream residents. On the other hand, participants were blocked from a comfortable
decision of doing nothing because of unresolved concerns about people remaining
exposed to dioxin. This type of analysis can yield valuable knowledge about the public’s
views. It is significant for practitioners also that not a lot of data must be gathered to get

quality insight. Simply asking an open ended question (important) about people’s

92

preferred solution before, after and one month following their involvement in a
deliberation can provide useful information.

One of the challenges faced by risk managers that this study attempted to find a
solution to is that of engaging and learning about the interests of the “silent majority.”
Personal, face to face invitations to forums being held in their neighborhoods at a time
convenient to them, and offering a $25 gift card, did not appear to be enough to get those
other than for whom the issue is most salient, to participate. One risk of engaging a non-
representative sample (people with views differing from the general population) is that it
is difficult to determine if the outcomes will serve the interests of those who did not
participate. The implication for future research is that more it may be necessary to do
more to attract people to participate. It may be important to raise awareness of the
problem in areas with lower literacy and higher poverty rates. Providing child care,
transportation, or recruitment by community members, known and trusted by residents
may also yield additional participation for underrepresented groups. Perhaps most
importantly for improving citizen participation in decision making is for the agency to
demonstrate taking the citizens views and preferred solutions into serious account in
deciding a clean-up plan of action.

Finally, settling on assessment criteria can be difficult because assessing
outcomes is inherently subjective. One implication of this tool is that it could be
replicated and used in other settings to evaluate group decisions and identify issues that
need further definition. The tool can help groups know when deliberations can cease (a
stable agreement is found), and take action on the chosen course. On the other hand,

when a decision is poorer, it can help practitioners identify those issues and attitudes that

93

need work to come to a decision. The assessment tool could possibly be adapted by
interchange other factors that the researcher or practitioner may wish to include. One
implication is that practitioners wishing to use an abbreviated method to reduce
transaction costs associated with group decision making, can reduce their data collection
to just one open ended question asked at three points in time.

Agencies need manageable public involvement methods that yield quality group
decisions. The engagement approach outlined in this thesis and the group decision
assessment tool gains ground toward the goal of developing methods that are usefirl for
public involvement in risk assessment and other difficult collective decision making

scenarios.

Research limitations & validity

A limitation of the study is that even though forum participants were reasonably
representative of their neighborhoods’ population, a representative sample does not imply
that the general public will be satisfied with the conclusions of the citizens in the issue
forums. Participating in the process may be a critical component to resolving cross-
pressures, so others may not be satisfied with the decision if they did not go the decision
making process, because they themselves did not think through the tradeoffs between the
various scenarios to come to group consensus.

The research artificially excludes special interest groups from the dialogue. In a
real clean-up decision making process, special interest groups would likely participate
and from this research we can’t predict how more polarized activist participants would

have affected the decision outcomes. Our design deviated from what would likely occur

94

in an open process where more powerful stakeholders would have access to manipulating
the situation.

The purpose is to gain insight about the effectiveness of the NIF for cases where
the tradeoffs are difficult and decision-makers wish to gain reliable information about the
affected public’s interests and concems. Generalization to a particular population is not
possible with purposeful sampling and small samples (Patton, 2002). From this study we
can explore the variation within this community, investigating the nuance and potential
patterns within the case.

Credibility of the research conclusions is sought through rigor of data collection
and analysis, triangulation of data and analysis, and by making researcher biases explicit.
During data collection, an effort was made to minimize reactivity. During the survey data
collection, the survey instruments remained the same and surveyors spoke with each
other to check that they were conveying the same information in the same manner.
Possible influences of differences in appearances such as gender, age, or race could not
be controlled for, but an attempt was made to reduce variation by wearing uniform
University shirts, wearing identifying name badges, and approaching homeowners in a
friendly and professional manner. The affiliation with Michigan State University aided in
being perceived by residents as a neutral and credible organization.

Different residents perceived surveyors in different ways: a few residents
expressed suspicion that surveyors were undercover spies for Dow or that we represented
environmental groups. Neutrality was attempted, but residents’ preconceptions affected
their reaction to researchers. For example, during forums, maintaining the neutrality of

the facilitators was attempted by not voicing personal opinions, and attempting to treat all

95

responses the same. Similarly, when administering surveys, surveyors avoided
expressing their opinions or validating the views of respondents. An attempt was made to
minimize participant reactivity to the survey by asking questions worded in the same
fashion at different times and asking the same questions of all participants.

