Citation
When what I think depends on who I am

Material Information

Title:
When what I think depends on who I am the role of social identity in consumer attitude formation
Creator:
Reed, Americus, 1970-
Publication Date:
Language:
English
Physical Description:
x, 149 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Consumer attitudes ( jstor )
Consumer research ( jstor )
Daughters ( jstor )
Handwriting ( jstor )
Psychological attitudes ( jstor )
Saliency ( jstor )
Social identity ( jstor )
Social psychology ( jstor )
Sons ( jstor )
Statistical relevance model ( jstor )
Dissertations, Academic -- Marketing -- UF ( lcsh )
Marketing thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2000.
Bibliography:
Includes bibliographical references (leaves 139-148).
Additional Physical Form:
Also available online.
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Americus Reed.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
026061292 ( ALEPH )
48449359 ( OCLC )

Downloads

This item has the following downloads:


Full Text










WHEN WHAT I THINK DEPENDS ON WHO I AM:
THE ROLE OF SOCIAL IDENTITY
IN CONSUMER ATTITUDE FORMATION















By

AMERICUS REED II













A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2000
































Like an iron fist in a velvet glove, you don't know what I'm made of.


Neil Peart














ACKNOWLEDGMENTS

The completion of this dissertation would not have been possible without the

assistance of several key people. First, I would like to thank my mother, father and sister

for their ongoing love and familial support. I would also like to thank my beautiful wife

Karen--whom I love dearly--for her timely encouragement, love and patience to put up

with me. Of course, my dissertation chair Joel Cohen has been both an inspiration and an

irreplaceable mentor, patiently guiding this work with a steady hand, spending

meticulous time and effort to facilitate my ongoing development as an aspiring scholar.

In that regard, my committee members must also be acknowledged for their copious and

insightful feedback: I thank Rich Lutz, Alan Sawyer, and Dolores Albarracin for their

commitment to this project and especially Barry Schlenker, who is both an incredible

human being and a scholar.

A Special thanks go to Karl Aquino--long time friend, mentor and colleague--for

taking me under his wing and exposing me to life in academia. I also appreciate the

encouragement and intellectual stimulation provided in my undergraduate consumer

behavior class taught by Lois Mohr, who was the first to introduce me to the interesting

concepts and ideas in consumer research and who was the first to suggest the University

of Florida as a good place to pursue my interests.

Speaking of the University of Florida, I would be remiss if I did not acknowledge

my friendship with two of the greatest people that I know: my life-long companions and

colleagues Lisa Bolton and Kevin Bradford. I should also mention three key people in


iii








the department, Julie Ritter, Jennifer Maynard, and Cathy Koenig for being so kind in

assisting me with all of the "devil in the details." I would like to thank all of my other

colleagues at the University of Florida for the numerous get-togethers, informal

discussions, round-robin basketball tournaments, fireside chats about research and life in

general, and for an academic environment that was both intellectually stimulating and

fun. Last but not least, thanks to Jonesy D. Cat--who does not really care about the

"significance of a value added contribution to the literature," but only that his food dish

remains full--for keeping the apartment animated and stress levels attenuated.





































iv















TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS .............................................. iii

ABSTRACT ......................................................... viii

CHAPTERS

1. INTRODUCTION AND OVERVIEW............................. 1

2. FACTORS THAT AFFECT A CONSUMER'S ATTITUDE ............ 4

Direct Experience with the Object................................. 4
Inferential Attitude Formation Based on Past Behavior ................ 5
Categorization Based Attitude Formation ........................... 5
Analytical Attitude Construction .................................. 6
Personal Values ........ .................. ................... 6
What About Social Identity? ..................................... 7

3. THE DEFINITION OF SOCIAL IDENTITY......................... 9

The Notion of Identity .......................................... 9
Social Identification Processes ................................... 10
Sociological Models ........................................... 11
Social Identity Theory........................................... 12
Social Categorization Theory .................................... 13
Definition of Social Identity and the Current Research ................. 14

4. SOCIAL IDENTITY BASED ATTITUDE FORMATION ............. 16

Is the Social Identity Salient?..................................... 16
Is the Social Identity Self-Important? ............................... 17
Is the Social Identity Object-Relevant? ............................ 19
Does the Social Identity Provide a Basis to Respond? .................. 19
Summary of the Framework...................................... 20

5. EXPERIMENT(S) 1A AND 1B: THE INTERACTIVE ROLE OF
SALIENCE AND OBJECT RELEVANCE...................... 25

Overview of the Studies ........................................ 25
Theoretical Premise of the Studies................................. 25


V








Pretest .................. .................................... 26
Experiment la.......... ................. .................... 32
Experiment Ib .................................................. 35

6. EXPERIMENT 2: THE ROLE OF SALIENCE, SELF-IMPORTANCE
AND OBJECT RELEVANCE ............................... 47

Purpose of the Study.......................... ................ 47
Theoretical Premise of the Study ................................. 47
Experiment 2................................................... 48

7. EXPERIMENT 3: THE ROLE OF SELF-IMPORTANCE AND
EVALUATIVE DIAGNOSTICITY ........................... 60

Purpose of the Study and Hypotheses to be Tested.................... 60
Overview of the Experiment ..................................... 61
Experim ent 3 .. ............................................... 61


8. GENERAL DISCUSSION AND CONCLUSIONS.................... 78

Contributions. ................................................ 84

APPENDICES

A. VISUAL IMAGERY STUDY ................................... 88

B. WORD SEARCH STUDY ..................................... 98

C. CONSUMER SELF DESCRIPTION STUDY........................ 100

D. NEW PRODUCT ASSESSMENT STUDY .......................... 101

E. HANDWRITING STUDY ...................................... 108

F. THE PANAS MOOD SCALE.................................... 119

G. WEB PAGE DESIGN STUDY .................................. 120

H. EXPERIMENT 1B: OMINBUS ANOVA OF EFFECTS OF IDENTITY
AND RELEVANCE ON PURCHASE LIKELIHOOD OF FOCAL
PRODUCT ............................................... 128

I. EXPERIMENT 1B: OMINBUS ANOVA OF EFFECTS OF IDENTITY
AND RELEVANCE ON PURCHASE LIKELIHOOD OF FILLER
PRODUCTS ............................................ 129



vi









J. EXPERIMENT 2: OMINBUS ANOVA OF EFFECTS OF SALIENCE 130
RELEVANCE AND SELF-IMPORTANCE ON PURCHASE
LIKELIHOOD OF FOCAL PRODUCT........................

