r11... :1. . 1.. ‘33

.7: 9.3%...

 

3.

u 1“ .Q~.‘.oo&~

M‘\l:ll.r‘4-l

 

 

 

513$ ..-...,..-..£.. La. ... . . ,. .. . . ..r..........1.. af..w§a§§¥é§a§

 

 

7...... lllllllllllllllllllllllllllll||lll|||llllllllllllllllllllllll

l 3 1293 02058

(7‘
A ‘0
\ J
C?)

This is to certify that the

thesis entitled

THE EFFECT OF AMBIGUITY
ON A SOCIAL CHANGE INITIATIVE

presented by

Boris H. J. M. Brummans

has been accepted towards fulfillment
of the requirements for

Masters degree in Communication

mma/A

Major profit/ssor

Date WM

0-? 639 MS U is an Affirmative Action/Equal Opportunity Institution

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11/00 c/CIRCIDanst-p.“

 

THE EFFECT OF AMBIGUITY ON A SOCIAL CHANGE INITIATIVE
By

Boris H. J. M. Brummans

A THESIS

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

MASTER OF ARTS

Department of Communication

19 April, 2000

ABSTRACT
THE EFFECT OF AMBIGUITY ON A SOCIAL CHANGE INITIATIVE
By

Boris H. J. M. Brummans

Based on research on diffusion of innovations, frame analysis, and strategic ambiguity,
the following hypothesis is tested: In a decentralized diffusion system, potential adopters
perceive high ambiguity in the meaning of a social change initiative as detrimental to the
initiative’s implementation. Ambiguity is assumed to occur in a decentralized diffusion
system, because it is likely that the change agency uses a set of ideas and assumptions to
define the meaning of the initiative that is different from the one the potential adopters
use. In total, 48 potential adopters from three communities were interviewed about their
experiences with the implementation of a particular initiative. The transcribed interviews
were searched for ambiguity-synonymous and ambiguity-antonymous words by means of
a Microsofi Word word-search. In total, 263 words were found and reliably content
analyzed by two coders in terms of high/low ambiguity (independent variable). In
addition, the coders coded whether the interviewee perceived this ambiguity as
detrimental/ beneficial to the implementation (dependent variable). Results showed a

significant, negative, association between the independent and dependent variable
(x2=1 17.048, 1 df, p=.000; <I>= -.667, p=.000; <I>2=.450). Thus, it is assumed that a

change agency and potential adopters of a change initiative need to establish a common

ground for defining the meaning of the initiative to achieve implementation success.

Copyright by
BORIS H. J. M. BRUMMANS
2000

PREFACE

I would like to thank the following people:

Professor James Dearing, for making sense of my ideas and writing, for giving me a great
opportunity to learn, and for anticipating and reinforcing the academic direction that I am
taking.

Professor Peter Manning, Professor William Donohue, and Professor James Dearing, for
being Splendid thesis committee members and for letting me stand on their shoulders.

Edward Roberts and Stefania Tortorici, for jolly good coding and friendship.

Stefania Tortorici, for showing me what is true and for connecting me to the things that
are worthwhile and beautiful.

IV

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

CHAPTER ONE: INTRODUCTION

Frame convergence or divergence in organizational
communication

Defining the meaning of a social change initiative
The difi’usion of an innovation

Defining the meaning of a social change initiative in a
decentralized diflusion system

CHAPTER TWO: THEORY
Using frames to define meaning
The process of defining meaning
The politics of defining the meaning of a social change
initiative
Ambiguity as a result of frame divergence
Ambiguity: Multiplicity and uncertainty in meaning

The effect of ambiguity on the implementation of a social
V change initiative in a decentralized diffusion system

CHAPTER THREE: THE COMPREHENSIVE COMMUNITY
HEALTH MODELS INITIATIVE

Diffusion of the Initiative in the counties of Calhoun, Muskegon,
and St. Clair

VIII

IX

11

11

11

16

17

17

18

20

23

Defining the meaning of the Initiative in different inter-
organizational subsystems
CHAPTER FOUR: METHODOLOGY
Quantitative content analysis
Use of quantitative content analysis
Type of data
Unit of analysis
Coding the transcripts
Coder training
CHAPTER FIVE: RESULTS
Coder reliability

Inter-coder reliability
Intra-coder reliability

Content analysis results

CHAPTER SIX: CONCLUSION
Review of results
Theoretical implications

Conceptualization of ambiguity
Re—consideration of the validity of the strategic ambiguity
argument
Reviewing the diffusion process in terms of the studied
initiative

Research limitations and implications for future research
Generalizing from a small number of words to a small
number of individuals to a community

Empirically investigating frame convergence and divergence
Generalizing across studies

VI

24

27
27
27
29
29
3O
39
42
42

42
43

43

51

51

52

52

53

58

60

6O

61
61

REFERENCES

APPENDIX A

VII

63

68

LIST OF TABLES

Table 1.
Thesaurus List of Ambiguity-Synonymous Words

Table 2.
Thesaurus List of Ambiguity—Antonymous Words

Table 3.
List of Underlined Words not Included in Table 1. & 2.

Table 4.
Ambiguity by Implementation Cross- Tabulation

Table 5.
Pearson 's Chi—Square T est for Ambiguity by Implementation

Table 6.
Pearson ’s Phi(—Square) Coefficient T est for Ambiguity by Implementation

Table 7.
Ambiguity by Implementation by Community Cross-Tabulation

Table 8.
Pearson 's Chi-Square T est for Ambiguity by Implementation per Community

Table 9.
Pearson 's Phi(-Square) Coefficient T est for Ambiguity by Implementation
per Community

VIII

32

34

69

45

45

45

48

50

LIST OF FIGURES

Figure 1. Rogers’ (1995) model of the innovation process in an organization.

Figure 2. Model representing the definition and redefinition of the meaning
of an object, event, idea, or action by communicating individuals in a social
system based on their convergent frames.

Figure 3. Model representing the definition and redefinition of the Initiative’s
meaning by communicating individuals in and beyond a particular

subsystem of the decentralized, inter-organizational system based on

their convergent frames.

Figure 4. Model of the studied initiative’s innovation process based on
Rogers’ (1995) model of the innovation process in an organization.

IX

15

25

58

CHAPTER ONE

INTRODUCTION

The present thesis investigates to what extent the ambiguity that results from not sharing
the same premises for determining the meaning of a social change initiative is perceived
as beneficial or detrimental to the implementation of the initiative. This first chapter

introduces the main theoretical ideas.

Frame convergence or divergence in organizational communication

Organizational communication scholars (e.g., Manning, 1992a/b; Pacanowsky &
O’Donnell-Trujillo, 1983; Putnam, 1982; Weick, 1979, 1995) have increasingly given

attention to the role of cognition and symbolism in organizing. Predominantly, symbolic

interactionistsI focus on how organizational members define their organizational context,
roles, and activities by communicating with each other.

Unlike functionalistic organizational communication scholars, symbolic
interactionists believe that an organization does not exist outside its members. An

organization is constructed through communication by individuals who are involved in

 

The symbolic interactionist perspective originated from the work of researchers such as
Bateson (1972), Berger and Luckmann (1966), Blumer (1969), and Mead (1934).

joint problem—solving to accomplish a common goalz. Rogers (1995) defines

communication as:

A process in which participants create and share information with one another in
order to reach a mutual understanding. This definition implies that communication
is a process of convergence (or divergence) as two or more individuals exchange
information in order to move toward each other (or apart) in the meanings they
give to certain events. (Rogers, 1995, p. 6)

The way in which communicating individuals jointly produce and re-produce
meaning depends on the “frames” they use during the definitional process. A frame is a
set of ideas and assumptions used by an individual to define what something means. By
influencing the frames of organizational members, organizational leaders influence how
these members define important matters such as organizational roles and goals (Bolman
& Deal, 1991; Hare & Blumberg, 1988; Johnson, 1997). Frames guide people toward
appropriate, “consensually validated” (Weick, 1979, p. 3), interpretations of what

something means:

Frames are both windows on the world and lenses that bring the world into focus.
Frames filter out some things while allowing others to pass through easily. Frames
help us to order experience and decide what action to take. (Bolman & Deal,
1991,p.11)

 

The term “member of an organization” is somewhat misplaced, because the symbolic
interactionist perspective presupposes the non-objective existence of an organization.
From this point of view, an organizational member perceives himself or herself as a
member of an organization through social construction. Durkheim (1950) notices the
importance of what he calls “social facts” such as organizations: Although social facts
like organizations are products of human interaction (and particularly of human
communication), individuals perceive these facts outside themselves and experience the
consequences of these facts as real and backed up by sanction.

Some researchers (e.g., F iske, 1997; Watanabe, 1993) mention that the frames of
those who are communicating with each other in a particular context converge or diverge.
Hare and Blumberg (1988) use a different term for frame convergence. They write about
a social interactionalfield, “a dynamic entity that includes all the individual perceptual
fields, plus the overt behavior by means of which people attempt to communicate to each
other the content of their perceptual fields” (p. 65). Snow, Rochford, Worden, and
Benford (1986) use the term frame alignment, referring to “the linkage or conjunction of
individual interpretive frameworks” (p. 467).

As Rogers’ (1995) definition of communication already suggested,
communicating individuals succeed or fail to understand each other depending on
whether their frames converge or diverge. Frame convergence refers to the extent to
which communicating individuals use similar sets of ideas and assumptions to define
what something means. Frame divergence, on the other hand, is defined as the extent to
which communicating individuals use dissimilar sets of ideas and assumptions to define
what something means.

Bolman and Deal (1991) argue that frame divergence often impedes effective
communication between people in an organization, because people are often unwilling or
unable to take a different point of view once they have ad0pted a particular frame. From a
diffusion of innovations perspective, it is argued in this study that people who are
involved in implementing a social change initiative are hindered by the ambiguity in

meaning that results from their diverging frames.

Defining the meaning of a social change initiative

The diffusion of an innovation

Rogers (1995) defines diffusion as “the process by which an innovation is
communicated through certain channels over time among the members of a social
system” (Rogers, 1995, p. 5). He explains that diffusion is the communication of an
innovation, “an idea, action, or object that is perceived as new by an individual or other

unit of adoption” (Rogers, 1995, p. 11). The meaning of an innovation is defined by

individuals who are communicating with each other in a social system3, “a set of
interrelated units that are engaged in joint problem-solving to accomplish a common
goal” (Rogers, 1995, p. 23). Within a social system, an innovation’s meaning is more or
less commonly accepted depending on the convergence or divergence between the frames
of the communicating individuals. Not only, do social system members define the
meaning of an innovation, but also the meaning of the social system itself, that is, their
roles, the problems they are attempting to solve, and the goals they are trying to achieve.
The diffusion of innovations theory pertains to the diffusion of “physical”
innovations such as new technologies or products, but also, in a broader sense, to social
change, “the process by which alteration occurs in the structure and function of a social
system” (Rogers, 1995, p. 6), involving a change in the decision making, attitudes, values
and/or knowledge of social system members. This change can be planned, as system

members anticipate or respond to events that are occurring outside or inside their own

 

“The members or units of a social system may be individuals, informal groups,
organizations, and/or subsystems” (Rogers, 1995, p. 23).

social system. However, social change can also happen spontaneously.
Sometimes, social change is organized in the form of a social change program or

social change initiative, referring to:

An organized effort conducted by one group (the change agent [i.e., the change
agency assisted by its change agents]), which intends to persuade others (the
target adopters) to accept, modify, or abandon certain ideas, attitudes, practices,
and behavior. (Kotler & Roberto, 1989, p. 6)

A change agency is an institution, usually initiating, sometimes designing, and
less commonly governing a social change program, seeking to increase the adoption of an
innovation by a targeted audience. The target audience is defined by the change agency
and is composed of potential innovation adopters. A potential adopter is an individual
who may make the decision to use an innovation because he or she perceives it as the best
course of action available (Rogers, 1995). A change agent is “an individual who
influences clients’ innovation-decisions in the direction deemed desirable by the change
agency” (Rogers, 1995, p. 335). He or she functions as the critical intermediary between
the change agency and the potential adopters with the aim to secure the adoption of an
innovation, but also to manage the adoption process in general.