In addition to these measures seeking to minimize participant reactivity,
researchers followed rigorous and systematic data collection procedures. Issue forum
groups were audio and video tape recorded and the facilitator used the same semi-

structured format for each issue forum.

96

APPENDIX

400.0 5 2.me 3202.2. mega: 0:080

 

93:0 00210:" 9.030 :3 «003. 260.3 82.3 0:0 3.9.26.3 0<2 :30

 

 

 

 

 

Mmmm $02.00 5309.20? 2H2 H.033 0:0 30:3 22
_ 35x 50< 0:050 00:00:56 0: :0H
Who—“mmnflmmfioflw .H = $0 00on - 226.0 503 0 0% 3:00. 30: . . .
0.0 m 2: 03.: m 52. 520: 8 000 a 50 _0<0_0 200 0<2 m 3:050 50 :2 00on 0:0 20850 503
00 0308200 :00? 20.0 00:00 2 :30 0:0 : 30< 00:; 50: 5. 0:0 000 a 520 .0 m:< 200. ._.:0: no :03
:m<0 H0 0 0 00305.8 00305.8 320 - ocH _ 55x 2 <0: :2 H30 520.
0005 z. ”2 00on 0<2 0:0 0<2 50 60.0 20:5
60
0.00: z :0. <<:0H0<2 .H 900 0 :0H 00on 0:0 .2 :220 mem :0
:m 8.2 8 002 z :0. 85.00. 9080 so :04 00040.
_H 00030 8 30 52 30 002 00.5.0:
52 2080 30 20mm ":2 0003 H0 £20 £050 00 0 00320350 8 :06 2.0520
HNN 25.00. ":0 :6302 00:00:42.0:0 0:0 06% 30 000<0 302.0:00 H0020. 0:0: mm .20.
3000 20mm 0<2 m 02.00 2 :30 226.8 50 2000 2 H30 :<2 50... 20
030.0200 8 00 30 302 00:500.
00:00::20 0: H30 :32 0006... 950
0030 8:: 2 00305303 200 0: 50 :9
30:22 0: 00300300 0: 10622 30 :2 0.0on 52 £050 00 00on H0 :06. .000 30. 2.200 :0 0.05:0
.30 ":0 :<2 is 000000. 5520 <0: 0050 000 30 302 0202020 2.03 003:0 58 30 m: 2 83:2 2258

 

30:22 30 «00.0030

 

0:006.

 

00S: :<2. 2230 50 :6:
030032.30 0: m <02? 2 502?
.05. 0:0 «0020 202000 _: 3000 2000

$520 050%000 £050 00 302 3200020.

 

97

400.0 >0 2.00 040002.00 moan: 0:080
0300100000? 000: 00000 00300... 330.00 :0. 00100 006300.303. .00. 0: 02. 0600a 00.20 00382.00 000.0

 

 

 

 

 

 

 