K. EXPERIMENT 2: OMINBUS ANOVA OF EFFECTS OF SALIENCE
RELEVANCE AND SELF-IMPORTANCE ON PURCHASE
LIKELIHOOD OF FILLER PRODUCTS........................ 131

L. EXPERIMENT 3: OMINBUS ANCOVA OF EFFECTS OF SALIENCE
RELEVANCE AND SELF-IMPORTANCE ON
EVALUATIVE DIMENSIONS ............................... 132


REFERENCES..................... .............................. 139

BIOGRAPHICAL SKETCH ........................... .................. 149



































vii














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements of the Degree of Doctor of Philosophy

WHEN WHAT I THINK DEPENDS ON WHO I AM:
THE ROLE OF SOCIAL IDENTITY
IN CONSUMER ATTITUDE FORMATION

By

Americus Reed II

December 2000

Chairperson: Joel B. Cohen
Major Department: Marketing

Attitudes are viewed as a particularly important intervening variable between a

consumer's awareness of a brand and the intention to purchase that brand. Since

consumers do not buy brands they do not like, the importance of creating favorable

attitudes toward products, brands and companies can hardly be over-emphasized.

However, subsequent to the introduction of multi-attribute models in the late 1960s,

practically all textbooks and journal articles relied on a componential view of consumer

attitudes. These models define attitudes as a weighted assembly of beliefs about the

features or benefits of the product.

It is worth noting that practitioners--who are actually responsible for creating

favorable attitudes toward their product offerings--appear to believe that consumers often

form attitudes in a very different way. Consider the prevalence of social identity oriented

advertising themes (e.g., a brand is held up to be the embodiment of a particular



viii









"lifestyle" associated with consumers grouped into some type of social category).

Moreover, consumers' social identification often becomes the focus of market

segmentation and product positioning strategies. Yet, a favorable attitude toward a

socially symbolized brand that is created using this approach is not well described via a

multi-attribute, feature/benefit-based process.

In some respects, the situation described above might be thought of as a

"disconnect" between scholarly research on consumer attitudes and practitioners'

frequently selected means of creating favorable attitudes. This research attempts to

address this gap. It takes the view that there are multiple bases for attitude formation.

Further, it is important to understand the precise conditions under which a particular basis

for attitude formation is likely to lead to an attitude.

Social identity is hypothesized to serve as a basis for attitude formation when a

social identity is 1) salient, 2) self-important, 3) object relevant, and provides 4) an

evaluative basis to respond. Three laboratory experiments examine the theoretical

underpinnings of this assertion. Evidence from experiment 1(a) suggests that a social

identity's salience can mediate consumers' momentary self-conceptions which when

combined with object relevance (experiment lb) leads to more favorable consumer

attitudes (when adopting a social identity would in fact provide a favorable basis to

respond to the product). Results from experiment 2 demonstrate the moderating role of

self-importance in forming an attitude based on social identification processes and data

from experiment 3 suggest that a social identity's evaluative content is more likely to

impact attitude formation when it is diagnostic for an evaluative response. The totality of

the empirical results, a comparison of the dissertation framework to other frameworks




ix










that describe similar phenomena, and the concluding implications of the framework for

attitude theory and marketing practice are then discussed.


























































x














CHAPTER 1
INTRODUCTION AND OVERVIEW

The "psychological tendency that is expressed by evaluating a particular entity

with some degree of favor or disfavor" (Eagly and Chaiken, 1993, page 1) is a critical

intervening variable between a consumer's awareness of a brand and the decision to buy

it. However, attitude research has been dominated by early multi-attribute frameworks

developed by Fishbein and Ajzen (1975) and others, structural models emphasizing the

cognitive components of attitude (c.f. Lutz, 1975; Rosenberg and Hovland, 1960),

information integration models and economic utility approaches (c.f. Kahn and Meyer,

1991; Lynch, 1985). These models imply a computational process of attitude formation

where in the most general sense, each of the attitude object's components is given a

rating (i.e., on level of performance, strength of association) and a weight (whose

meaning and measurement varies with the model).

Of course there are situations where consumers might form attitudes in this way.

For example, a consumer might purchase a car because she believes the car gets good gas

mileage, is modestly priced and is also safe. Additionally, she thinks it is important for a

car to have those product features. However, this conceptualization does not adequately

describe other processes that might lead to the formation of a consumer attitude. For

instance, marketing practitioners seem to believe that a consumer's social identification

with a particular subculture, ethnic identity, family, gender, set of peers or even

spokesperson, can be the basis for a favorable consumer attitude. This is often reflected



I





2


in "life style" advertising and certainly in various product positioning and market

segmentation strategies.

This intuition demonstrated by marketing practitioners has not totally gone

unnoticed in academic research. Various studies looking at the persuasiveness of

spokespersons, (Deshpande and Stayman 1994) gender and ethnic differences in

advertisements, (Meyers-Levy, 1988; Jaffe, 1991; Forehand and Deshpande, in press)

food consumption, (Wooten, 1995; Stayman and Deshpande 1989; 0' Guinn and Meyer,

1984; Wallendorf and Reilly, 1983; Hirschman, 1981) media usage, (Saegert, Hoover and

Hilger, 1985) brand loyalty / organizational patronage, (Deshpande, Hoyer and Donthu,

1986) and even information search behavior have demonstrated that classifying

consumers based on a particular social categorization can lead to differences in the

descriptions of how consumers behave.

But merely knowing that consumers linked to different social categories

sometimes respond differently sheds very little light on when and how this is likely to

occur. Despite the early work by functional theorists who pointed out that social identity

matters in motivating people to hold certain attitudes, (Shavitt, 1990; Katz, 1960; Smith,

Bruner and White, 1956)--numerous discussions by social scientists regarding reference

group influence in terms of people's willingness and ability to "adopt a standpoint,"

(Turner, 1956)--and more recent interests in social identity and inter-group relations,

(Tajfel, 1981; van Knippenberg and Ellemers, 1990) relatively little work has been done





3


to examine a process by which social identification is likely to lead to the formation of an

attitude' (see Eagly and Chaiken 1993, page 660 on this point).

This research attempts to build on prior work by incorporating social identity

within an attitude framework i.e., outlining the conditions of when and how a consumer's

social identification is likely to be the basis for attitude generation. The rest of this

dissertation is organized as follows. In chapter 2, a brief review is given of the different

factors that affect how a consumer might form an attitude. Chapter 2 concludes with a

discussion of social identity as an under researched alternative basis by which a consumer

may form an attitude. Chapter 3 is a discussion and review of the concept of "social

identity," i.e., how it has been defined across past literatures and how the term social

identity will be used in this dissertation. Chapter 4 is the theoretical background for this

dissertation. It describes a proposed process by which a consumer may form an attitude

based on social identification processes including the precise factors that are likely to

impact the process. Chapters 5 7 discuss the overview, design and results of empirical

work intended to test the theoretical assertions made in Chapter 4. Chapter 8 concludes

with a general discussion of how the empirical results conducted in this dissertation relate

to previous conceptual frameworks that describe similar phenomena and the implications

of this research to attitude theory and marketing practice.