The process of social change in an organizational social system is characterized by
two sub-processes (Rogers, 1995) (see Figure 1.). The first sub-process, initiation, “all of
the information gathering conceptualizing, and planning for the adoption of an
innovation, leading up to the decision to adopt” (Rogers, 1995, p. 392), consists of (1)

agenda-setting, “the way in which needs, problems, and issues bubble up through a

 

#1.

#2.

Agenda-setting:

General problem

that may create

a perceived need

for innovation is defined

Matching:
General problem is fit
with an innovation

 

 

 

#3.

#5.

Redefining/
Restructuring:
Re-invention and
modification to fit
the organization,
and alteration of
organizational
structure

Clarifying:

Clearer definition

of relationship between
organization and
innovation

Routinizing:
Innovation becomes
ongoing element in
organization’s
activities, and loses
its identity

INITIATION

DECISION TO ADOPT

IMPLEMENTATION

Figure 1. Rogers’ (1995) model of the innovation process in an organization.

system and are prioritized in a hierarchy for attention” (Rogers, 1995, p. 391), and (2)
matching, “the stage in the innovation process at which a problem from the organization’s
agenda is fit with an innovation, and this match is planned and designed” (Rogers, 1995,
p. 394). Implementation, the second sub-process, refers to “all the events, actions, and
decisions involved in putting an innovation into use” (Rogers, 1995, p. 392) and includes
(1) redefining, the re-invention of the innovation to accommodate the organization’s
needs and structure more closely, and the modification of the organization’s structure to
fit with the innovation, (2) clarifying, the putting “into more widespread use of the
innovation in an organization so that the meaning of the new idea gradually becomes
clearer to the organization’s members” (Rogers, 1995, 399), and (3) routinizing, the
incorporation of the innovation into the regular organizational activities while the

innovation gradually loses its separate identity.

Defining the meaning of a social change initiative in a decentralized diffusion system
According to the classical diffusion model, an innovation originates from a
centralized source and then diffuses to the potential adopters (Rogers, 1995). The
centralized diffusion model, proposing the vertical-to-horizontal communication of a new
idea, was challenged by Schon (1967). He claimed that the classical diffusion model
failed “to capture the complexity of relatively decentralized diffusion systems in which
innovations originate from numerous sources and then evolve as they diffuse via
horizontal networks” (Rogers, 1995, p. 364). Comparing a decentralized with a

centralized diffusion system in terms of communication, Rogers (1995) writes:

In general, centralized diffusion systems are based on a more linear, one-way
model of communication. Decentralized diffusion systems more closely follow a
convergence model of communication, in which participants create and share
information with one another to reach mutual understanding. (p. 365)

Rogers’ (1995) quotation enables the introduction of Leonard-Barton’s (1988)
concept of mutual adaptation, which she defines as “the re-invention of the technology
[i.e., the innovation] and the simultaneous adaptation of the organization” (p. 253).
Although her article predominantly focuses on the implementation of new technologies,
some of her remarks and findings also pertain to the implementation of a social change
initiative. Mutual adaptation is necessary, Leonard-Barton (1988) argues, because a new
idea almost never fits perfectly into the adopter environment. Misalignments between the
innovation and those adopting are common. These misalignments can impede the
implementation stage in the diffusion process and therefore “must be addressed if the
implementation is to succeed” (Leonard-Barton, 1988, p. 252). Leonard-Barton ( 1988)
defines misalignments as “poor fits” (Leonard-Barton, 1988, p. 252), but does not discuss
the misalignment between the frames (i.e., frame divergence) of the change agency and
those of the potential adopters.

Because frame convergence is important for the functioning of an organization
(Bolman & Deal, 1991; Hare & Blumberg, 1988; Johnson, 1997; Snow, Rochford,
Worden, & Benford, 1986), mutual adaptation does not only have to happen by altering
the innovation itself and the adopting environment, as Leonard-Barton (1988) argues, but
also by ensuring, through communication, that the frames of the change agency and the

potential adopters become similar. If this convergence is not initiated and maintained

throughout the diffusion process, the meaning of the innovation will be defined
dissimilarly and ambiguity might be the result.

Some researchers argue that frame convergence can contribute to organized
activity and understanding (Bolman & Deal, 1991; Hare & Blumberg, 1988; Johnson,
1997; Snow, Rochford, Worden, & Benford, 1986). From a conflict negotiation
perspective, Drake and Donohue (1996) state that “disputants operating under divergent
frames increase Ihe likelihood of distributive interaction (Bateson, 1972; Donohue, 1991;
Johnstone, 1989; Putnam et al., 1986), escalation (Pruitt & Rubin, 1986), and protracted
disagreement (Lax & Sebenius, 1986; Pillar, 1983)” (p. 301).

Others contend that frame divergence can lead to ambiguity which can be
productive for an organization, when used strategically, because it provides room for
differential, self-serving interpretation by each person involved in the organizational
communication process (Sussman, 1975; Eisenberg, 1984; Eveland, 1987). Eveland

(1987) writes:

Sharing information among people (and organizations) requires that all be
operating on somewhat the same general level of abstraction, and be using
something like the same variety of metaphors. It does not require perfect
information, or precise specificity, to be effective-sometimes ambiguity and
generality can be very effective, particularly if one does not know just what sort of
metaphors an information recipient is applying. (p. 307)

With respect to the ambiguity caused by frame divergence, Weick (1995), like
Goffman (1974), suggests that ambiguity can be understood as confusion created by
multiple meanings or as ignorance caused by insufficient information, which is similar to

uncertainty. Eisenberg (1984) argues that ambiguity can be used strategically “to foster

agreement on abstractions without limiting specific interpretations” (p. 231), and,
therefore, can be used to facilitate social change. Similarly, looking at superior-
subordinate dyads, Sussman (1975) notes that frame divergence does not necessarily lead
to communication breakdown but is “a ‘natural’ and often ‘healthy’ state of such dyads”
(p. 192).

Based on Sussman’s (1975), Eisenberg’s (1984), and Eveland’s (1987) notions, it
is plausible that frame convergence is not a necessary condition for the successful
diffusion of a social change initiative in a decentralized diffusion system. In a
decentralized diffusion system, the meaning of a change initiative is not pre-defined by
the change agency, but “negotiated” by the change agency and the potential adopters.

This theoretical assumption is incongruent with the idea that frame misalignments
may impede mutual adaptation during the implementation of a change initiative. In
addition, those involved in assisting the planning and implementation of the social
change initiative that is investigated in this thesis, indicated that the potential adopters of
the initiative perceived ambiguity in the initiative’s meaning as detrimental to the
implementation. The following research question guides the investigation of this

incongruency:

RQ: Do potential adopters perceive high ambiguity in the meaning of a social
change initiative as beneficial or detrimental to the initiative’s
implementation in a decentralized diffusion system?

10

CHAPTER TWO

THEORY

This chapter presents the theoretical basis for studying the following research question:

RQ: Do potential adopters perceive high ambiguity in the meaning of a social
change initiative as beneficial or detrimental to the initiative’s
implementation in a decentralized diffusion system?

Using frames to define meaning

The process of defining meaning
A sign is “a discemable indication of what is not itself directly perceptible”

(Merriam-Webster’s Collegiate Dictionary, 1998. P. 1091). Fiske (1997) explains:

A Sign is something physical, perceivable by our senses; it refers to something
other than itself; and it depends upon a recognition by its users that it is a

Sign. . .the word CAR (marks on paper or sounds in air) has a mental concept
attached to it. Mine will be broadly the same as yours, though there may be some
individual differences. This shared concept then relates to a class of objects in
reality. This is so straightforward as to seem obvious, but there can be problems.
My wife and I, for example, frequently argue over whether something is blue or
green. We share the same language, we are looking at the same piece of external
reality: the difference lies in the concepts of blueness or greenness that link our
words to that reality. (p. 41)

11

A sign is composed when a signifier, the sign’s perceived image (e.g., the sight of
a board next to the road with the following signs: “ATTENTION! IL Y A DES OURS
SUR LA ROUTE!”4), and a signified, the mental concept to which the Sign refers (e. g.,
someone’s mental concept of what the signs “ATTENTION! IL Y A DES OURS SUR
LA ROUTE!” mean) are mentally connected with each other (F iske, 1997). Si gnifiers and
signifieds are socially constructed by people who are communicating with each other in a
particular context (i.e., someone who knows the Roman alphabet and language rules,
interprets the signs “ATTENTION! IL Y A DES OURS SUR LA ROUTE!” as words
that are separated by exclamation marks and spaces, whereas someone who is not familiar
with this alphabet and these language rules might interpret the signs as expressions of a
modern artist; French people, in addition, connect the perceived signifier
“ATTENTION!” to the signified “BE CAREFUL!”, but a Russian might not make this
connection).

A Sign does not have meaning in itself unless it is connected to other signs (e.g.,
for understanding the board with the signs “ATTENTION! IL Y A DES OURS SUR LA
ROUTE!”, someone has to connect the perceived signs with each other by using the
conventions of the French language; only someone who knows these conventions will
define the meaning of the marks as: “BE CAREFUL! THERE ARE BEARS ON THE

ROAD!”).

 

A Sign may be a mark on a piece of paper such as a character (e.g., “A”), a number of
these marks combined as a word (e.g., “ATTENTION”), a sentence, or a complete story
(in the example mentioned above, a Sign is considered to be a word). Because a sign can
be defined in terms of many different levels of analysis, it seems difficult to pinpoint its
exact meaning.

12

Over time, communicating individuals define the meaning of an object, event,
idea, or action by developing social conventions to determine how signifieds and
signifiers that relate to the object, event, idea, or action need to be constructed, how

signifiers need to be connected with signifieds, and how signs need to be connected with

other Signs. One of the important assumptions5 that Fiske (1997) makes, is that the
mentioned social conventions are broadly shared by people who are communicating with
each other in a social system (or social group). Communicating individuals are able to
understand each other, because they agree, by convention, on what is meaningful and

what is not. Fiske (1997) describes the communication process as follows:

When I communicate with you, you understand, more or less accurately, what my
message means. For communication to take place I have to create a message out
of signs. This message stimulates you to create a meaning for yourself that relates
in some way to the meaning that I generated in my message in the first place. The
more we share the same codes [i.e., the same social conventions], the more we use
the same sign systems, the closer our two ‘meanings’ of the message will
approximate. (p. 42)

Instead of speaking about a social convention, the term “frame” is used in this
thesis to denote the set of assumptions that people use to construct signifiers and
signifieds, to connect signifiers with signifieds, and to connect signs to other signs. A
frame is “a set of ideas, conditions, or assumptions that determine how something will be
approached, perceived, or understood” (Merriam-Webster’s Collegiate Dictionary, 1998,
p. 463). Goffman (1974) defines frames similarly, as “principles of organization which

govern events-at least social ones-and our subjective involvement in them” (p. 10-11).

 

This assumption is also made by Watanabe (1993), and Hare and Blumberg (1988).

13

Based on these two definitions, a frame is defined in this thesis as a set of ideas and
assumptions used by an individual to define what something means.

The legitimacy of a particular frame depends on the agreement between
communicating individuals as well as on their shared background and experience. For
this reason, a frame might be shared or not shared. Frame convergence refers to the
extent to which communicating individuals use similar sets of ideas and assumptions to
define what something means. Frame divergence is defined as the extent to which
communicating individuals use dissimilar sets of ideas and assumptions to define what
something means.