0000 #3 m000...:0 3300.030? 0.401033 0:0 303: .00:
0033103603 :00 8 :9... 00<0.00303. .0. :0 :05.
”muwmwmwmoflflwwmfifiwmg :0300. 0030. 000000 0300 00. 0000.00 <3: :05.
0003 : 0a 0 .00 0.. 003 .0 > 00303000: 0. :0:¢0..~.:0 3:0 00.. 0:0 0:03.00. #003303 0.. 30 00.. 0:0 200.. 53.0: .2... 00
w. w o 000 x - £0. a 0 3 0:0: 0 3... 0 : 0 ~22 s3: :05. 6030.00.00. 0:03.000 0<0..00_0 .: 0 .05 .000 .0 3:0 300. 000. 0200320
:_0«0 00.2 o\o .: 0.0:: 0 0:0 0:0 0: 000: 00000 00300: 30300 0. 000.30 53: 3:0 030.03. 203 0030
.2030.“ 30.30.3000: 0:203... .: ”:0 0000 0:0 000.0 0000
. 00000 0.. 0030 300.0030 .30 030 0200.00.
. . . . m.3:0.. .03 :30 0.00: 3:0 080.03 .5000
wmu Wm...“ .mofithmMWMwnw: HMS..N.:0 0:03.000 00: 00 08<0: 23:30 0:050 00 00:0.
00305000: 0.. 00302 0:0 30300000
wwo 00:4 .30.... 300330 0. 0:03.000 0:0 3.062002030 20350. 00300“ 53: 00.000: "000:0 0.. 00.00.00. 0:0
.: 030.. 0300 20. ..30 003:0 s3: :.0: 0:03.00. 00033030 8 000.?
03003000?
. 00..0<00 <3: 30.0 .303000: 0: 0.00 04000 3:03 3:0
. . . . . . . 0:03.00. 0.00: :0 .2000 00 0003. ._.:0: 0030.: 0:0
ww. 0.00: ...0 ":0 30... 0.00030 0:03.00. ...3: 320 58:30:09 «00.000 ":0 00.. 0303.1: “00. 003.0. 0600 530.0
000.0 0030 : 00303 <3: ":0 00.. 0<0Q 000..
> 00303000: 0." 000: 00000 00302 0:0
0000.0 0:050 00 30.0 :030..~.:0 0.90:0 <.0 0:03.000 0:2 . . . .
wwm 0.206 0.. 3:0 00:0.. 00:4 ":0 ”00:300.. :00 000: 33:0. <<03.:0:c0d._ 3% "”3300" "mwwfidm3 :0w. 000: 00:03
.30.: 0050000 0:0 3030 .0 .302: 0003 ":0 «0000.0 0: 0: 000 < _ 0 0 08 03.
0.00 0.4000.
:0<0 :00303 00:00 0<<0<
.403 .:0 :<0: Z0<0.. 0030
03.330 000.? ._.:0<.<0 00.
uwu .0 03<0 3:03 .0 0: 00:0. 0: 03. 00. 0 0 I z>

 

...:0< 00:." :0<0 03. .000.
00s. .0..00 ":0 :<0: <<00 00
000 .3 £00.03 .400N0.

 

 

 

98

400.0 >0 00......
0.000 0000.03 000: 00000 00300.. :0:..0..~0 :0E000 003300.003.“ .00. 0: 030.000.... 00.0.0 :030 2.00 000.0

 

 

 

0000 0. 0000.30 33030.0? 0.40.. .033 0:0 30:... 0.0..
_ .:.:x .:0 0.00 0:050 00 0.00:00 :0. :030 .:0
.003 .. 00. 00 3330. 2033.030 0:03.000 0. :00. 0.0:0 5...: 30300000 03:00.0. .:0 :00 0. .:0.30_ $030.00..
000 00:.. 00.0.0. .0. :0.c.0 .0x0 000: 00000 00300. ...00 0:0 0000.02 ..000 .0 333.00 .:0 0.90:0 0. 0000.0... 0:03.00.
003 0. .. .: ..30 00:..0. 30000.00. :0:..0_.N.:0 00030. _. .:000 0000.0._...00 :0<0 000:
.0300 0:0 :0<0 ...._0 0. :0 003.00 0.00 0.4000.
00.0:000 :030 00 00: 303 .33 830330.00 0.00
00... ...00. 0.. :0300. 003. 00... z 0:. 3.0 0 0. 0x.: 5...: 000: 00000 00300.. 0. 0 0:00 .0 00. 0 00300300 :030 _: 0 :0:-0..00.00

 

0.00: :3.

 

 

0.00. N- .0. 0.00: :0 0.. 0.000000 00.:..0:0 :000 .0
00 30030300.

 

99

...00.0 >0 0mm» 03.02.00 00.5.0: 000.00
0.000 000.033 0.0000 00. 000.0 3 ..<0. 0. 00000.03. ..000 00:20.. 30:3 30.00:.0 000.:0. ..3000. 000 .02
00:00:..0..0: 0.000

 

 

 

 

 