It is always possible to redefine any factor or variable (e.g., social identification) as an
attribute or benefit and then move it into the attitude portion of a multi-attribute model.
Each input is simply reduced to the status of a belief element. However, the model would
not attempt to distinguish between these different inputs and would provide no insight
into the process or when this type of influence is likely to occur.














CHAPTER 2
FACTORS THAT AFFECT A CONSUMER'S ATTITUDE

The consumer attitude literature is replete with various factors that are thought to

impact a consumer's attitude toward an object or issue. These different factors may

reflect different processes by a which a consumer may form an attitude and are briefly

reviewed here to motivate and provide context for the subsequent dissertation work.

Direct Experience with the Object

A consumer might form an attitude on the basis of direct experience with the

attitude object (Regan and Fazio, 1977; Songer-Nocks, 1976; Fazio and Zanna, 1978a,

1978b; Fazio, Chen, McDonel and Sherman, 1982) or by an imagined attempt to

experience the attitude object (Keller and McGill, 1994; Pham, 1998; Schwartz, 1990).

Physical or imagined interaction with the attitude object results in a confluence of

cognitive, affective and perhaps behavioral information. This information can be used as

inputs to object evaluation. For instance, consider a consumer at a Thai restaurant trying

to decide on what dish to consume for dinner. S/he might actually try a sample of some

particular dish on the menu and generate an attitude toward that meal based on direct

experience. As another example, consider the husband and wife who are trying to decide

on a particular locale for their impending vacation. They might imagine themselves in

various situations (sipping drinks on a beach in Jamaica, cozying up to a fire in the snow-

capped mountains of Colorado, etc.). By asking the question "how do I feel about it?,"





4





5


the couple can generate attitudes toward each alternative based on the unique affective

information derived from "trying on the consumer episode" (Pham, 1998).

Inferential Attitude Formation Based on Past Behavior

The work by Bem (1967, 1972) on "self-perception" theory suggests that there

may be instances in which a consumer will infer her attitude based on thinking about past

behavior. For example, consider the consumer who has purchased a brand of dishwasher

repeatedly. Her behavior might have become habitualized in that she no longer thinks

about the purchase decision. She just routinely buys the brand. Yet, when queried as to

her attitude toward the particular brand of dishwasher detergent, she indicates a favorable

attitude because she infers one based on the implications of her consistent behavior. In

these instances, attitudes are a function of some retrospective consideration of sequences

of past behavioral choices.

Categorization Based Attitude Formation

The categorization and consumer learning literatures suggests that a person may

generate an attitude by considering its similarity to other liked or disliked objects (Lingle

and Ostrom, 1979; Fiske and Pavelchak, 1986; Gregan-Paxton and Roedder John 1997).

In these instances evaluative information associated with an attitude object similar to the

attitude object in question is brought to mind as a basis for generating an attitude toward

the specific attitude object. So in the restaurant example, the consumer might already

possess an attitude toward a similar meal (e.g., Chinese food), and based on an anchoring

and adjustment type of categorization mechanism, come up with an attitude toward some

Thai dish on the menu. Or in the other example, the couple may generate evaluative

information toward the Jamaica trip by considering the similarity of an attitude toward a





6


previous trip taken to the Bahamas last summer. In general, the process of classifying an

object and/or thinking about how it relates to some analogous entity can bring to mind

attitudes based on its perceived similarity to other objects for which the consumer holds

preexisting attitudes.

Analytical Attitude Construction

Of course, much of the research in consumer behavior and attitude theory

suggests that an attitude may be analytically constructed. In other words, a process

whereby some set of cognitions and beliefs weighted by their evaluations leads to an

attitude and subsequent behavioral tendency (Fishbein and Ajzen, 1975; Ajzen and

Fishbein, 1980). The combinatorial attitude models imply that evaluative information is

generated when a subset of the universe of beliefs about the object becomes salient.

Depending upon what beliefs become accessible, the attitude can be computed vis a vis

the assessed implications of those beliefs. For example, the person in the restaurant can

generate an attitude toward the Thai dish by considering specific features of the attitude

object (e.g., this dish contains fish, cilantro, a bit of sodium, curry etc.) and the evaluative

implications of those features (I like fish and cilantro, but I am watching my sodium

intake, etc.)

Personal Values

An attitude may also come from one's personal belief system (Rokeach, 1968,

1973, 1980). In these instances, evaluative information comes to mind as a function of

referencing one's personal values or beliefs about the consequences of an emergent

attitude object for one's central value system (Stem, Deitz, Kalof and Guagnano, 1995).

For example, a consumer might possess a higher order value structure that leads to the





7


belief that protecting the environment is an important act because it is "the right thing to

do." Therefore, attitudes can be generated by thinking about the convictions that are

linked to central and basic values espoused by the consumer (Rosenberg, 1956).

What About Social Identity?

In classic attitude research, a number of researchers focused on the motivational

antecedents of attitudes. Ironically, social identity was identified as one important factor

as to why some people hold certain opinions (Katz, 1960; Smith, Bruner and White,

1956). The term social adjustment function, which was later termed more generally as

the "social identity function," (Shavitt, 1990, 1989) denoted how some attitudes can

mediate self-other relations and establish one's social identity:

the function of social adjustment served by holding an opinion is at once more
subtle and more complex. For it is by holding certain views that one identifies
with, or indeed differentiates oneself from various 'reference groups' those
groups in terms of whose standards the individual judges himself and with which
he identifies with or feels kinship. (page 42)


To the extent that an attitude embodies a unique social classification or reference group,

the attitude serves a social identity function (Shavitt, 1990). For example, an attitude

regarding the American flag may symbolize the nations' values and represent a social

classification with which consumers may identify with.

The functional approach is a direct lead in to the current research because the

functional approach (e.g., identifying and measuring different motivational underpinnings

of attitude) made the point that social identity may motivate people to hold certain

attitudes. This work established why social identity matters in attitude formation. But if

consumers hold certain attitudes in order to establish or maintain a social identity, then it





8


makes sense to examine a process whereby consumers form an attitude based on social

identity.

The current research builds on the work by functional theorists by spelling out

when and how the motivational factor of social identity is likely to lead to attitude

formation, including the precise factors that are likely to lead a person to rely on social

identification as the basis for their attitude. However, in order to succinctly talk about a

process that involves "social identity," as its driving mechanism, one must first establish

a clear and concise definition. In the next chapter, social identity is briefly reviewed in

order to set forth a consistent definition of social identity that will be used throughout this

dissertation.