Figure 2. presents a simplified representation of the meaning definition process in
a social system. It shows that individuals define and redefine the meaning of a particular
object, idea, event, or action in a similar way depending on the convergence of their
frames. The commonality of meaning is established and maintained through
communication between the individuals in the social system.

Hare and Blumberg (1988) argue that the meaning definition process is not

delimited by a starting or ending point:

Just as every action begins with a definition of the situation, an idea that governs
the action (Stebbins, 1986), so every action ends with a new definition of the
Situation (Perinbanayagam, 1974). (p. 91)

In line with Hare and Bl‘umberg’s (1988) remark, Figure 2. indicates that the definition of

meaning depends on the convergence of frames, but that this convergence is reconsidered

constantly. In this way, the object’s, idea’s, event’s, or action’s meaning is redefined on a

14

continuing basis. When meaning is redefined, it might change or remain more or less

 

 

 

unaltered.
Individuals in a social system
Communication within social system
Object’s, Convergent frames
event’s, idea’s, ‘ of the individuals in
or action’s (Re-)definition the social system with
meaning regard to an object,
event, idea, or action

 

 

 

Figure 2. Model representing the definition and redefinition of the meaning of an object, event,
idea, or action by communicating individuals in a social system based on their convergent
frames.

It is assumed that the meaning of a social change initiative6 changes quickly
during the initiation sub-process (see Figure 1.), because its meaning has not been
established yet. Particularly during the clarifying stage, the frame of the change agency
should start to converge with those of the potential adopters, so that a common meaning

of the change initiative can be agreed upon. Rogers (1995) writes, in this regard:

 

A social change initiative can be conceptualized as one sign, but also as a set of
interconnected Signs. The decision to conceptualize a Sign in a particular way (on a
particular level of analysis) depends on the point of view of the researcher (see footnote
5). In this study, a social change initiative is assumed to be a set of interconnected signs.
The meaning of a social change initiative is defined by interconnecting signs in a way
that is meaningful to those who are communicating with each other in a particular social
system.

15

The clarifying stage in the innovation process in an organization consists of social
construction. When a new idea is first implemented in an organization, it has little
meaning to the organization’s members. The innovation is surrounded by
uncertainty...Through a process of the people in an organization talking about the
innovation, they gradually gain a common understanding of it. Thus their meaning
of the innovation is constructed over time through a social process of human
interaction. (p. 399)

By the time the routinizing stage has been reached, the basis for defining the meaning of
the initiative should be commonly accepted, although this is not always the case in

reality.

The politics of defining the meaning of a social change initiative

Although it is useful to describe the meaning definition process by using a model,
a discussion of this process in terms of politics is important for providing a more accurate
description.

A social system is often a conglomerate of social subsystems that are defined as
one entity by the individuals in these subsystems. When a particular change initiative is
implemented in such a coalition of subsystems, the way in which the initiative is
commonly defined may become problematic (Eveland, 1987), because the change
agency’s frame and those of the potential adopters might diverge. If this is the case, the
determination of the initiative’s meaning can be difficult.

Politics, “the interaction of interests among parties in a relationship” (Eveland,
1987, p. 311), plays an important role in the definition of an initiative’s meaning. For
some, it is important to influence how an initiative is defined. Some scholars (Drake &

Donohue, 1996; Gamson & Wolfsfeld, 1993; Snow, Rochford, Worden, & Benford,

16

1986) speak, in this regard, in more or less the same terms, of “negotiation over
meaning” (Gamson & Wolfsfeld, 1993, p. 117).

Those who create and control the ambiguity involved in the definitional process
can manipulate those who have less control during these negotiations over meaning
(Manning, 1992b). Furthermore, those who control the definition process can highly
influence the way in which an organization’s agenda is set. This agenda determines the
organizational priorities.

Besides the influence of politics on the way in which the meaning of a change
initiative is defined, Eveland (1987) remarks that the definition of the initiative’s
implementation can be subject to political power too. Those who control the definitional
process by couching the innovation in terms of the political and strategic imperatives of

the organization (Johnson, 1997) determine what, how, and when is implemented.

Ambiguity as a result of frame divergence

Ambiguity: Multiplicity and uncertainty in meaning

Often, ambiguity results from frame divergence. Ambiguity can be defined as the
quality of being “[1] doubtful or uncertain esp. from obscurity or indistinctness. . .[2]
understood in two or more possible senses or ways” (Merriam-Webster’s Collegiate
Dictionary, 1998, p. 36). Goffman (1974) defines ambiguity in a similar way. He
distinguishes two types of ambiguity, “one, where there is a question as to what could

possibly be going on; the other as to which one of two or more clearly possible things is

17

going on” (p. 302-303). Both these conceptualizations suggest that the meaning of an
object, idea, event, or action can be uncertain or that there can be multiple meanings.
Uncertainty refers to the quality of being “indefinite... problematical
...untrustworthy [unreliable]. . .dubious. . .doubtful. . .not clearly identified or defined. . ..
variable, fitful” (Merriam-Webster’s Collegiate Dictionary, 1998, p. 1285). Based on this
definition, uncertainty in meaning depends on the degree to which the meaning of an
object, idea, event, or action is clearly defined, definite, or determinate. An object, event,
idea, or action has high uncertainty in meaning when it is unclearly (or not at all) defined
and low uncertainty in meaning when it is clearly defined. Multiplicity in meaning, on the
other hand, refers to the degree to which the meaning of an object, idea, event, or action is
defined in more than one way. An object, event, idea, or action has high multiplicity in
meaning when it is defined in more than one way and low multiplicity in meaning when

it is defined in only one way.

The effect of ambiguity on the implementation of a social change initiative in a
decentralized diffusion system

To achieve successful mutual adaptation during the clarifying stage, the meaning
of a social change needs to become defined in a similar way by the change agency and the
potential adopters. Such similarity can only be achieved by communicating individuals if
they come to an agreement about the rules for defining the initiative’s meaning, that is, if
they communicate to achieve frame convergence.

In a decentralized diffusion system, mutual adaptation of a change initiative’s
meaning cannot always be achieved, because the change agency purposively takes a

hands-off approach in providing direction to the definitional process. A common meaning

18

of the initiative needs to be negotiated in a decentralized diffusion system and, as a result
of the frame divergence that often occurs during this negotiation process, potential
adopters encounter ambiguity in the meaning of the initiative.

In the strategic ambiguity literature, the argument is made that “strategic
ambiguity is essential to organizing, because it allows for multiple interpretations to exist
among people who contend that they are attending to the same message” (Eisenberg,
1984, p. 231). This argument is incongruent with Leonard-Barton’s (1988) argument that
misalignments can be detrimental to the process of mutual adaptation. Furthermore, the
conclusions reached by some of those involved in assisting the planning and
implementation of the change initiative that is investigated in this thesis, indicated that
the potential adopters of the initiative perceived ambiguity in the initiative’s meaning as
detrimental to the implementation. To guide the investigation of the strategic ambiguity
argument’s validity against the perceptions of those involved in implementing the

initiative of study, the following hypothesis is tested:

H: In a decentralized diffusion system, potential adopters perceive high
ambiguity in the meaning of a social change initiative as detrimental to the
initiative’s implementation.

19

CHAPTER THREE

THE COMPREHENSIVE COMMUNITY HEALTH MODELS INITIATIVE

In 1992, the W. K. Kellogg Foundation (hereinafter referred to as “the Foundation” or as
“Kellogg”) introduced a new community health care reform program, the
“Comprehensive Community Health Models Initiative” (referred to as “the Initiative”), in
order to assist three communities to improve the health of residents through social and
institutional change.

Major assumptions proposed for testing the Initiative were that (1) communities
could identify community health care needs and effectively manage health care resources
to address these needs, and (2), in order to improve health care provision, local
consumers, purchasers, and providers of health care needed to be equally engaged in
decision making with respect to the allocation and use of health care resources. These
assumptions implied the Foundation’s expectation that local communities did not lack
resources for improvement, but rather coordination and mechanisms for more inclusive
decision making.

In order to operationalize the Initiative, the Foundation established formal
partnerships with the community foundations of Calhoun, Muskegon, and St. Clair
counties. The selection of these particular communities was based on the assessment of
six criteria: (1) One to three community hospitals needed to exist in the community, (2) a
union needed to be present, (3) there needed to be indications of problems in the health

care system, (4) at least one corporate employer needed to be present, (5) the community

20

needed to have a population of 75,000 to 200,000, and (6) there needed to be a significant
population of minorities or under-served.

As a primary communication channel between the Foundation and the partnering
communities, the Foundation established an Operations Office. In addition to its primary
function, the Operations Office facilitated project planning and community development,
provided technical assistance, coordinated consultants, supported the Initiative’s
evaluation process, and facilitated federal and state governmental relations to assist the
three communities in implementing policy changes.

To establish inclusive, participatory community decision-making, a community
governing board, a community-level decision making unit composed of consumers,
providers, and purchasers of community health care, was established in each community.
Its purpose was to determine how health care services needed to be organized and what
resources needed to be assigned to health care.

As intermediaries between the Foundation and the Operations Office on the one
hand, and the communities on the other, local program offices, consisting of organized
groups of professional staff members, were instituted in each community. From the start
of envisioning the intended community improvement, the Initiative had six objectives,
regarding (1) community decision making (governance), (2) community health
assessment, (3) community-wide coverage, (4) an integrated administrative structure, (5)
a community-based information system, and (6) an integrated health delivery system.

The Initiative was composed of four phases. In the introduction phase (three to six
months), local community foundations established a dialogue to inform community

members about the Foundation’s invitation to become a partner.

21

During phase two, the visioning and planning phase (12 to 18 months), each of
the partnering community foundations instituted a planning committee, whose members
were drawn from the community at-large, and consisted of consumers, providers, and
payers of health care. In every community, the planning committee envisioned the
community’s health care fisture. In turn, this vision served as a basis for developing a 12-
month health investment plan that described how the community proposed to address
health care reform. In essence, each health investment plan was a community’s
interpretation of the six objectives.

The first objective stated the need for community decision making or governance
to assure that the legitimate interests of the payers of care, the providers of care, and the
needs of community residents were responsibly met. This meant that each of these groups
needed to participate equally in the community decision making process established to
oversee the health care system of the community. The second objective pertained to
achieving community-wide coverage through the expansion of affordable insurance
coverage and better access to health care services. Developing an integrated health
delivery system that elevated the roles of disease prevention and health promotion, and
integrated medical, health, and human services, was the third objective. Fourth was the
objective to establish an integrated administrative structure. Such a structure gave a
community the possibility to create a single set of rules and regulations to reduce system
fragmentation and waste, and to create a centralized database for monitoring system
performance. The fifth objective stated the need for establishing a community—based
information system in each community. The final objective pertained to conducting

periodic community health assessments to assess community residents’ health status and

22

each community’s health care delivery system, and to provide benchmarks for monitoring
change while the Initiative was implemented.

In the endorsement phase (three to four months), the communities sought broad
public support for the health investment plan by having various interest groups and the
community in general review the proposed plan.

During the implementation phase (three to five years), the health investment plan
was refined, updated, and put into use. The main goal of this period was to assure
complete integration and sustainability of the health investment plan’s major elements

once the Foundation’s grant had ended.

Diffusion of the Initiative in the counties of Calhoun, Muskegon, and St. Clair

The Initiative was an innovation, because it proposed a significant redesign of the
communities’ health care systems at the local level. Because such plans threatened to
change long-established patterns of resource distribution and decision making, the
Initiative could be considered a somewhat radical innovation, embedding high degrees of
perceived uncertainty and risk.

With regard to the diffusion of the Initiative in the counties of Calhoun,
Muskegon, and St. Clair, the Foundation, assisted by the Operations Office, was the
change agency that initiated social change. Indirectly, Initiative participants, and directly,
governing board members, represented the potential adopters of the Initiative in each

community.