0000 0. 0000.30 3300.03? 0.0.0: .033 000 300.0 .03.,
.u .000 0033.. 5.030... 05:00..
. 00.3: 0. 30:3:00 0. 00.00300.
rdwmjmowmm. finnfifim 0. 00.0 0< 00s. .. :0. .:.0.00.00 3 0.00: :0 0.00 .03. 0.000 0.. 330<30
000 00 0.0 0. 0 0 «..0 0:. 0:. 0 305:0. 00:..0. 0. 20000.03 :00 00:.03.:0..0:. 0.000 00030 .0 0.00 20.0.
.00 gm. . 0 0 5...: 0 8300:0082 000x000 .0 031.02 .00. 200.0 00000 300:.0330..0:.
. :0. 0.30. 0 00.00:.0 <<0< 0.
03000333030.
0.00: :0 00030 0.300. 20.2 .00: .. E0 03 0:3 0. 0:0 .
000 3000033. 00:.. .392 30c... m0.0:._0: 00:00.
O0..00..<0 0.003 300... 00:..0. ...00030 0:0 0300.:0 0.0.. .
~00 00:0.:0.0:. 00:.. .302 00:00:..0..0: 0.000. 3000.000.
0030.0.0 0.00: :0. 0.00
00.2 ..03 03.030
000 03.330. .. 0000:.. 0.000 0033. 0:0 0.00: :0 ._.0.0_ 0.00: :0.

 

30x0 00:00 .00. 00$.
00: 0.... 0:30.

 

 

 

100

...00.0 >0 .u<m 03.02.00 00.5.0: 0:030
03:0 000.033 30:00 0x000:3. ..000 0033.. 000: 00000. 0.0000 :0. 000.0 0:0 5.0.. .0. 0.0.0300.0..0: .0. .:0
00000.03 530: 0330.0

 

 

 

 

0000 0 0000.30 3300.03... 0.3.. .033 0:0 30...: .03.
> 0.00 0.. 0.00 0.00000
0.0.30 5...: 0.000 0.0.:
. . . .<.0:00030:. 0.0030030 .0 303.... 5.3.0 305 0x.0. 0:0 30.3 0.4030 .0
A ”“4“” 0 ”M0JM030..0._..0..._0..“MM0 :0<0. 00.000. >00000 0<0.0 0x000:3 .0 303. 00:33 .0 .00:0
00:0 ...0. 0 : 0 0. . 0.0::30 0:0 .30.... .0 .00::0.00.00.-0.0_00.00_ 3000.0: 30.:000 .0
. . 3303. 0. 0.9.30 0. :0003 0.90: 00:.03.:0..0:.
300000.... 003030
5..0.03300.0..0: 0.00000.
.. .. :00 .0 00 33030 ..03 .:0 :3. 00.83 _
. . . 5.050 ...3 .0 000 .:0 :3. 30.30.00 .0 .:0
moire ”mmn:000w.<w0n.v_mm< 0.. X02333 .020. 0< 0.0030 0 00:0. 0:0 0 .05.
0mm3.000. .0 0: 0 : : 0.. : ..< 0 . 0.30 5:0: .:0 5.0.0. ..05. .0 0.300. a. 3 .:0
m .. <0: 03000 2:03 00 <0: 0 : 0 0::. .. 05. 0.. .0 ..<m. .: 0:330.. .30: .:0 :3. 0:0 20000.03 00:.0
0:. ..0 0:330. .30: 0: 00:. 0 00 02 0.00000. _ 0.00 33.. :05. 3030.00.00
0. 0 0 : : .: 0. 0 «0.0 0 :0 0 0 :. . 0. 00:.0 003 0:. 0.00.030 3 .:0 :00. .:.:3 0:0
.. 00 0 .0.“: 0 003 :0 0 .0. 0. 30:05 0:0 :0. :0<0 .0 0.0000
0 . 0. 303 0000.0.
. . 0:000 02 0300.:0 00 .0:0 00 0300:..0:0
_ 0:000 0... 030030 0:. .
.0 03000 00.:0 00.0 .0 .30. .:0 00.. 00 E03 .03: .0 .300 .:0 0:0. 0:0 0:.. ..03

 

 

.. 305 00 :000 000.:

 

0.05.30 0330 0:0 00:0.:0 303 0.00.030.