CHAPTER 3
THE DEFINITION OF SOCIAL IDENTITY

The Notion of Identity

The concept of "identity" was born within the psychoanalytic tradition. Neo-

Freudian psychoanalyst Erik Erikson (1950, 1956a, 1956b) investigated the struggle of

children to synthesize the continual bodily changes of youth into a meaningful and

integrated personality structure. He introduced the term "ego identity" into the literature

as "the awareness of the fact that there is a self sameness and continuity to the ego's

synthesizing methods and that these methods are effective in safeguarding the sameness

and continuity of one's meaning for others" (1946 page 23). In his later work, Erikson

(1974, 1978) goes on to write about issues related to the development of identity over the

life course.

The idea of identity seemed to be a useful concept that cut across research

disciplines. For example, in the early 1960s, two influential statements on identity were

introduced within the symbolic interaction tradition. "Appearances and the self," by

Gregory P. Stone (1990), offered one of the first concise definition of identity widely

used within sociology. Identity, according to Stone, is a meaning that a self acquires

when "situated." In other words, a person's identity is cast in the shape of a social object

by the acknowledgment of his participation or membership in social relations. In this

regard, identity is dynamic in that it is continuously changing. It is "intrinsically

associated with all the joinings and departures of social life (page 94)." This dynamic



9





10


nature of identity lead theorists to conceptualize how the search for identity could lead to

conflict. Heinz Lichtenstein (1977) published a compilation of work on identity that was

organized under the idea of identity as an existential "dilemma." The premise here is that

conflict associated with different identities emerged from a simultaneously and

irreducibly private and public reality.

Within the symbolic interaction tradition, Goffman (1959, 1961a) was perhaps

one of the first scholars to use the term "social identity" as distinguished from personal

identity and ego identity. His dramaturgical analysis begins by describing the entrance of

a stranger and the process of interpreting appearances that "enable us to anticipate his

category and attributes," i.e., his social identity. Other theorists like Berger and

Luckmann (1966, 1971), conceptualized identity as a social meaning constructed like

other meanings, but with the uniquely existential dimension of being anchored in an

individual's body (Weigert, 1983). In a more positivist framework, Burke (1980)

conceptualized identity as the subjective component of a role. In his analysis, interrelated

multiple role identities constitute an individual's self-concept.

Social Identification Processes

To socially identify implies a psychological connection with some other person or

group (Deaux 1997). This notion of identification has developed in a wide variety of

different research traditions. For example, psychoanalytic models define identification as

a process by which the individual develops ties with another person or group. Freud

(1955, 1985) stated that identification can "arise with any new perception of a common

quality shared with some other person" (page 137). Early in life, identification is

considered an unconscious process of imitating referent others who serve as models for





11


beliefs, values and behaviors. In later stages of maturation, the identification process

involves conscious choice and discrimination among possible identities (Higgins, Loeb

and Moretti, 1995). For example, Kelman's (1958, 1961, 1974b) typology of social

influence uses the term 'identification' to describe situations where an individual

willingly adopts an attitudinal position recommended by some referent other. The

motivation for adopting the position is that it establishes or maintains a positive self-

defining relationship with the referent other. Conceptually similar to French and Raven's

(1959) idea of referent power, this type of social relationship may take the form of the

target desiring to adopt the role of the referent other. It may also take the form of a

reciprocal role relationship in which the target desires to participate in activities vis-a-vis

the referent other (Eagly and Chaiken, 1993).' Identification processes are invoked only

in the context of the role relationship with the referent other that is the basis of the

identification. Therefore, once the relationship no longer becomes important to the

individual, identification ceases to exist. Foote (1951) expanded this analysis beyond

dyads when he interpreted human motivation as a consequence of identification with an

important reference group. He conceptualized identification as the "appropriation of and

commitment to a particular identity or series of identities" (page 14).

Sociological Models

Whereas psychological development theories of identification focus on members

of the family unit as identification targets, sociological models emphasize a broader set of


'An identification process (Kelman, 1958) is very different from a compliance based
process in that it is not dependent on the referent other's direct surveillance of the
individual's behavior. Neither is it contingent on the mediation of punishments
associated with not adopting the referent other's particular attitudinal position. However,
Kelman's notion of identification is a function of the salience of the self-defining
relationship with the referent other.





12


social categories that are subsets of the social system (Stryker and Statham, 1985). These

categories are large and the members are not necessarily known to one another; yet every

occupant of a particular role category shares behavior and expectations with other

occupants of the category. In this way, roles exist in the broader social system and the

outcome of conscious role selection is a defined position relative to others. These models

conceptualize identification as an interdependent process, whereby the relationships are

carried out via cooperating participants (Deaux, 1997). Symbolic interaction models

(Mead, 1934) including Stryker's (1980) identity theory and the role-identity model of

McCall and Simmons (1978) offer a similar perspective on social identification. These

models posit that social interaction operates through the enactment of roles that the

individual chooses to play (Deaux, 1997). However, these theories not only take into

account standard roles like mother and child (Stryker and Statham, 1985; Thoits and

Virshup, 1995), but also identities not so readily defined in terms of specific systems,

(e.g., mentor and student).

Social Identity Theory

In another developing research stream, social psychology has expanded its

conceptualization of the self concept to capture aspects of inter-group and intra-group

processes that are linked to important, collective social identities (Hogg, 1996; Hogg and

Abrams, 1988; Tajfel, 1978, 1982; Tajfel and Turner, 1979). Tajfel's (1959, 1969b)

work has generated empirical findings suggesting that an individual can define him or

herself in various ways. Accordingly, membership within a self-inclusive social category

provides a category-congruent self-definition that constitutes an element of the self

concept (see also Turner, 1982; Turner and Giles, 1981). In an elaboration of social





13


identity theory, Tajfel (1978, 1981; Tajfel and Turner, 1979) presumes a process of self

definition impacted by specific membership groups. The essence of self-definition is

"that part of the individual's self concept which derives from knowledge of his

membership of a social group (or groups) together with the value and emotional

significance attached to that membership" (Tajfel, 1981, page 255). Social identity

theory explicitly recognizes the importance of considering both basic motivational and

cognitive processes in order to account for inter-group perceptions and behavior. As

Tajfel and Turner (1979) put it, "social categorization entails much more than cognitive

classification of events, objects or people. It is a process impregnated by values, culture

and social representations" (page 114). Social identity theory was originally developed

to capture aspects of inter-group categorization and group dynamics associated with large

scale social identities such as races, nations and so forth.