23

Each community’s professional staff, organized in local program offices,
functioned as change agents. They operated under the governance of the change agency to
(1) provide community accountability, (2) establish communication channels in the
community, and (3) maintain linkages to the community through work groups.

The partnering communities were given broad discretion in operationalizing the
Initiative’s six key objectives. The Foundation’s approach, therefore, may be

characterized as a decentralized, rather than centralized, diffusion system.

Defining the meaning of the Initiative in different inter-organizational subsystems

Johnson (1997) states that the meaning of an innovation often is constructed on
different levels of an organizational system and that these levels “entail considerably
different ways of viewing, interpreting, and discussing the innovation” (p. 17). In this
study, the diffusion of the Initiative occurred in a decentralized, inter-organizational
system. The system consisted of two levels, the level of the change agency and the level
of the three communities. The level of the change agency was considered to be a
subsystem that included the Foundation and the Operations Office. Each community was
considered to be a separate subsystem consisting of the potential adopters of the
Initiative. Each community was dependent on the change agency when it came to
financial resources, yet independent with regard to envisioning, organizing, and

implementing the Initiative.

24

To simplify how the meaning of the Initiative was defined in the inter-

organizational system, the model that was proposed in Figure 2. is extended to Figure 3.

 

’ Subsystem of Foundation &
Operations Office members

Intra-subsystem communication

 

 

v

Initiative’s Convergent frames of
meaning { Foundation & technical
assistance office
(Re-)definition members with regard
to the Initiative

 

 

 

Inter-subsystem communication

 

—> Potential adopters in each community subsystem
(Calhoun, Muskegon, & St. Clair)

Inna-subsystem communication

 

 

i

 

Initiative’s Convergent frames of
meaning < the individuals in each
community subsystem
(Re-)definition with regard to the
Initiative

 

 

 

Figure 3. Model representing the definition and redefinition of the Initiative’s meaning by
communicating individuals in and beyond a particular subsystem of the decentralized, inter-
organizational system based on their convergent frames.

25

Figure 3. indicates that the individuals in each subsystem defined and redefined the
meaning of the Initiative more or less similarly, depending on the convergence of their
frames. The commonality of the Initiative’s meaning was established and maintained
within each subsystem through intra-subsystem communication. A common meaning of
the Initiative could only be established between the change agency and the potential
adopters in a community, if both parties engaged in inter-subsystem communication. That
is, if both parties communicated about the frames they used to define the Initiative’s
meaning within their subsystem, and if both parties intended to achieve frame

convergence through this communication.

26

CHAPTER FOUR

METHODOLOGY

This chapter provides an operational framework for testing the following hypothesis:

H: In a decentralized diffusion system, potential adopters perceive high
ambiguity in the meaning of a social change initiative as detrimental to the
initiative’s implementation.

In this hypothesis, the degree of ambiguity and the effect of this ambiguity on

implementation are assumed to be nominal variables.

Quantitative content analysis

Use of quantitative content analysis

Content analysis refers to “any technique for making inferences by objectively
and systematically identifying Specified characteristics of messages” (Holsti, 1969, p. 14).
Message content can be expressed verbally, textually, or in the form of an image (Riffe,
Lacy, & Fico, 1998).

Communication researchers use quantitative content analysis in particular
situations or when a particular question needs to be addressed. Holsti (1969) writes that

content analysis is especially useful when it is difficult to access data and when a

27

researcher is limited to archival evidence (e. g., when a researcher wants to study Cuban
people’s attitude toward Fidel Castro’s regime during the Cuban missile crisis). In
addition, content analysis can be used to investigate the particular use of words in
discourse (e. g., when someone intends to study the meaning of the word “like” in
different contexts of discourse) or when the volume of data surpasses the investigator’s
ability to conduct research (Holsti, 1969).

Quantitative content analysis provides a researcher with the opportunity to
unobtrusively investigate the use of specific words, the frequency that particular variables
occur, or the re-occurrence of certain themes in people’s accounts of a particular object,
event, idea, or action. An interviewer could ask someone who is involved in the
implementation of an initiative, for example, if he or she thinks a high degree of
ambiguity is beneficial or detrimental to the implementation. However, when the
interviewee responds to such a question, only a one-time thought might be expressed,
rather than an underlying theme in his or her opinion. Moreover, asking such specific
question would coerce the interviewee to think about an issue that he or she might not
have thought about spontaneously, because he or she did not perceive it as an issue. By
conducting quantitative content analysis, the underlining theme in the thoughts of
interviewees can be observed and recorded for analysis, without prompting for Specific

issues.

28

Type of data

The present thesis is one component of a larger project7. Five members of the
research team who were involved in this project conducted 48 interviews with 48
potential adopters from Calhoun, Muskegon, and St. Clair. The interviewees had been
particularly active in the implementation of the initiative. The goal of the team was to

gain understanding about four constructs: (1) Decision making, (2) leadership, (3)

changes over time in the initiative’s visions, local organizations, or institutions and
partnership members, and (4) sustainability of the partnership vision. A1148 interviews
(i.e., 18 from Calhoun, 12 from Muskegon, and 18 from St. Clair) were transcribed for

content analysis.

Unit of analysis

Two coders were trained to code recording units, specific segments of content that
are characterized by placing in given categories (Holsti, 1969). In this study, a recording
unit was any word that indicated high or low ambiguity. However, not all words
indicating ambiguity functioned as recording units, only those that referred, in any way,
to the implementation of the initiative and to ambiguity in meaning. Adjectives like

9‘) 6‘

“equivocal”, “ambivalent”, or “two-edged”, nouns like “equivocality , ambivalence”, or

 

James Dearing, Principal Investigator, Implementation Phase Cluster Evaluation of the
Comprehensive Community Health Models Initiative, sponsored by the W. K. Kellogg
Foundation, Battle Creek, Michigan, C. Patrick Babcock, Program Officer.

Referring to the early W. K. Kellogg Foundation idea that improving local health care
systems should proceed from a community partnership that brings together consumers,
payers, and providers to make implementation decisions.

29

“uncertainty”, and verbs like “obscure” or “doubt”, indicated high ambiguity in meaning,
whereas adjectives like “clear”, “lucid”, or “definite”, nouns like “clarity” or “lucidity”,
and verbs like “clarify” or “understand” suggested low ambiguity in meaning.

The context unit is the largest body of content that can be searched to characterize
or interpret the classification of a recording unit (Holsti, 1969). In this study, the context
unit was the answer in which an interviewee mentioned the recording unit (an answer was
indicated by “A:” in the transcripts, whereas a question was indicated by “Q:”). The
coders used the context unit to classify the recording unit in terms of a particular level of
ambiguity in the meaning of the initiative (independent variable). After they had coded
the level of ambiguity, the coders classified whether the potential adopter perceived the
recording unit as beneficial or detrimental to the initiative’s implementation (dependent

variable).

Coding the transcripts

All words that indicated ambiguity were located in the transcripts by using the
Microsoft Word word-search (Ctrl+F) function. Which words were searched for, was
determined by using Merriam-Webster’s Collegiate Thesaurus (1993). A list of words
that were synonymous with “ambiguous”, “arnbiguously”, or “ambiguity” (high
ambiguity), or were similar in meaning, was comprised (see Table 1.). Not only the
synonyms, indicated with “syn” in Merriam-Webster’s Collegiate Thesaurus (1993), were
included in the list. Also related words, indicated with “rel”, and idiomatic equivalents,
indicated with “idiom”, were included, because it was likely that the interviewees used

words that were not exactly synonymous with the word “ambiguous” or with its

30

synonyms. Limited use words, indicated by

66””

(double bars) were not included in the list
because Merriam-Webster’s Collegiate Dictionary (1993) mentions that “the employment
of such a term may involve a problem of diction too complex for presentation in a
thesaurus, or a restriction in usage” (p. 26a). A list of antonyms was also comprised by
means of the Merriam-Webster’s Thesaurus (1993), in order to locate words that
indicated low ambiguity in meaning (see Table 2.).

All words that appeared in each transcript by using this search method, were
underlined with a pencil in a printed copy of the interview. However, some words were
underlined in the transcripts that did not appear in Table 1. or 2., because they represented
different forms of particular words that were included in Table 1. or 2., and were relevant
indicators of ambiguity. These words were included in Table 3. (see Appendix A). The
adjective “complicated” or the noun “complication” were underlined, for example,
although only the verb “complicate” occurred in Table 1. Furthermore, a few antonyms
were included in Table 3. that were not taken from Merriam-Webster’s Collegiate
Thesaurus (1993), but appeared during the search and possibly were relevant recording
units. Together, Tables 1., 2., and 3. present an exhaustive list of the words that were
underlined in the texts.

After the word-search, a researcher who had been working with the research team
for a year and who was well informed about the studied initiative determined which
underlined words could be accepted as valid recording units. Each time a valid recording
unit was located, the word was highlighted in the printed interview by the researcher. In
total, 263 recording units were found (i.e., 117 for Calhoun, 67 for Muskegon, and 79 for

St. Clair).

31

Table 1.

Thesaurus List of Ambiguity-Synonymous Words

 

 

 

Headword Synonym

AMBIGUOUS

(adjective) OBSCURE 3, amphibological, equivocal, opaque, tenebrous, uncertain,
unclear, unexplicit, unintelligible, vague, dubious, fishy, problematic,
suspect, unsettled, unsure, questionable, DOUBTFUL 1

OBSCURE 3 double-edged, double-faced, dusky, murky, nubilous, sibylline, difficult,
incomprehensible, inexplicable, puzzling, unfathomable, illegible,
abstruse, Delphian, enigmatic, esoteric, inscrutable, mysterious, mystic,
mystical, inconclusive, indecisive, indefinite

DOUBTFUL 1 borderline, clouded, doubtable, dubitable, impugnable, open, precarious,
queasy, shady, shaky, undecided, uneasy, unstable, question-begging,
touch-and-go, chancy, insecure, speculative, hazy, unlikely, contingent,
iffy, at issue, in dispute, in doubt, in question, MOOT, IMPROBABLE l

MOOT arguable, debatable, disputable, mootable, controversial

IMPROBABLE 1

AMBIGUITY
(noun)

 

UNCERTAINTY

OBSCURE
(verb)

no synonym(s) different from the above

amphibology, double entendre, double meaning, equivocality,
equivocation, equivoque, tergiversation, dodge, evasion, hedge, quibble,
shift, subterfuge, cavil, haggling, hairsplitting, quibbling, obscurity,
uncertainty, vagueness

UNCERTAINTY, concern, dubiety, dubiosity, incertitude, mistrust,
skepticism, suspicion, uncertitude, wonder, dubiousness,
questionableness, disbelief, incredulity, unbelief

doubtfulness, dubitancy, query, anxiety, bother, disquiet, trouble, worry,
agitation, distress, perturbation, uneasiness, disfaith, hesitation, reserve,
salt

adumbrate, becloud, bedim, befog, cloud, darken, dim, dislimn, eclipse,
fog, gloom, haze, mist, murk, obfuscate, obnumbilate, overcast,
overcloud, overshadow, shadow, blear, blur, fuzz, blind, conceal, dim
out, hide, screen, shade, shroud, bemask, camouflage, cloak, cover,
disguise, mask, veil, belie, falsify, misrepresent

32

DOUBT

 

(verb) QUESTION 2, challenge, dispute, mistrust, DISTRUST, misdoubt,
suspect

QUESTION 2 suspicion, hesitate (over), wonder (about)

DISTRUST disbelieve, discredit, unbelieve

CONFUSE 3, 4

(verb) PUZZLE, bewilder, muddy, complicate, confound, involve, mix up, lose
in a fog

PUZZLE baffle, foil, frustrate, befuddle

 