 

101

...00.0 >0 0030
03:0 00033:. 30:00 0x000:.0. ..000 0033.. 000: 00000. 0.0000 :0. 000.0 0:0 5.0.. .0. 0.0.0300.0..0: .0. 30
..0000.03 5.3: 03:00.0

 

 

 

 

 

 

 

 

 

 

0000... 00.03 .033 >3. 0:0 303: .03..
0. 0000.30 3300.03... 0.3. .033 0:0 303: .03.
0.. <<:0.0<0. .0 :00000. 0.0.0300.0..0: .......%..mw0 0. 3.030.002030 .0 0.00.. 005.: 30
20330 00: 00. 0.00 30 . 0:.000 .. .0 0 <0.< 0.00. :00_.: 0.00.03 . 33..
.. 00:00. ..003 3300 00 30< 0.0 03335900050 00 .0.. 0.0:0.
. _ 00:. 33.. 303 .0 ..:0. 0:0 .
.0 00:. ..:05. 00.33:. ._.:0. _. 00: 00 0.00:00 :0.
...:0< :000 .0 00 0.3.30
wo:....w.w..0 Immw.00.WM<0<:00_MMM.M 0.0000 :0 0.. 003033030 00..0 0:0
300 00:...0 00 3 0 : 5...: 53 0 2.0.0 0.:0< :0000 .0 00 00:0. 000.3030. 0.00000 0.3.0 5003 3 0 00.0
mm .0 .M< 0. 00 30 30403 .. < 0000.0 0350 0.0< 0<<0< .33 0.30330 5.0<. 00 :0. .0 00303303 0:<
30.0030 00:. 00 0 0 003.0330 30 0.00. 0.:0. 030. .. 00s. 00:.0 5.0... 0: 0 0:03.00.
. . . 5.0 5. 0 3 0. 0 0: 0. . 0: 0 00.3.0: 5.0.0: 00:.0 0.00: :0 30 :3..
0.00: 5.0.0..
_..0.0 0. 30.30. 0000.0 0: 30 .
00 ..3. 03:3 00 30.0 ”.00..“0M3 :3. 0:0 0.< 0
33.30. 0 .
00 00:.. _.:05.. 0.0.0300.0..0: 0.< 030030.
0 .. 0.0 _. :0 :05. 0000 00 <0: 20 . 33.. 0.000.:0 .0 30 5.0< .0 00.

 

0%...

 

 

 

102

...00.0 0> 20> 033.30 00.33: 0330
0.0:0 000.03? 0033. .3003? 0.. 0.00%0. 0.0.0300.0..0: 0: .00 00000.03.

 

 

 

 

 

 

 

0000 .0 0000.30 3300.03... 0.3. 32.3 0.3 303... .03.
0... 0.000.:0 0:0 00:..0. .00
3000.030. 0... 0.00030 0:... ..
<<:03 .0 .. 00330 .33... 030 .0030 03.. m 30300.0 0:00.. . . . .
.mo :05. 5.0.0 0. 0.00000. 0:0 003.. 3. m ...0 .00: 0:::0.. >_00 “mmfiwwgco: 0. 030030 5...: :05.
0.00: :0 5.00. .0 .003. 0.0.0300.0..0: 3. .00 00.. .0 0..
003 .00 .0300 .00. 0.0 0.300.
033333300. .300. 0.0.
. . 0... 0.000.:0 .. .00 0000.0 33.30 3 .00
mm. M00 0:30.:0. 3.3. 5.03.. 0.. 0.000.:0. 0... 0.000.:0 00: 00 000.00 :0 3 0330.3
00: :0 ..000. P00.
2.0..0 0:3 .003 .0 80$ .30 0.. .00
.00 0 . 03000.00 ..3. 0:0 0:. 0 0030. 0.00.03 005. 0003.00. :00 00:000 0:00
0:. 0:05.. . . . . .
_: 00.5.00: 0:0 0:. 00:0 03.... 00 _: 0:30:0. 0030.0. 5.03.. 00..
030.030 5.0 0.0 :0<30 0.0::0 00.0.
..003 5.0.. 0:0:0: 0.0:0 0:0 30.3 0000.0
..:0.03:0 .0 .03.. .00. 00: 0030 .500 0. ..000 0033.. 0.. 05.03 0. .00 00:03: 00 .00. 00:.0 000000
...0 00. .. 0:. 0. .003. 0....30 .. 0.000.:0 0:0 0.0.00.00. 0. 5.00.00. .0 ..3 0: 0. :00. 0. :00 .00 0.00
:0 5.... 00.000 .. .0 0.00. 0003.00. :00 3 0: 000:. 3. 30303:. . .03.. .00 3303. .000 .0
..0..00. 00.003. 0300.... 30000.03 00000.0... 3 .00 00000.03
0.000.
00303 .. 0:0 0.00 .. .33 00.:0
..m 20 0:05.00 3.30..
<<:003. .0 0. 3:... 005. 0:0
mm» 0.03 003.:303. 0:050 20