Social Categorization Theory

Later, Turner (1985) and his colleagues developed social categorization theory to

conceptualize aspects of intra-group dynamics related to more minimal groups. These

groups are primarily characterized by frequent face to face interaction among all

members. Self-categorization essentially expands the operation of the categorization

process as the cognitive basis of intra-group behavior. This theory suggests that people

can categorize themselves and others at a number of different levels of abstraction that

can define one's social identity. For self-categorization theory, the depersonalization of

self-perception is the basic process underlying group phenomena.2 This includes various


2 The term depersonalization is not meant to arouse a negative connotation. It does not
contain any of the implications of'dehumanization' or de-individuation, but simply refers
to a contextual change in the level of identity and not a loss of identity.





14


elements of collective behavior, shared group attitudes and mutual influence processes

(Turner 1985, pages 99-100). Therefore, self categorizations are defined as cognitive

groupings of oneself and some class of stimuli as the same, in contrast to other classes of

stimuli that are different (Turner, Hogg, Oaks, Reicher and Wetherell, 1987). The

cognitive aspects of identification can be considered from two perspectives: The process

of categorization itself, and an analysis of the beliefs associated with a self designated

category (Deaux, 1997). Turner (1982, 1984) increases the weight given to cognitive

aspects of identification. He frames his basic premise as "a cognitive elaboration of the

nature of social identity as a higher order level of abstraction in the perception of self and

others" (1987 page 42). The definition of social identity by Tajfel makes it possible to

conceptualize a social identity without reference to an out-group. In contrast, however;

Turner's definition requires a we-they distinction (Deaux, 1997). In any event, social

identity and self-categorization theory emphasize the independence and in the case of

social categorization, the possibility of competition between in group and out groups in

the identification process.

Definition of Social Identity in the Current Research

It appears clear that throughout the life course, socialization within a culture

causes a person to become aware of an infinite number of social categories in the external

environment. Some of these potential bases for self-definition are more permanent (e.g.,

mother, daughter, friend, etc.) while others may be more transitory (e.g., Republican,

athlete, graduate student etc.) This research starts with the idea that consumers may

perceive themselves in terms of these various levels of abstraction (Tajfel, 1959; Tajfel

and Turner, 1979; Turner and Oakes, 1986) and at any given point in time will have





15


available a subset of social categories that can become a part of their "working" or

"spontaneous" self-concept (Markus and Kunda, 1986; McGuire, McGuire and Winton,

1979). These social categories are "cognitive structures" whose meaning comes about

over time through socialization and many of the other processes described in this chapter.

A consumer may adopt a social category as one of her social identities in order to

think about various actions or judgments. For clarity of exposition, in this research, the

term "social category" refers to the infinite number of potential social constructions that

may come from culture, society, marketers, peer groups, etc. The term "social identity"

refers to the actuated perspective or frame of reference that a consumer possesses as part

of the repertoire of who they are or want to appear to be.3 The next chapter describes a

process of how an attitude may come to be formed on the basis of "social identity."




















3 The current research does not seek to determine the different social identities available
to a person. Instead, it takes the view that for any social identity, a specified set of factors
determines whether it is likely to generate an attitude. The emphasis is on those factors.
A social identity needs a social referent, but that could be a group, an abstracted ideal, an
individual--or any social construction. For instance, for a child, a comic book hero may
well be a social identity. As long as the child attempts to see the world through the eyes
of that entity, he has adopted that social identity to make some type of assessment.














CHAPTER 4
SOCIAL IDENTITY BASED ATTITUDE FORMATION

Consider the social identity of "athlete." What does it mean to "be" an athlete?

Thinking about this social identity might trigger a mental representation or distinct

picture of what an athlete thinks (attitudes) and does (behaviors). A consumer might

adopt this social identity and use its associated evaluative content as the basis to form an

attitude (e.g., attitude toward Nike sports shoes or attitudes toward exercise). This is

likely to result in a collectively anchored attitude that is formed via identification

processes (Kelman, 1958) and is held, expressed or used as a guide for behavior in order

to establish, maintain or even communicate a social identity to others (Shavitt, 1990).

This chapter describes this process and the key factors that increase the likelihood

of its occurrence. A consumer's social identity is more likely to affect her attitude when

a social identity is 1) salient, 2) self-important 3) relevant to the object and 4) provides a

basis to evaluatively respond. The framework is graphically pictured in figure 1. To

motivate the forthcoming empirical work, each of the moderating factors is now briefly

described.



Is the Social Identity Salient?

Social categories are internal mental representations that can become a basic part

of how consumers view themselves. For example, a consumer may see herself as a

"democrat" a "professor," "tomboy" or "working mother." However, no matter how



16





17


extensive is knowledge about them, social identities can have little impact on consumer

attitudes and behaviors unless social identity information is accessed:

Postulate 1: If a social identity is salient, then there is an increased likelihood that the
evaluative content linked to that social identity will impact the formation
of a consumer attitude.

This model is using the term salience to refer simply to the extent that a social identity is

an "activated" conceptual structure in the consumer's working self-concept. If a social

identity is salient, it can bring to mind attitudes and behaviors consistent with the social

identity (e.g., the social identity "Democrat" might be linked to perceptions of

favorableness toward government sponsored social programs).

Consider an African American consumer in a grocery store full of Caucasian

consumers. His ethnic identity might be salient. If the same African American consumer

were in a grocery store full of African American women, his gender (male) identity might

be more salient (see McGuire et al., 1979). The salience of his social identity might

trigger attitudes useful for consumer decision-making. For example, if standing in front

of the grocery store magazine rack, the African American consumer might be more likely

to peruse Ebony Magazine in the first case and GQ in the second. In these cases, the

consumer perceives a stronger link between the social identity (e.g., ethnic identity vs.

gender identity) and the object (Ebony vs. GQ). According to the framework, salience of

a social identity moderates the relationship between evaluative content linked to a social

identity and attitudes generated on the basis of it (see Figure 1).

Is the Social Identity Self-Important?

Not all social identities are alike. Consider the following example. Two

consumers Alan and Rich both consider themselves "athletes." Alan is a former high





18


school track star who has moved on to other things. He is now a weekend-warrior-type,

playing the occasional game of tennis to stay in shape. In general, he tries to eat healthy.

Rich on the other hand is also a former high school track star, but is a former Olympic

silver medal winner in the 100 meters. He is a fierce competitor who anxiously awaits

the opportunity to reclaim his prior days of glory.

Both of these consumers probably possess "athlete" as part of their sense of who

they are. But because of past experience and future aspirations, Rich's identity as an

athlete might carry more personal consumer value to him. The fact that this identity is

much more engulfing to Rich may lead to a higher likelihood that many of Rich's

attitudes will be based on the "athlete" aspect of his social identity. Additionally, holding

all else constant, Rich is probably more likely to be more favorable toward and object

(e.g., Nike brand shoes vs. Keds) that is linked to his athlete identity. Therefore, even if a

social identity is salient, it might not be the basis by which a consumer attitude is formed.