Note. The table lists words that are synonymous, related in meaning, or idiomatically equivalent,
to the adjective “ambiguous”, to the noun “ambiguity”, or to a verb related to “to be ambiguous”,
taken from Merriam-Webster’s Collegiate Dictionary (1993). Capitalized words indicate main-
entry headwords that directed the decision of which words to include in the list. For example,
starting with capitalized main-entry headword “AMBIGUOUS”, all the words that were
synonymous, related in meaning, or idiomatically equivalent, to the word “ambiguous” were
included in the list as well as those words that were synonymous, related in meaning, or
idiomatically equivalent, to its other capitalized main-entry headwords; that is, “OBSCURE”
(sense number 3) and “DOUBTFUL” (sense number 1). “DOUBTFUL”, again, had two main—
entry headwords, “MOOT” and “IMPROBABLE” (sense number 1). Underlined capitalized
words indicate the first main-entry headword that signifies the beginning of a sequence of
synonyms. If the same word occurred more than once, it was omitted from the list. The list of
noun synonyms is not limited to the main-entry headword “AMBIGUITY”. Looking at adjective
synonyms, Merriam-Webster’s Collegiate Thesaurus (1993) also mentioned the noun “DOUBT”
(related to “DOUBTFUL 1”) as acceptable main-entry headword. “To be ambiguous” does not
exist as a verb in Merriam-Webster’s Collegiate Thesaurus (1993). Therefore, also those verbs
that were associated with adjectives mentioned in the list such as “OBSCURE” (related to the
adjective “OBSCURE”), “DOUBT” (related to the adjective “DOUBTFUL 1”), and “CONFUSE
_3_,_4” were included. A list of words that were synonymous, related in meaning, or idiomatically
equivalent, to adverbs such as “ambiguously” were not included in the list. However, they were
included in the Microsoft Word word-search. While searching for the adjective “ambiguous”,
Microsoft Word automatically detected the adverb “ambiguously”. Sometimes, it was necessary
to type only a few basic characters of a particular word in order to find all the forms of that word.
If typing “ambig”, for example, Microsoft Word searched “ambiguity”, “ambiguous”, and
“ambiguously”.

33

Table 2.

Thesaurus List of Ambiguity-Antonymous Words

 

 

Headword Synonym (antonymous to “ambiguous”, “ambiguity”)

EXPLICIT

(adjective) categorical, clean-cut, clear-cut, definite, definitive, express, specific,
unambiguous, certain, clear, distinct, perspicuous, plain, sure,
unequivocal, accurate, correct, exact, precise

LUCID 3

(adjective) UNDERSTANDABLE, apprehensible, comprehendible, comprehensible,
fathomable, graspable, intelligible, knowable, luminous

UNDER-

STANDABLE unblurred, simple, straightforward, exoteric, lay, popular

INDUBITABLE l

(adjective)

POSITIVE 3

EXPLICITNESS
(noun)

LUCIDITY
(noun)

CLARITY

UNDER-
STANDING 3
(noun)

MEANING 1

POSITIVE 3, incontestable, indisputable, irrefutable, uncontestable,
uncontrovertible, undeniable, unquestionable

inarguable, incontrovertible, indubitable, irrebuttable, indisputable,
undoubtable, unequivocal, unquestionable, assured, decisive

no synonym(s)

CLARITY, cleamess, limpidity, perspicuity, plainness,
comprehensibility, intelligibility, understandability, distinctness

articulateness, articulation, care, exactitude, fussiness, meticulousness,
nicety, precision, accuracy, correctitude, propriety

MEANING 1, acceptation, import, intendment, message, purport, sense,
significance, Significancy, signification

intent, sum and substance, drift, effect, essence, tenor, force, point, value,

hint, implication, intimation, suggestion, connotation, definition,
denotation

34

ILLUMINATE 3
(verb) CLARIFY 2, clear, clear up, elucidate, explain, illustrate, construe,
define, dramatize, expound, express, gloss, interpret, shed light (up)on

CLARIFY 2 settle, straighten out, delineate, formulate, analyze, break down, simplify,
make plain

UNDER-

STAND l

(verb) APPREHEND 1, accept, catch, comprehend, get the drift

APPREHEND l compass, conceive, cotton (on)to, realize, recognize, sense, absorb,

digest, seize, catch on, wise (up), penetrate, catch the drift of, get the
idea, get through one’s head, make head or tail of

 

Note. The table lists words that are synonymous, related in meaning, or idiomatically equivalent,
to the opposite of the adjectives “ambiguous” “obscure” “doubtful”, the noun “ambiguity”, or to
a verb related to “to be ambiguous”, taken from Merriam-Webster’s Collegiate Dictionary
(1993). Capitalized words indicate main-entry headwords that directed the decision of which
words to include in the list. Underlined capitalized words indicate the first main-entry headword
that signifies the beginning of a sequence of synonyms. If the same word occurred more than
once, it was omitted from the list. The list of noun synonyms is not limited to the opposite of the
main-entry headword “AMBIGUITY”. Looking at the adjective antonyms, Merriam-Webster’s
Collegiate Thesaurus (1993) also mentioned the nouns “LUCIDITY” (related to the adjective
“LUCID 3”) and “UNDERSTANDING” (related to the adjective “UNDERSTANDABLE”) as
acceptable main-emu)I headwords. Because “to be ambiguous” does not exist as a verb in
Merriam-Webster’s Collegiate Thesaurus (1993), also those verbs that were associated to
antonymous adjectives mentioned in the list such as “ILLUMINATE 3” (related to the adjective
“LUCID 3”) and “UNDERSTAND” (related to the adjective “UNDERSTANDABLE”) were
included.

Hereafter, the coders determined the level of ambiguity in meaning for each
recording unit, while reading the entire interview. They made this decision based on the
conceptual definition of the two levels of ambiguity and based on the context of the
answer in which the word appeared (i.e., the context unit). Sentences like, “We all agreed
on what the Initiative meant in Calhoun” or “It was clear from the start what had to

happen. I’m sure most of us knew” indicated low ambiguity in meaning. Examples of

35

sentences that suggested high ambiguity in meaning were: “The confiising thing was that
we all had a different understanding of the Initiative” or “What was very complicated was
that Kellogg did not give us many guidelines for how to organize this whole Initiative.
Too many things were left open.” The coders wrote “H” in the margin next to the
recording unit in question, if the word pertained to “High ambiguity in meaning”, and
“L” was written for “Low ambiguity in meaning”. The coders were instructed to choose
one category only.

Sometimes, the context unit did not provide the coders with sufficient information
to make an accurate coding decision. For this reason, they were instructed to read the
entire interview and base their decisions on what made sense in the context of the
complete interview.

After determining the level of ambiguity, the coders decided if the interviewee
thought the particular level of ambiguity in meaning had been (or still was) beneficial or
detrimental to the implementation of the initiative in his or her community, based on the
context unit. If the context unit contained words such as “beneficial”, “positive”,
“helpful”, “constructive”, “good”, “favorable”, “effective”, or “practical”, the recording
unit was coded as “beneficial to the implementation”. A sentence like, “It was helpful for
the whole process that the Foundation stayed away from telling us what to do”, suggested
that high ambiguity was beneficial. Adjectives, like “negative”, “detrimental”, “bad”,
“destructive”, “ineffective”, or “unfavorable”, lead the coders to classify the recording
unit as “detrimental to the implementation of the initiative”. For example, the verb
“know” in the sentence, “It was vague. I still don’t know what Kellogg wants us to do”,

was coded as high ambiguity that was detrimental to implementation. “B” was written in

36

the margin, if the word indicated that the level of ambiguity in meaning was “Beneficial”,
“D” was written, if the word indicated that the level was “Detrimental”. Also here, the
coders were instructed to choose only one category.

Oftentimes, several recording units occurred in one interview answer. Only if a
recording unit indicated a different instance of high or low ambiguity, or a difference in
whether the interviewee thought ambiguity had been (or still was) beneficial or
detrimental, the unit was coded differently. If several recording units in one interview
answer indicated no difference, these recording units were coded as one recording unit.
The sentence, “Well, I don’t think we understood what was going on. This lack of
understanding was a major source of frustration during the implementation”, provides an
illustration. In this case, “understood”, “understanding”, and “fi'ustration” were three
recording units, but they received only one code (i.e., “HD”). A bracket was used to
indicate to which collection of recording units one code pertained.

Occasionally, the interviewer wanted to assure that he or She had understood the
interviewee’s answer. If an answer contained recording unit(s) that was/were identical
(i.e., exactly the same word) to the previous answer, both answers were coded as one
recording unit. If there was a separation of the answers in the text by an answer that did
not contain the same recording unit(s), they were coded as multiple recording units. For
example, “Q: So there was no vision?” “A: Yes, I don’t think we knew.” “Q: You don’t
think you knew?” “A: Yes, I don’t think we knew how to implement the Initiative and
this lead to great confusion.” In this case, “knew” occurred in two different answers that

immediately followed each other, but was coded as one recording unit (i.e., as “HD”).

37

Also here, a bracket was used to indicate to which collection of answers one code
pertained.

Sometimes, a recording unit was highlighted in the question of one of the
interviewers. If this occurred, the answer to the question or the answer that preceded the
question was used to code the recording unit. If a recording unit occurred in the question
and in the answer, only the recording unit in the answer was coded. For example, “Q: Do
you think that having no vision was confusing during the implementation?” “A: Yes, I
do.” In this case, “confusing” was the recording unit, but the answer of the interviewee
lead the coders in their coding.

To enable the measurement of inter-coder reliability, five randomly selected
interviews were duplicated and given to both coders. The remaining interviews were
randomly and equally distributed among the coders. This meant that each coder had an
equal amount of transcripts from each community. Before the coding started, two of the
five duplicated transcripts, coded by coder one, were randomly selected for data analysis.
In the same way, three transcripts, coded by coder two, were randomly selected for data
analysis. Three interviews were repeated after the coders had coded all their transcripts to
measure the degree of intra-coder reliability. The three repeated interviews were not used
for data analysis.

The coders were instructed not to code longer than three hours per day. In
addition, they were asked to return a transcript as soon as they finished coding it, to
assure that they treated every transcript independently. Finally, the coders were instructed

not to communicate with each other about anything related to their task.

38

Only after the coding was completed, numerical codes were given to the
characters written next to the recording units in the margin. The character “L” was coded
as “1”, “H” received code “2”; “D” received code “1” and “B” received a “2”. The
numerical values for the independent and dependent variable were then entered into two
separate columns in the “Statistical Package for the Social Sciences” software (Student
Version 7.5, 1996) (referred to as “SPSS”).

After entering the data, the content analysis results were cross-tabulated by means
of SPSS to display the association between the independent variable and the dependent

variable in the form of cell counts and cell percentages (see chapter five). By means of
SPSS, the Pearson ’s Chi-Square T est9 (X2) was computed to assess whether or not the

nominal dependent variable depended significantly on the nominal independent variable.

Pearson ’s Phi Coefficient (CID) was computed to calculate the correlation between the
independent and dependent nominal variables. Pearson ’s Phi-Square Coeflicientlo ((1)2)

was also calculated to analyze what proportion of variability in the dependent variable
was accounted for by the independent variable. The Phi-Square Coefficient was

computed by means of a calculator.

Coder training

Before starting their task, the coders were trained until they reached a sufficient

 

See Richards (1998) for further explanation of the Chi-Square test.

See Richards (1998) for further explanation of the Pearson’s Phi Coefficient and
Pearson’s Phi-Square Coefficient.

39

degree of inter-coder reliability. Agreement on coding the independent and dependent
variable was recorded by calculating the Ratio of Coding Agreementsll (Holsti, 1969) and

Cohen ’s Kappa12 (1960). It is important to report the Ratio of Coding Agreements
together with Cohen’s Kappa, because the Ratio of Coding Agreements “does not take
into account the extent of inter-coder agreement which may result from chance (Bennett,
Alpert, & Goldstein, 1954)” (Holsti, p. 140).