 

30. .00 0... .0 0.00: .. :0.

 

 

 

103

...00.0 >.. 0.00.. 033.30 00.53: 0:030
03:0 000.03? 3.33... 3.00000: 3 :.0: 003033000: 0300. 5.00.0.0 0.00. 00 :0330 0.005.003 0:. 00 30:..0.

 

 

 

 

 

 

 

 

:000:
0000 .0 0000.30 3303.03... 0.3. 333 0:0 30:... .03.
0:30. 030.. .0 000 .. .003 .0 0 0.00 00.. 0. 00:00. .0
mum 0 0:.. .3 0 5. 0:30. 0.3.. 0. .00 000.3 .3000. 003.0 0000.0. _. 30.0 .0 0 ..0... .00: 0000.00. 0000.0.0 0.00:
. 30<.:0. 0.000.:0. 0.0. :0 0. 8300:0000: .0 .0000 0.3030 00. 3.30.0 503
::.:000.0 0.00 3:0.
3300.03... 303 0000.0 0:. o. 0.00
. . . 0.00... 00:.03.:030 0:0 0.. 0300.... 0:.. 0:. 0300.... 3.000003 0:0 30:03. 02:00:00.3:
N00 “Mm.” mwmmmmflfimflfiflmmmw. 000:.0 00 .0030 302.03.... 0. 0.33:. 0:0 003.030 000.3 0.3.. 0. 30.0030 0:0 .00..
. 000.00. 00 000.3 030.00 .00.... 0.. 00:03:.
30.00:.0 0:0 00.00.30.
..:0 5.0.. .00.. 0.0 00.00 .0 0..
..00. - - 00.:0 .00 080... .
mum ”wwfiwhwfimm ”mafia.“ ”NW. m<0.:03 000.3 :0.. 0: 0000.0 0.0.. <<00. 2000 30.0 0:05.03 .0 005. 0000.0 0:0 5.00.0.0 0.0
00. 0 ..00. 3 .0 0 0... 0. 0 .0 0. .03. 00: 5.0 ..3 503.. <<0.. 0:0 000. 0.3030.
30.. 0.35.00 30 30:..0. _
000.03 0.. .00.. 000.0000.
. . . .oo ... 0000 0.0.: .0 00 .0... 0.0:0. <00... ...0 .0303 003033000: 3 000030 0.000 0:0 003.
~00 0.00: _. :0. 10030: 0000:. 00000.03 .0 00 0.00:00 .5 :0:-0000 0.000.
.00.. 5.0:. .0.. ..0: .. :0. 0:.0
NNO 5.03 003033030. 00:. 20 20
..:05..
_ 000.. .03.. .00.. 00: 00
03.330. _. 5.0:.0 00 .00
N00 000.... <<00. 0.0 .00.. 00.:0 .0 20 20
00. 303 :0 0:....
N00 .503 0.3.00. 00. 0x00 .03. 00300. 3300. m:0:3 000.3 030.00.

 

0:05.003. 00:. 0:0 005..

 

053.0 0.0 0030.030

 

 

104

BIBLIOGRAPHY

Abel, T. D., & Stephan, M. (2000). The limits of civic environmentalism. American
Behavioral Scientist, 44(4), 614-628.

Abelson, J ., Eyles, J., McLeod, C. B., Collins, P., McMullan, C., & Forest, P. G. (2003).
Does deliberation make a difference? Results from a citizens panel study of health
goals priority setting. Health Policy, 66(1), 95-106.