The consumer must identify with that social identity (i.e. the social identity must be self-

important):

Postulate 2: As a social identity becomes more self-important, then there is an
increased likelihood that the evaluative content linked to that social
identity will impact the formation of a consumer attitude.

A consumer may be drawn to a social identity for various motivations. At one

extreme, the self-importance of a social identity may manifest as an impression motivated

and temporary public standpoint (Schlenker, 1985). At another extreme, the social

identity may serve as a "phenomenological lens" that deeply engulfs the individual as a

powerful basis for self-definition (as it does for Rich in the example above).





19




Is the Social Identity Object-Relevant?

Even if a social identity is salient and self-important to the consumer, it might not

be a basis for attitude formation. The social identity must be relevant to the object that is

to be evaluated. The object-relevance of a social identity increases the likelihood that the

attitude object will be thought of in terms of the social identity:

Postulate 3: As a social identity becomes more object relevant, then there is an
increased likelihood that the evaluative content linked to that social
identity will impact the formation of a consumer attitude.

For example, a consumer who perceives herself as a "working mother" may be more

favorable to an automobile that emphasizes safety and practicality. As another example,

an attitude toward affirmative action may be more relevant to a consumer's ethnic social

identity if the person perceives affirmative action as a means of achieving an important

identity based goal such as insuring equitable employment (Kravitz, 1995). The

perception of relevance may be momentarily heightened if the person is concerned about

his or her immediate, upcoming opportunities on the job market. Therefore, active goals

and current concerns may impact the relevance of the attitude object to the social identity

and thereby motivate the consumer to consider the attitude object in relationship to some

social identity.

Does the Social Identity Provide a Basis to Respond?

Even if a social identity is salient, self-important and object relevant, it might not

be the basis by which a consumer forms an attitude. This model uses the term

"evaluatively diagnostic" to refer to the extent to which the evaluative content of the

social identity has sufficient clarity and specificity to inform the consumer's evaluation of





20


the object and guide a behavioral response. In other words, does the social identity point

a person in a meaningful direction?

Postulate 4: If a social identity is evaluatively diagnostic, there is an increased
likelihood that the evaluative content linked to that social identity will
impact the formation of a consumer attitude.

For example, suppose a consumer is evaluating a consideration set of shoe brands.

Hence, an attitude is called for. Further suppose that the person's social identity (e.g.,

"Urban Youth") is highly salient at the time an attitude is generated (e.g., the person is

watching a program on Black Entertainment Television). Even if the consumer is

evaluating a set of brands (e.g., sketcher's, FUBU) that is clearly relevant to her self-

important social identity, she may not be clear on which brand best embodies her "Urban

Youth" identity. In the current terminology, her social identity in this case would be

evaluatively non-diagnostic relative to evaluating the brands of shoes (c.f., Feldman and

Lynch, 1988). Her social identity as an urban youth and the fact that there is no clear

identity related norm (Kallgren, Reno and Cialdini, 2000) provides her with an

inadequate basis to respond to the object (i.e., in this example, choose among the two

brands).

Summary of the Framework

The framework of this dissertation research suggests that a consumer can think

about herself in terms of various social identities that can connect her to a brand, product

or behavior. These social identities may lead to the formation of an attitude, but only

under specific conditions. There are four key moderating factors that determine if social

identity, (defined as the evaluative content of a social identity) will impact consumer

attitude formation. If when an attitude is called for, a social identity is salient (the





21


consumer is thinking about herself as being that kind of person), self-important (the

person strongly identifies with that social identity), relevant to the object to be evaluated

(is functionally linked to the object), and provides a basis to respond (gives meaningful

evaluative direction for the consumer), then there is an increased likelihood that an

attitude will be generated on the basis of social identification processes.

Intended contribution of the framework. The framework integrates past research

traditions in social identity theory and functional attitude theory to develop a theory of

social identity based attitude formation. For example, recent theorizing in social

psychology assumes that much of the working self-concept's content is socially

embedded. Specific social self-categorizations are brought into play in the social field

(Turner, Hogg, Oakes, Reicher and Wetherell, 1987). These social identities become the

basis of perception conduct and behavior (e.g., Giles and Johnson, 1987; Hinkle and

Brown, 1990; Hogg, 1992). Therefore, this framework argues that the salience of a social

identity is important in bringing to mind evaluative content (i.e., attitudes and behaviors)

that is linked to that social identity. However, the salience of a social identity is a

necessary but not sufficient condition for social identity based attitude formation. A

consumer must also be drawn to a social identity as a basis for self-definition. The more

a consumer identifies with a social identity, the more likely that social identity will

impact her attitudes. However, a salient and self-important social identity will not lead to

an attitude unless it is linked to the object that is being evaluated. The object must be

"relevant" to the social identity in that an attitude generated on the basis of the social

identity serves a psychological function of self-expressiveness (Katz, 1960; Smith,

Bruner and White, 1956) or what has been more recently called a "social identity"





22


function (Shavitt, 1990). Consequently, the object relevance of the social identity

moderates the impact of the social identity's evaluative content on forming an attitude

toward the object. Lastly, the social identity must provide a clear basis to respond to the

object. Even if a salient, self-important and object relevant identity is not diagnostic

(Feldman and Lynch, 1988) to a judgment, it is unlikely that social identity will be the

basis for attitude formation. Each of the above theoretical assertions have been discussed

in past literatures in different theoretical domains. But taken together, the current model

attempts to make a contribution by integrating these conceptual underpinnings in a way

that describes a process of how and when social identity is likely to impact consumer

attitude formation.

Importance of the framework. Past research has often defined the process of

attitude formation as specific belief elements and their evaluative implications. The

result has been a focus on features / aspects of the attitude object and how these attributes

should be measured and represented in some algebraic form. Consumers might form

attitudes in this way. However, there might be situations where the driving mechanism

behind an attitude has less to do with product features or benefits and more to do with

social identity. If being of some social identity is salient, and a consumer is thinking

about herself in relation to that particular social identity, then it is possible for that frame

of reference to dominate how she interprets the world around her. If one assumes that it

is important to distinguish the different bases by which a consumer forms an attitude in

order to understand why consumers react to a product offering in a certain way or to

identify the most effective ways to modify a consumer's response, then conceptualizing a

social identity based process of attitude formation and defining the contributing factors






23


that impact the process should lead to important implications for attitude theory,

marketing practice and public policy.