In the first training session, in which the coders were instructed to read a coder
protocol13 , 23 pre-constructed examples were coded. A Ratio of Coding Agreements of

.96 was recorded14 for coding the level of ambiguity and a Cohen’s Kappa of .83. For the
beneficial/detrimental coding decisions, a Ratio of Coding Agreements of .87 was
recorded and a Cohen’s Kappa of .60.

On the second day of training, after a discussion of the decisions on which the
coders disagreed in the first test, a second test was conducted, using 26 pre—constructed

recording units. This time, the Ratio of Coding Agreements for the agreement on level of

 

The Ratio of Coding Agreements=2M/N1+N2. “M” is the number of coding decisions on
which the two coders agree. “N1” and “N2” refer to the number of coding decisions that
coder one and coder two make, respectively.

Cohen’s Kappa=Po-Pe/ l-Pe. “Po” refers to the degree of observed agreement (which is
exactly the same as calculating the Ratio of Coding Agreements); “Pe” refers to the

degree of expected agreement between the coders based on chance (Riffe, Lacy, & Fico,
1998).

The protocol has not been included in this thesis, because all the instructions and
examples used in the protocol are included in the present chapter.

The cross-tabulation function of SPSS was used to compute Kappa. For each coder, the
numerical codes (1,2) belonging to the characters (L, H/D, B) in the margin were entered
in separate columns for the independent and dependent variable. A calculator was used
to compute the Ratio of Coding Agreements.

40

ambiguity was .92, while Cohen’s Kappa was .71. For the agreement on
beneficial/detrimental decisions, a Ratio of Coding Agreements of .96 was recorded and a
Cohen’s Kappa of .84.

After this second session, the coders started the official coding of the transcripts,
even though a relatively low Cohen’s Kappa for the agreement on beneficial/detrimental
decisions was recorded. Riffe, Lacy, and Fico (1998) mention, in this regard, that
research on concepts that have not been investigated intensely may continue with
reliability levels that are somewhat below the .80 to .90 range. Inter-coder reliability
levels below .70 are unacceptable, however, because they are hard to interpret and it
becomes difficult to replicate the study.

The coders mainly disagreed, because they coded whether an interviewee thought
ambiguity had been or was beneficial or detrimental to implementation in terms of what
the interviewee thought would have been beneficial or detrimental for the implementation
of the initiative, not in terms of what had actually happened according to the interviewee.
This issue was included as one of the “important notes” in the coder protocol.
Furthermore, the coders needed time to learn to code the recording units in terms of the
broader context of the interview answer. The coders were reminded about this issue in the

“important notes” section of the coder protocol.

41

CHAPTER FIVE

RESULTS

This chapter presents the results of the quantitative content analysis in cross-tables. First,
however, the inter-coder reliability and intra-coder reliability measures are presented.
Several tests are conducted to analyze the significance of the hypothesized associations
between perceived ambiguity in meaning and the perceived effect of this ambiguity on

the implementation of the studied initiative in Calhoun, Muskegon, and St. Clair.

Coder reliability

Inter-coder reliability

In total, 64 recording units, from five duplicated transcripts, were coded by both
coders. The Ratio of Coding Agreements between the coders was .97 for the independent
variable, while Cohen’s Kappa was .86. With regard to the dependent variable, the Ratio
of Coding Agreements between the coders was .95. Cohen’s Kappa was .70.

The results of the inter-coder reliability tests showed high Ratios of Coding
Agreements for each variable. In two out of 64 cases, the coders disagreed when coding
the independent variable. In three out of 64 cases, they disagreed when coding the
dependent variable. A substantial difference in the measurement of Cohen’s Kappa was

found. This difference indicates a low, but still acceptable, reliability-level for the

42

dependent variable. Because the distribution of recording units was highly skewed (i.e.,
most of the recording units were coded as “High/Detrimental”), the probability that the
coders agreed by chance was high; thus, disagreeing even three times decreased Cohen’s

Kappa substantially.

Intra-coder reliability

In total, 36 recording units were coded twice by each coder. For each coder, a
Ratio of Coding Agreements of 1.0 and a Cohen’s Kappa of 1.0 were recorded for the
independent variable. For the dependent variable, also a Ratio of Coding Agreements of
1.0 and a Cohen’s Kappa of 1.0 were recorded for each coder. The intra-coder reliability
measures indicate that the coders were completely consistent with their own coding over

time.

Content analysis results

The results of the content analysis were cross-tabulated in order to compare the
number of recording units per cell (indicated as “cell counts”) and cell percentages (the
number of recording units in a cell/the total number of recording units”). Table 4. shows
that low ambiguity in the meaning of the initiative was perceived as detrimental to the

initiative’s implementation in 5.323% (n=14/N=263) of the cases, whereas in 8.365%

 

15 The character “n” indicates the number of recording units in a cell; “N” indicates the

total number of recording units for the three communities or per community.

43

(n=22/N=263) of the cases, low ambiguity was perceived as beneficial. On the other
hand, in 84.411% (n=222/N=263) of the cases, high ambiguity was perceived as
detrimental and in 1.901% (n=5/N=263) of the cases high ambiguity was perceived as
beneficial.

The outcomes provide evidence that high ambiguity was associated with a
negative effect on implementation. This evidence is supported by the outcomes of the
Pearson’s Chi-Square test (see Table 5.), the Pearson’s Phi Coefficient test (see Table 6.),

and the Pearson’s Phi-Square Coefficient test (see Table 6.). For the content analysis data,
Pearson’s Chi-Square was found to be significant (X2=117.048, 1 df, p=.000). This

implies that the dependent variable significantly depended on the independent variable.
Furthermore, a significant, negative, Pearson’s Phi Coefficient (CI): -.667, p=.000) was
found. This suggests that there was a relatively strong negative correlation between the

independent and dependent variable. The calculation of Pearson’s Phi-Square ((D2=.450)

suggests that a relatively large proportion (45.000%) of the variability in the dependent
variable was accounted for by the independent variable. By comparing the number of
recording units per cell, it is shown that instances of high ambiguity in the meaning of the
initiative occurred more frequently than instances of low ambiguity.

Although these findings suggest a causal relationship between the independent
and dependent variable, conclusions about causality need to be regarded with caution,
because the correlations that were found do not imply causation (Riffe, F ico, & Lacy,
1998). Causality cannot be inferred, because (1) the independent variable was not
controlled, (2) rival explanations were not controlled, and (3) time-order was not

investigated by means of an experimental pretest posttest design.

44

Table 4.

Ambiguity by Implementation Cross- Tabulation

 

 

 

 

 

 

 

 

 

Implementation
Detrimental Beneficial Total
Ambiguity
Low Count 14 22 36
Percentage 5.323% 8.365% 13.688%
High Count 222 5 227
Percentage 84.41 1% 1.901% 86.312%
Total Count 236 27 263
Percentage 89.734% 10.266% 100.0%
Table 5.
Pearson ’s Chi-Square T est for Ambiguity by Implementation
Value df Asymp. Sig.
(2-sided)
Pearson’s Chi-Square 117.048 1 .000*
N of Valid Cases 263
*p < .05
Table 6.
Pearson 's Phi(-Square) Coefficient T est for Ambiguity by Implementation
Value Approx.
Sig.
Pearson’s Phi Coefficient -.667 .000*
Pearson’s Phi-Square Coefficient .450
N of Valid Cases 263
*p < .05

45

Besides testing the overall relationship between the degree of perceived ambiguity
in the meaning of the initiative and the perceived effect of this ambiguity on
implementation, each community’s data were separately analyzed to detect possible
differences between communities.

In Calhoun, low ambiguity was perceived as detrimental in 5.983% (n=7/N=117)
of the cases, whereas in 9.402% (n=11/N=117) of the cases, low ambiguity was perceived
as beneficial. In 82.051% (n=96/N=117) of the cases high ambiguity was perceived as
detrimental and in 2.564% (n=3/N=117) of the cases, high ambiguity was perceived as
beneficial (see Table 7.). These outcomes suggest consistency with the combined results

for the three communities. The evidence for this consistency is substantiated, because
Pearson’s Chi-Square was found to be significant (X2=48.775, 1 df, p=.000) (see Table
8.). Furthermore, a significant, relatively strong Pearson’s Phi was found ((1): -.646,

p=.000) (see Table 9.). The independent variable accounted for 42.000% (<I>2=.420) of the

variability in the dependent variable (see Table 9.).

In Muskegon, low ambiguity was perceived as detrimental in 4.478% (n=3/N=67)
of the cases, whereas in 5.970% (n=4/N=67) of the cases, high ambiguity was perceived
as detrimental. In 86.567% (n=5 8/N=67) of the cases, high ambiguity was perceived as

detrimental and in 2.985% (n=2/N=67) of the cases, high ambiguity was perceived as

beneficial (see Table 7.). A significant Chi-Square (X2=22.262, 1 df, p=.000) was
calculated (see Table 8.) and Pearson’s Phi was significantly strong (CI): -.576, p=.000)
(see Table 9.). The independent variable accounted for 33.200% (<I>Z=.332) of the

variability in the dependent variable (see Table 9.).

46

In St. Clair, low ambiguity was perceived as detrimental in 5.063% (n=4/N=79) of
the cases, whereas in 8.861% (n=7/N=79) of the cases, low ambiguity was perceived as
beneficial. In 86.076% (n=68/N=79) of the cases, high ambiguity was perceived as
detrimental and in .000% (n=0/N=79) of the cases, high ambiguity was perceived as
beneficial (see Table 7.). These outcomes show the strongest evidence for the association
between ambiguity in the meaning of the initiative and the effect of this ambiguity on

implementation. Again, this evidence was supported by a significant Chi-Square
(X2=47.48O, 1 df, p=.000) (see Table 8.) and a quite strong negative, significant Pearson’s
Phi (<I>= -775., p=.000) (see Table 9.). In St. Clair, the independent variable accounted for
60.100% (432:.601) of the variability in the dependent variable (see Table 9.).

In sum, although there were differences between the three communities,
significant results were found in all three cases for the same association between the
independent and dependent variable. The strengths of the associations between the
communities might be different, because the number of recording units per community
varied considerably. In addition, the relatively low number of recording units that were
coded might have contributed to these differences. In line with the combined results for
the three communities, instances of high ambiguity in the meaning of the initiative

occurred more frequently than instances of low ambiguity in each community.

47

Table 7.

Ambiguity by Implementation by Community Cross- Tabulation

 

 

 

Implementation
COMMUNITY Detrimental Beneficial Total
CALHOUN
Ambiguity
Low Count 7 l l 18
Percentage 5.983% 9.402% 15.385%
High Count 96 3 99
Percentage 82.051% 2.564% 84.615%
Total Count 103 14 l 17
Percentage 88.034% 1 1.966% 100.0%
MUSKEGON
Ambiguity
Low Count 3 4 7
Percentage 4.478% 5 .970% 10.448%
High Count 58 2 60
Percentage 86.567% 2.985% 89.552%
Total Count 61 6 67
Percentage 91 .045% 8.955% 100.0%
ST. CLAIR
Ambiguity
Low Count 4 7 l 1
Percentage 5.063% 8.861% 13.924%
High Count 68 0 68
Percentage 86.076% .000% 86.076%
Total Count 72 7 79
Percentage 91 . 139% 8.861% 100.0%

 

48

Table 8.
Pearson ’s Chi-Square T est for Ambiguity by Implementation per Community

 

 

 

 

 

COMMUNITY Value df Asymp. Sig.
(2-sided)
CALHOUN
Pearson’s Chi-Square 48.775 1 .000*
N of Valid Cases 117
MUSKEGON
Pearson’s Chi-Square 22.262 1 .000”
N of Valid Cases 67
ST. CLAIR
Pearson’s Chi-Square 47.480 1 .000***
N of Valid Cases 79
*p < .05
**p < .05
***p < .05

49

Table 9.