Amstein, S. R. (1968). A ladder of citizen participation. Journal of the American institute
of planners, 35(4), 216-225.

Arvai, J., & Gregory, R. (2003). Testing alternative decision approaches for identifying
cleanup priorities at contaminated sites. Environmental Science & Technology,
37(9), 1469-1476.

 

Beierle, T. C. (2002). The quality of stakeholder-based decisions. Risk Analysis, 22(4),
739-749.

Council, N. R. (1996). Understanding Risk: Informing Decisions in a Democratic
Society. Washington, DC: National Academy Press.

Diebel, A. Chibucos, M. & Beckwith, J. (2005). Dioxin Contamination of the
Tittabawasssee River and Flood Plain: Overview of Issue Forum Outcomes. East
Lansing: Michigan State University.

EPA, M. S. U. U. (2003). Dioxin Contamination and the Tittabawassee River and
Floodplain: Managing the Risks: An Issue Guide.

Fischer, F. (2000). Citizens, Experts, and the Environment: The politics of local
knowledge. Durham and London: Duke University Press.

Gregory, R. (2000). Using stakeholder values to make smarter environmental decisions.
Environment, 42(5), 34-44.

Gregory, R. S. (2002). Incorporating Value Trade-offs into Community-Based
Environmental Risk Decisions. Environmental Values, 1 I, 461-488.

Habermas, J . (2006). Political Communication in Media Society: Does Democracy Still
Enjoy an Epistemic Dimension? The Impact of Normative Theory on Empirical
Research. Communication Theory, 16(4), 411-426.

Jasanoff, S. (1991). Acceptable evidence in a pluralistic society. In D. G. Mayo & R. D.
Hollander (Eds), Acceptable Evidence: Science and Values in Risk Management
(pp. 29-47). New York: Oxford University Press.

105

Kelly, T. (2004). Unlocking the Iron Cage Public Administration in the Deliberative
Democratic Theory of J iirgen Habermas. Administration & Society, 36(1), 38-61.

Kothari, B. C. a. U. (Ed). (2001). Participation the new tyranny? London: Zed Books.

Lachapelle, P., McCool, S., & Patterson, M. (2003). Barriers to effective natural resource
planning in a "messy" world. Society & Natural Resources, 16(6), 473-490.

Mathews, D. (1994). Politics for People. Urbana: University of Illinois Press.

Mathews, D., & McAfee, N. (2000). Making choices together: the power of public
deliberation. Kettering, OH: Ruffolo Press.

McComas, K. C. W. S. (1998). Reassessing Public Meetings as Participation in Risk
Management Decisions. Risk: Health, Safety & Environment, 9, 347-360.

Patton, M. Q. (2002). Qualitative Research & Evaluation Methods (3 ed.). Thousand
Oaks: Sage Publications.

Pellizzoni, L. (2003). Uncertainty and Participatory Democracy. Environmental Values,
12, 195-224.

Putnam, R. D. (1995). Bowling Alone: America's Declining Social Capital. Journal of
Democracy, 6(1), 65-78.

Renn, O. W., Wiedemann, P. (Ed.). (1995). Fairness and competence in citizen
participation: Evaluating models for environmental discourse. Dordrecht:
Kluwer Academic Publishers.

Slovic, P. (1997). Public perception of risk. Journal of Environmental Health, 59(9), 22-
23.

Smith, G. (2003). Deliberative democracy and the environment. London, New York:
Routledge.

Spano, S. (2001). Public Dialogue and Participatory Democracy. Cresskill, New Jersey:
Hampton Press, Inc.

Thompson, P. (2004). Seminar on Ethical Considerations in Risk Assessment.

Webler, T., & Tuler, S. (2000). Fairness and competence in citizen participation:
theoretical reflections from a case study. Administration & Society, 32(5), 566-
595.

Webler, T., Tuler, S., & Krueger, R. (2001). What is a good public participation process?
Five perspectives from the public. Environmental Management, 27(3), 435-450.

Yankelovich, D. (1991). Coming to public judgment: making democracy work in a
complex world (1 st ed.). Syracuse: Syracuse Universrty Press.

106

   

utillliltiifllflljlllllflljitl