24





Consumer
Needs an
attitude


,--- -- -- -- -- -- ----
FACTORS THAT AFFECT
EVALUATIVE ATTITUDE:
IMPLICATIONS
OF A SOCIAL 1) Direct Experience /
IDENTITY Imagined Experience
2) Similarity to Other Objects
3) Past Behavior
4) Object Features
5) Personal Values
SALIENCE-







SELF-
IMPORTANCE

MODERATING
FACTORS


OBJECT
RELEVANCE







DIRECTION -




ATTITUDE







Figure 1.
Conceptual framework of social identity based attitude formation.














CHAPTER 5
EXPERIMENTS 1A AND 1B: THE INTERACTIVE ROLE OF
SALIENCE AND OBJECT RELEVANCE

Overview of the Studies

This chapter describes the results from a pretest that assesses the effectiveness of

a salience induction manipulation. The salience manipulation was used in two

experiments--also described in this chapter--that carefully examine attitude formation

when a social identity is both salient and relevant to an attitude object.

Theoretical Premise of the Studies

Consumers are exposed daily to various marketing stimuli and consumer

situations. Any of these can heighten a particular basis for social identification. For

example, a female consumer in a store full of males might be likely to have her "gender"

identity salient (c.f., McGuire, McGuire and Winton 1979; McGuire, McGuire, Child and

Fujioka 1978). Or a commercial that shows a concentration of ethnically homogenous

individuals may make ethnic identity salient to consumers who view that ad and share

that identity (Forehand and Deshpande, in press). Marketers can in a sense, alter how a

consumer momentarily thinks about herself by making a particular social identity salient.

In experiment la, the following hypothesis will be tested:

HI(a): Heightening the salience of an identity increases the likelihood that a
consumer will see herself in terms of that identity.

However, according to the framework, the salience of a social identity alone is not

enough to ensure that an attitude will be based on the evaluative content of a social



25





26


identity. If the social identity is salient and relevant to the brand or product, then social

identification is more likely to connect the consumer to the brand or product. This

increases the likelihood of social identity leading to the formation of a consumer attitude.

In experiment lb, the following hypothesis will be tested:


Hl(b): Heightening the salience of an identity leads to higher (lower)
evaluations of an identity relevant (identity irrelevant) product when
adopting such an identity would provide a favorable basis to respond to
the product.

When a social identity is made salient, it is at least temporarily a part of a person's

"working" or "spontaneous" self-concept (Markus and Kunda 1986; McGuire and

Padawer-Singer 1978). This heightened salience combined with object relevance (i.e.,

when the particular social identity is relevant to the stimulus to be evaluated) should

increase the likelihood that the object will be thought of in terms of the particular social

identity. This should increase the likelihood of attitude generation based on social

identity.



Pretest

Purpose and overview of pretest. The subsequently described pretest examines

the effectiveness of the salience manipulation used in experiments la and Ib. In the

pretest, one of two different social identities was made salient between subjects. Subjects

had these social identities made salient in one of two different ways. Subjects then

participated in an ostensibly unrelated task intended to assess the effectiveness of the

salience induction technique. More details follow.





27


Procedure. In the first task, subjects participated in a "visual imagery study" that

involved evaluating "Internet typeface formats" for use in electronic communications

(e.g., Web page designs). Subjects evaluated different multiple typeface formats on

several dimensions (e.g., warm, memorable, fun and common) in terms of how a

"random" printed word came across to them. Unbeknownst to the subjects, the task made

either their friend (family) social identity salient by requiring the subjects to evaluate the

word "friend" ("son" / "daughter") seven different times in seven different typefaces'.

This is a pure salience manipulation intended to prime the particular social identity of

interest.

Additionally, two different types of salience were induced between subjects.

These two different types of salience have no substantive significance in the theoretical

framework, rather, these two different types of salience were tested to assess possible

differences in the extent to which a particular depth of processing was needed to invoke

the salience of the social identity in question. Therefore, in the concept salience

condition, subjects evaluated the focal word (friend, son or daughter) in no particular

evaluative context. In other words, the preamble to the focal word was evaluatively

neutral (e.g., think about the word "friend."). In the idea salience condition, the

evaluative context was manipulated in the lead in (e.g., think about the importance of

being a "friend.") priming not only the social identity, but also a valenced idea linked to

that social identity. A portion of an example idea salience manipulation is shown in


This pretest involves two social identities that have been established in prior work.
They are the "friend" and "family" social identities that are associated with peer versus
familial role contexts (cf. Baldwin and Holmes, 1987; Solomon, 1983). Because males
and females participated in the study, the family social identity was further broken down
into the precise social identities of "son" and "daughter" depending on the gender of the
subject.





28


figure 22. Subjects rated the type faces and to bolster the cover story, answered several

questions regarding their experience with Internet communications (e.g., time spent

surfing web sites, etc.)

Immediately afterwards, subjects participated in an ostensibly unrelated exercise

whose cover story concerned assessing familiarity with "word search games". Subjects

were instructed to peruse a word search grid on a subsequent page that contained eighteen

words (including the focal word "friend", "son" and "daughter") amidst randomly

arrayed letters in a 20 x 20 word search grid. Subjects were given approximately ten

minutes to find as many words as possible.3 Subjects were then debriefed and dismissed.

Predictions. This pretest was designed to assess the extent to which the salience

manipulation in the visual imagery study was able to heighten the salience of a particular

social identity. In a figure ground sense, it was expected that heightening the salience

of either the "friend," or "family" ("son" / "daughter") social identity would increase the

likelihood that subjects would be able to locate the focal word (friend, son or daughter)

associated with that salient social identity. This might be particularly the case if the

salience induction contained an evaluatively valenced context (idea salience conditions)

because of the additional depth of processing presumably evoked by the idea salience

conditions.




2 The full stimulus materials for these salience manipulations are included in Appendix A
at the end of this dissertation.
3 Subjects were further told that every time they found a word, to 1) circle the word in the
grid and then to 2) write down the word in one of eighteen spaces provided. Subjects
were told to stop after ten minutes. The instructions encouraged subjects not to cheat.
Also, the experimenter verbally indicated that the exercise was intended to be fun and
was not a test of their abilities in any way. Subjects were also told that the data would be
useful only if instructions were carefully followed.






29


Experimental design. Subjects were assigned to a 3 (focal word that was

evaluated: friend, son, daughter) x 2 (type of salience: concept vs. idea salience) between

subjects factorial.4 The dependent measure of the pretest was the total number of

instances in which the word "friend," "son" or "daughter" was identified (circled) in the

word search grid.

Results. One hundred and ten (N=110) subjects from a southeastern university

participated for class credit and were assigned to experimental treatments. For each

social identity made salient, the percentage was calculated of instances in which subjects

identified any of the focal words "friend," "son" or "daughter." These results (collapsed

across type of salience induced conditions) are presented in figure 3.