Pearson ’s Phi(-Square) Coefficient Test for Ambiguity by Implementation per Community

 

 

 

 

COMMUNITY Value Approx. Sig.
CALHOUN
Pearson’s Phi Coefficient -.667 .000*
Pearson’s Phi-Square Coefficient .420
N of Valid Cases 117
MUSKEGON
Pearson’s Phi Coefficient -.576 .000**
Pearson’s Phi-Square Coefficient .332
N of Valid Cases 67
ST. CLAIR
Pearson’s Phi Coefficient -.775 .000***
Pearson’s Phi-Square Coefficient .601
N of Valid Cases 79
*p < .05
**p < .05
***p < .05

50

CHAPTER SIX

CONCLUSION

In this last chapter, the results of this study are reviewed. The empirical findings are
interpreted in terms of the theoretical concepts that were explicated in chapters one and
two. Furthermore, the methodological and theoretical limitations of this study are

discussed in conjunction with implications for future research.

Review of results

Based on the content analysis of the transcripts, the following can be concluded:

In the decentralized diffusion system of study, potential adopters perceived high
ambiguity in the meaning of the studied social change initiative as detrimental to
the initiative’s implementation.

Although there were differences between the three communities, significant
results were found in all cases for the same association between the independent and
dependent variable. The differences in the strengths of the associations between the
communities may have been caused by the variability of the number of recording units
across the three communities. Furthermore, the relatively low number of recording units

that were coded might have contributed to these differences. Causality cannot be inferred

51

from any of these outcomes, because (1) the independent variable was not controlled, (2)
rival explanations were not controlled, and (3) time-order was not investigated by means
of an experimental pretest posttest design. In each community, instances of high

ambiguity in the meaning of the initiative occurred more frequently than instances of low

ambiguity.

Theoretical implications

Conceptualization of ambiguity

Initially, ambiguity was operationalized according to Goffman’s (1974) and
Weick’s (1995) dimensions, termed “multiplicity in meaning” and “uncertainty in
meaning”. However, during the training of the coders, this operationalization was
changed, because the coders could not reliably discriminate between these two
dimensions. The coders asked how they could know whether or not a struggle over the
definition of the initiative could be attributed to the fact that the Foundation did not
provide a definition (which would be an indication of high uncertainty in meaning) or to
the fact that multiple meanings existed. Also, difficulties concerning causality impeded
coding the recording units according to the two proposed dimensions. It was questioned
whether low multiplicity in meaning implied low uncertainty in meaning, whether low
uncertainty implied low multiplicity in meaning, whether high multiplicity in meaning
implied high uncertainty in meaning, and/or whether high uncertainty implied high

multiplicity in meaning.

52

The questions that arose during coder training lead to an operationalization of
ambiguity without distinguishing between dimensions. Uncertainty and multiplicity in
meaning were treated as different conceptual aspects of ambiguity instead of different
dimensions. The questions that the coders raised offer an opportunity for reconsidering
the validity of Goffman’s (1974) and Weick’s (1995) initial conceptualization. It can be
questioned whether multiplicity and uncertainty in meaning mutually exclude each other
or converge to such a great extent that they cannot be considered separate dimensions of
ambiguity. As far as causality is concerned, it is questionable to what extent the two
aspects of ambiguity co-vary and whether or not there is a particular time-order in which

these aspects occur.

Re-consideration of the validity of the strategic ambiguity argument

Eisenberg (1984) defines ambiguity as “non-clarity”, that is, as a communication
process variable. Clarity refers, in this regard, to “a continuum which reflects the degree
to which a source has narrowed down the possible interpretations of a message and
succeeded in achieving a correspondence between his or her intentions and the
interpretation of the receiver” (p. 229-230).

In this thesis, ambiguity has been defined as an outcome of the communication
process. The meaning of an object, idea, event, or action (in this thesis, the initiative of
study) is ambiguous, because its meaning is defined more or less dissimilarly by the
source and the receiver based on their divergent frames. Furthermore, the organizational
approach that the Foundation toke toward the communication process with the potential

adopters, suggests that strategic ambiguity also was a process variable. The fact that the

53

Foundation chose to be only somewhat directive in the definition of the initiative’s
meaning by organizing the inter-organizational initiative in a decentralized way, indicates
a strategic ambiguity approach. Strategic ambiguity, therefore, seems to have been an
outcome variable and a process variable, in the case of the investigated initiative.

With regard to the issue whether ambiguity is uncertainty in meaning and/or
multiplicity in meaning, Eisenberg’s (1984) conceptualization suggests that ambiguity
pertains to not defining something clearly, but also to defining something in multiple
ways, regardless of whether ambiguity is a communication process or outcome variable.

The results of the content analysis suggest that the argument that ambiguity can be
productive, if used strategically, may be questioned. The validity of the argument can be
discussed in terms of the following three assumptions, made by Eisenberg (1984):
Strategic ambiguity “(1) promotes unified diversity, (2) facilitates organizational change,
and (3) amplifies existing source attributions and preserves privileged positions” (p. 227).

First of all, Eisenberg (1984) argues that strategic ambiguity fosters agreement on
abstractions without limiting specific interpretations in organizations. Furthermore, he
argues that the strategic use of ambiguity fosters the existence of multiple viewpoints
through the creative use of organizational symbols.

In the interviews, some potential adopters reported that multiple viewpoints

impeded the implementation process. One interviewee from Muskegon said:

...My understanding of what the Foundation was thinking about was an integrated
method of delivering services to people and when they said integrated they meant
an integration that included human needs, social needs, as well as strictly medical
needs. As we talked about integrated delivery systems, I found that at least two-
thirds of the people involved in our conversation thought of integrated delivery
meaning integrating let’s say, inpatient care, outpatient care, home health services,

54

pharmacy services. In other words, they were thinking of ways to integrate the
delivery of purely medical services as opposed to integrating human services. And
until it finally dawned on me that not everyone was seeing integration the way
that I was seeing integration I was having a lot of trouble talking...

In addition, when it came to defining abstract concepts such as “consumer”,
“payer”, or “provider” of health care, some interviewees reported that this definitional
was difficult, time-consuming, and de-motivating. One of the potential adopters from

Calhoun reported:

...One of the difficulties that we ran into later on. . .that manifested itself on the
board [i.e., the community’s governing board consisting of payers, providers, and
consumers of health care] was the arbitrary decision on definitions. What is a
consumer? What is a provider? What is a payer? And we had people who had
great difficulty saying well wait a minute, you call me a provider but I’m also a
consumer.

In the second place, Eisenberg (1984) assumes that ambiguity can be used
strategically to facilitate organizational change. In this regard, organizational goals are to
be expressed arnbiguously in order to allow flexible alteration of operations. Furthermore,
Eisenberg (1984) implies that strategic ambiguity can “facilitate relational development
through the emergence of a restricted code to which only certain individuals are privy” (p.
233)

With regard to defining specific goals, the interview data suggested that some
potential adopters thought that not having well-articulated goals impeded the
implementation of the investigated initiative. One interviewee from St. Clair said the

following:

55

I think they [i.e., the Foundation] have to be more articulate in their goal... and be
able to back it up all the way down the line. If it changes a little bit, I think you
have to change the goal. You have to make it clear that this is what we thought,
it’s going in this direction, and this is what it is now. . .Usually, when people write
grants, they want things, specific things. This [i.e., the initiative that was studied
in this thesis] was not that Specific and it was a wonderful idea, but it was either
too broad or it was too vague.

In addition, some interviewees said that they did not know how to implement the

initiative. One interviewee from St. Clair, for example, said:

...I think if we had had a list of 20 areas that were acceptable kinds of things that
they [i.e., the Foundation] wanted us to do, we probably would have saved a year
and a half of peoples time and we could have, would have been much more
effective. We would have had a much better concept of the thing. . .they had in
mind.

Eisenberg’s (1984) writing about the use of a restricted code suggests that frame
convergence between people in an organizational system is beneficial to social change.
However, the case of the present inter-organizational initiative indicates that, when an
organizational system is decentralized and consists of different subsystems, people may
not communicate beyond a frame that is restricted to their own subsystem. In this regard,
some of the interviewed potential adopters spoke about the Foundation as an
organizational system that was almost removed from their own system.

In the third place, Eisenberg (1984) argues that highly ambiguous communication
is easier to deny than communication that has a low degree of ambiguity. Disclosing
information in an unequivocal way may limit options and prematurely endanger plans

(Bok, 1983). Contrary to this assumption, several potential adopters of the studied

56

initiative thought that the Foundation had a hidden agenda and perceived not disclosing
information as detrimental to the implementation of the initiative. An interviewee from

Muskegon said:

...I don’t think that Kellogg was very effective at saying-it’s wide open, you guys
can do almost anything that you want. Just the whole review process, you know-
you submit a budget, they send it back and you’d submit it and they’d send it
back. And you had the feeling that there was some hidden agenda there and they
needed to be a little straightforward about that.

In sum, the case of the studied initiative provides empirical evidence that
contradicts the three main assumptions made by Eisenberg (1984). The findings suggest
that, even in a decentralized diffusion system, potential adopters of a social change
initiative prefer adopting a clear and unified definition of the initiative’s meaning over
receiving great freedom in the definitional process. Potential adopters not only demand
more specific information about the implementation in terms of operational definitions,
but also need to have a clear concept of what the initiative is. For this reason, it is
important that the change agency and the potential adopters establish a common frame for
defining the meaning of the social change initiative in order to achieve implementation

SUCCCSS.

Reviewing the diffusion process in terms of the studied initiative
In line with Rogers’ (1995) model of the innovation process in an organization
(see Figure 1.), the implementation of the studied initiative can be conceptualized as a

sub-process of the initiative’s diffusion in the three communities (see Figure 4.). Figure 4.

57

 

I. INITIATION OF THE
INITIATIVE

#1 . Agenda-setting:
General problem that
in health care decision-
making the voice of the
people is lost was
defined

#2. Matching:
General problem was fit
with potential partnering
communities

 

DECISION TO ADOPT
THE INITIATIVE

 

 

II. IMPLEMENTATION OF
THE INITIATIVE

#3. Redefining/
Restructuring:
Communities created
organizational
structures such as the
governing boards and
program offices

#4. Clarifying:
The initiative needed to be
put into more widespread
use so that its meaning could
become more established

#5. Routinizing:
The initiative needed to
become more incorporated
into the communities’
activities and lose its

V separate identity

 

Figure 4. Model of the studied initiative’s innovation process based on Rogers’ (1995) model of
the innovation process in an organization.

58

starts with the agenda-setting and matching stages in the initiation of the initiative.
Agenda-setting involved the conceptualization of the Foundation’s mission. This mission
targeted grant making to provide increasing access to integrated, comprehensive health
care systems organized around public health, prevention, and primary care. In the
matching stage, the actual initiative was developed and introduced to the three
communities.

After having decided to adopt the initiative as a way of “doing business”, the
redefining and restructuring stage, in which the communities instituted organizational
structures like the community governing boards and program offices, was not considered
to be particularly problematic by the potential adopters. Activities involving the
clarification of the meaning of the initiative, however, and the initiation of the intended
social change in the communities at-large, were perceived as complicated tasks that
hindered implementation.

Particularly unclear understanding about what was expected in terms of re-
designing the communities’ health care systems, indicating a high degree of ambiguity in
the meaning of the initiative, impeded implementation. The Foundation’s desire to remain

non-prescriptive may have hampered the implementation in this case.

59

Research limitations and implications for future research

Generalizing from a small number of words to a small number of individuals to a
community

The content analysis of words used in discourse generally involves the coding of
hundreds or thousands of words. In this study, a relatively small number of words, spoken
by individuals, were used as recording units. In some interviews, more recording units
were found than in other. On an average, five to ten recording units were found per
interview. Also, some interviews were longer than others. The average number of pages
per transcript was between 20 and 25, but there were also transcripts of 50 pages.

For these reasons, the results may have been influenced more by the perceptions
of those interviewees, who were more verbose and/or thought ambiguity was an
important issue. Because all data were analyzed together, the possibly confounding
influence of verbosity and amount of recording units used per interview were not taken
into consideration.

The inference that most interviewees perceived ambiguity as detrimental must be
regarded carefully, because the total number of recording units may not be a true
reflection of the average perceptual convergence between the interviewees. This problem
becomes more serious, if the results are used to make an inference about the entire
community, because only a small number of people’s perceptions may have emerged

through the way this study was operationalized.

60

Empirically investigating frame convergence and divergence

In this study, the meaning definition process has remained empirically
uninvestigated, although it has functioned as an important basis for making inferences.
The meaning definition process will need to be studied experimentally, in order to make
valid assumptions about causality.

One could imagine, for example, a study in which a homogeneous group of people
is compared with a heterogeneous group on the way in which they define the meaning of
an object, idea, event, or action (both groups receiving the same instructions and
background information) at different points in time. At T0, no establishment of similar
meaning would be expected; at T1, after communicating with each other for one hour,
supposedly some common meaning would be noticed; at T2, after communicating five
hours, even more similarity in meaning (than at T1) would probably be seen. At the
various points in time, each individual’s frame used for defining meaning could be
investigated by means of a short survey and it would be interesting to see if the frames

converge over time in the homogeneous group and if they diverge in the heterogeneous

group.

Generalizing across studies

In order to increase the confidence in the findings of this thesis, the study will
need to be replicated. In addition, increasing the generalizability of this study’s results, by
comparing outcomes of different studies on decentralized change initiatives that are

similar to the social change initiative of study, will also be important.

61

Opportunities for studying similar initiatives should present themselves, because,
besides private foundations such as the W. K. Kellogg Foundation, the Annie E. Casey
Foundation, and the Robert Wood Johnson Foundation, US. government agencies like
the US. Agency for Toxic Substances and Disease Registry, the Department of the
Interior, the Department of Agriculture, the National Oceanic and Atmospheric
Administration, the Environmental Protection Agency, and the Department of Energy, are
experimenting with change initiatives to increase community engagement in policy
decision making. To increase the generalizability of findings across studies, and to learn
more about the causal relationship between frame divergence and the implementation
success of community change initiatives, it will be important to use the opportunities that
are presenting themselves to conduct field experiments with several communities,

involving a multitude of people.

62

REFERENCES

63

 

Bateson, G. (1972). Steps to an ecology of mind. New York: Ballentine Books.

Bennett, E. M., Alpert, R., & Goldstein, A. C. (1954). Communications through limited
response questioning. Public Opinion Quarterly, 18, 303-308.

Berger, P., & Luckman, T. (1966). The social construction of reality. Garden City, New
York: Doubleday.

Bok, S. (1983). Secrets: 0n the ethics of concealment and revelation. New York:
Pantheon Books.

Bolman, L. G., & Deal, T. E. (1991). Reframing organizations: Artistry, choice, and
leadership. San Francisco, CA: Jossey-Bass.

Blumer, H. (1969). Symbolic interactionism: Perspective and method. Englewood Cliffs,
NJ: Prentice-Hall.

Cohen, J. A. (1960). Coefficient of agreement for nominal scales. Educational and
Psychological Measurement, 20, 37-46.

Donohue, W. A. (1991). Communication, marital dispute, and divorce mediation.
Hillsdale, NJ: Lawrence Erlbaum.

Drake, L. E., & Donohue, W. A. (1996). Communicative framing theory in conflict
resolution. Communication Research, 23, 297-322.

Durkheim, E. (1950). The rules of sociological method. Glencoe, IL: Free Press.

Eisenberg, E. M. (1984). Ambiguity as strategy in organizational communication.
Communication Monographs, 51, 227-242.

Eveland, J. D. (1987). Diffusion, technology transfer, and implementation: Thinking and
talking about change. Knowledge: Creation, Diffusion, Utilization, 8, 303-322.

Fiske, J. (1997). Introduction to communication studies (2nd ed.). London: Routledge.

64

Gamson, W. A., & Wolfsfeld, G. (1993). Movements and media as interacting systems.
Annals of the American Academy of Political and Social Sciences, 528, 114-125.

Goffman, E. (1974). Frame analysis. New York: Harper and Row.

Hare, A. P., & Blumberg, H. H. (1988). Dramaturgical analysis of social interaction.
New York: Praeger Publishers.

Holsti, O. R. (1969). Content analysis for the social sciences and humanities. Menlo
Park, CA: Addison-Wesley Publishing Company.

Johnson, J. D. (1997). A multiple level approach to innovation implementation.
Unpublished paper submitted for possible’publication in the Special Topic Forum
of Academy of Management Review focusing on Multilevel Theory and Research
Approaches, November 1997.

Johnstone, B. (1989). Linguistic strategies and cultural styles for persuasive discourse. In
S. Ting-Toomey & F. Korzenny (Eds), Language, communication, and culture
(pp. 139-156). Newbury Park, CA: Sage Publications.

Kotler, P., & Roberto, E. L. (1989). Social Marketing. New York: Free Press.

Lax, D. A., & Sebenius, J. K. (1986). The power of alternatives or the limits of
negotiation. Negotiation Journal, 1, 163-179.

Leonard-Barton, D. (1988). Implementation as mutual adaptation of technology and
organization. Research Policy, 17, 251-267.

Manning, P. K. (1992a). Organizational communication. New York: Aldine de Gruyter.

Manning, P. K. (1992b). Technological dramas and the police: Statement and

counterstatement in organizational analysis. Criminology, 30, 327-346.

65

Mead, G. H. (1934). Mind, self and society from the standpoint of a social behaviorist.
Chicago, IL: University of Chicago Press.

Men'iam-Webster’s Collegiate Dictionary (10th ed.). (1998). Springfield, MA: Merriam-
Webster.

Men'iam-Webster’s Collegiate Thesaurus (1993). Springfield, MA: Merriam-Webster.

Pacanowsky, M. E., & O’Donnell-Trujillo, N. (1983). Organizational communication as
cultural performance. Communication Monographs, 50, 126-147.

Perinbanayagam, R. S. (1974). The definition of the situation: An analysis of the
ethnomethodological and dramaturgical view. The Sociological Quarterly, 15,
521-541.

Pillar, P. R. (1983). Negotiating peace: War termination as a bargaining process.
Princeton, NJ: Princeton University Press.

Pruitt, D. G., & Rubin, J. Z. (1986). Social conflicts: Escalation, stalemate, and
settlement. New York: Random House.

Putnam, L. L. (1982). Paradigms for organizational communication research: An
overview and synthesis. The Western Journal of Speech Communication, 46, 192-
206.

Putnam, L. L., Wilson, S. R., Waltman, M. S., & Turner, D. (1986). The evolution of case
arguments in teachers’ bargaining. Journal of the American Forensic Association,
23, 63-81.

Richards, W. D., Jr. (1998). The zen of empirical research. Cresskill, NJ: Hampton Press.

Riffe, D., Lacy, S., & Fico, F. G. (1998). Analyzing media messages: Using quantitative

content analysis research. Mahwah, NJ: Lawrence Erlbaum.

66

Rogers, E. M. (1995). Diffusion of innovations (4th ed.). New York: Free Press.

Schon, D. A. (1967). Technology and change: The new Heraclitus. New York: Delacorte
Press.

Snow, D. A., Rochford, E. B., Jr., Worden, S K., & Benford, R. D. (1986). Frame
alignment process, micromobilization, and movement participation. American
Sociological Review, 51, 464-481.

Statistical Package for the Social Sciences for Windows Student Version 7.5 [computer
software]. (1996). Chicago, IL: SPSS Inc.

Stebbins, R. A. (1986). The definition of the situation: A review. In A. Fumham (Ed),
Social behavior in context (pp. 134-154). Boston: Allyn and Bacon.

Sussman, L. (1975). Communication in organizational hierarchies: The fallacy of
perceptual congruence. The Western Journal of Speech Communication, 3, 191-
199.

Watanabe, S. (1993). Cultural differences in framing: American and Japanese group
discussions. In D. B. Tannen (Ed.), Framing in discourse (pp. 176-209). New
York: Oxford University Press.

Weick, K. E. (1979). The social psychology of organizing (2nd ed.). New York:
McGraw-Hill.

Weick, K. E. (1995). Sensemaking in organizations. Thousand Oaks, CA: Sage

Publications.

67

APPENDIX A

68

Table 3.

List of Underlined Words not Included in Table 1. & 2.

 

Appears in Table 1. & 2. as

(Also) appears in text as

 

accurate (adjective)

agitation (noun)
anxiety (noun)
apprehend (verb)
blind (verb)

bother (noun)

challenge (verb)
clarify (verb)
clean-cut (adjective)
cloud (verb)
complicate (verb)

conceive (verb)
concern (noun)
confuse (verb)

controversial (adjective)
correct (adjective)
define (verb)

difficult (adjective)
distinct (adjective)
distrust (verb)
esoteric (adjective)
explain (verb)

fog (verb)
formulate (verb)

frustrate (verb)

graspable (adjective)

get (verb) the idea (noun)
hesitate (verb)

hide (verb)

indecisive (adjective)
indefinite (adjective)

69

inaccurate (adjective, antonym)/
inaccuracy (noun, antonym)
agitator (noun)

anxious (adjective)
apprehension (noun)

blind (adjective)/

blindsided (adjective)

bother (verb)/

bothersome (adjective)
challenge (noun)
clarification (verb)

clean (adjective)

cloud (noun)

complicated (adjective)/
complication (noun)
pre-conceive (verb)

concern (verb)

confusion (noun)/

confused (adjective)/
confusing (adjective)
controversy (noun)

incorrect (adjective, antonym)
definition (noun)/

redefine (verb)

difficulties (noun)

distinction (noun)

distrust (noun)

esoterica (noun)

explanation (noun)

foggy (adjective)
formulation (noun)/ formula (noun)/
reformulate (verb)

frustrated (adjective)/
frustration (noun)

grasp (verb)

yield (verb) the idea (noun)/
come to grips with (verb) (the
idea) (noun)/ idea (noun)
hesitant (adjective)/
hesitancy (noun)

hidden (adjective)

decisive (adjective, antonym)
definite (adjective, antonym)

insecure (adjective)
intent (noun)
interpret (verb)

knowable (adjective)

lay (adjective)
meaning (noun)
muddy (verb)

mysterious (adjective)

open (adjective)

precise (adjective)
questionable (adjective)
realize (verb)

recognize (verb)

reserve (noun)
simple (adjective)
skepticism (noun)
specific (adjective)

straightforward (adjective)

sum (noun) and substance (noun)
sure (adjective)

suspicion (noun)
trouble (noun)
undeniable (adjective)
unexplicit (adjective)
worry (noun)

secure (adjective, antonym)/
secure (verb, antonym)
intentional (adjective)/
intend (verb)

interpretation (noun)/
misinterpretation (noun)
know (verb)/knowledge (noun)/
lcnowledgeable (adjective)/unknown
(adjective)/knew (verb)
layman (noun)

meaningful (adjective)
muddy (adjective)/

muddled (adjective)
mystery (noun)/

demystify (verb)

open (verb)/ re-open (verb)
openness (noun)/open-
ended (adjective)

precision (noun)

question (noun)/question mark (noun)
realization (noun)
recognition (noun)/
recognizable (adjective)
reserve (verb)

simplification (noun)
skeptical (adjective)
specification (noun)/
specifics noun)

straight (adjective)/
straighten (verb)

substance (noun)

assure (verb)/ensure (verb)
reassure (verb)

suspicious (adjective)
trouble (verb)

deny (verb)

explicit (adjective, antonym)
worry (verb)

 

7O

    

IIIIIIIIIIIIIIIIIIIIIIII

11111 11112111 1111111 11111115 111111 11111111 1171 11