A logistic regression was performed defining the dependent variable as whether

or not a subject identified one of the focal words. The type of salience (concept vs. idea)

and which social identity was salient (friend, vs. family (son / daughter)) were the

explanatory variables in a model including main effects and all possible higher order

interactions.

An analysis examining whether or not subjects identified the word "friend" in the

word search grid produced the following results. There was a main effect of type of

salience (X2=4.06, p=0.0439), indicating that the percentage in which subjects picked out

the focal word "friend" depended on which type of salience was induced. Collapsing

across which social identity was salient showed that 22% vs. 39% of the subjects found



4 Subjects in the "son" conditions were all male, subjects in the "daughter" conditions
were all female. However, the son vs. daughter distinction in this analysis has no
substantive significance as the theory has no a priori predictions as to differences within
the "family" social identity conditions. However, the results will be described in terms of
all three different social identities.





30


the focal word "friend" in the concept and idea salience conditions, respectively. When

subjects in the first task examined any of the three focal words in an evaluatively

valenced context, they identified the focal word "friend" more often in the subsequent

word search grid. All other effects were not significant.

An analysis examined whether or not subjects identified the word "son" in the

word search grid. A main effect of which focal word was evaluated was found (X2=6.62,

p=0.0366). The percentage in which subjects found the focal word "son" in the word

search grid depended on which social identity was salient (focal word was evaluated). In

a follow up analysis, a planned contrast showed that more subjects found the focal word

"son" when son was salient (57%) as compared to when "friend" was salient (29%) or

when "daughter" was salient (48%), (X=3.49, p=0.0616). No other effects were

significant.

Finally, an analysis examined whether or not subjects identified the word

"daughter" in the word search grid. A marginal main effect of type of which focal word

was evaluated was found (X2=5.47, p=0.0651). The percentage in which subjects found

the focal word "daughter" in the word search grid depended on which social identity was

salient (focal word was evaluated). A planned contrast showed that more subjects found

the focal word "daughter" when daughter was salient (39%) as compared to both when

"friend" was salient (19%) and when "son" was salient (17%), (X2=5.49, p=0.0194).

Again, no other effects were significant.

Discussion. The results of this pretest are consistent with the idea that

heightening the salience of a social identity (having subjects rehearse the focal word)

made the focal word stand out. In other words, the salience manipulations increased the





31


likelihood that subjects would locate the focal word arrayed in the subsequent word

search grid. These effects were found for the focal words associated with the "family"

social identity (i.e., "sons" and "daughter") but not for the "friend" social identity.5

Heightening the salience of "friend" had no significant effect on subjects finding the

focal word "friend."

Given the subtlety of the salience manipulation and that subjects were college

students participating in a task located in their school environment, it might have been

more difficult to heighten the salience of their "friend" social identity because it may

already be salient, minimizing the impact of the "friend" salience manipulation on

subjects ability to identify "friend" in the word search grid. In contrast, in the

aforementioned experimental situation, it might be easier to heighten subjects' familial

("son" and "daughter") social identity because subjects were certainly in a context in

which that identity might not likely be already activated (c.f. Bem, 1980).

It also appears that the concept and idea salience conditions had no impact on the

likelihood of subjects identifying the focal words "son" and "daughter." However, the

concept and idea salience inductions produced differential effects on the extent to which

the focal word "friend" was found. When salience of "friend," "son" or "daughter"

included an evaluatively valenced context, subjects were more likely to spot the word

friend. Although purely speculative, this could be the result of the fact that when subjects

rehearsed the word "son" and "daughter" in an evaluatively valenced context, (e.g., think



5 This pretest is based on the assumption that the more salient a focal word, the more
likely it will be identified in the word search grid. Of course, there are other stimulus
factors (e.g., word usage frequency, number of letters in the focal word, where the word
is arrayed in the grid, etc.) that may increase or decrease subjects ability to identify a
focal word. Hence, the analysis presented contrasts across a particular focal word.





32


about the importance of being a daughter) they were more likely to bring to mind

associations that also tapped the "friend" identity. For example, consider a person who is

so close to her/his parents that s/he sees them as friends rather than in the role of parent

vs. daughter/son. Nonetheless, the overall results of this pretest do suggest that the visual

task manipulations were particularly effective at heightening the salience of the "family"

(son and daughter) social identity.



Experiment la

Purpose and overview of Experiment la. Experiment la tests hypothesis la by

investigating the impact of salient social identities on altering consumers' momentary

sense of self-definition. Exactly as in the pretest, one of two different social identities

(friend vs. family) was again made salient between subjects. As in the pretest, subjects

had these social identities made salient in one of two different ways. Subjects then

participated in an ostensibly unrelated task intended to assess the extent to which the

salience induction manipulations shifted subjects' spontaneous self-concepts. More

details follow.

Procedure. In the first task, subjects participated in the same visual imagery study

described in the pretest. Unbeknownst to the subjects, the task made either their friend

(family) social identity salient by requiring the subjects to evaluate the word "friend"

("son" / "daughter") seven different times in the seven different typefaces. The two

different types of salience used in the pretest were also used in experiment la6


6These two manipulations are included to again asses the level of processing associated with
making a particular social identity salient, and to shed possible light on the results from the
pretest that suggested differential effects of the concept vs. idea salience manipulations on
locating the focal word "friend."





33


Immediately after completing the visual imagery study, subjects completed an ostensibly

unrelated consumer "self-description study" (c.f. Mcguire et al., 1979) where subjects

were asked to "please take a minute to describe yourself and to provide as much detail as

you wish." This open ended probe was the only instruction given. Subjects completed

the one page questionnaire and were debriefed and dismissed. This questionnaire is

included in Appendix C.

Predictions. Experiment la predicts that when a particular social identity (friend

vs. family) is made salient, then the person is more likely to "see themselves" in terms of

that social identity. These differences should be evidenced in the content of a person's

spontaneous self-descriptions. It is expected that when subjects have their "friend" social

identity made salient (by rehearsing and evaluating the word "friend" in the first task),

they will be more likely to spontaneously describe themselves in terms of their "friend"

social identity. Conversely, it was also expected that when subjects have their "family"

social identity made salient (by rehearsing and evaluating either the focal word "son" or

"daughter"), they would be more likely to spontaneously describe themselves in terms of

their "family" social identity. The effects described above might be particularly likely if

subjects rehearsed the focal word in an evaluatively valenced context (idea salience

conditions).

Experimental design. Subjects in experiment la were assigned to a 3 (focal word

that was evaluated: friend, son daughter) x 2 (type of salience: concept vs. idea salience)

between subjects factorial. The dependent measure was the number of instances in which

subjects did or did not explicitly mentioned some aspect of their relationship/connection

to either their "family" or "friends" in their spontaneous self-descriptions.





34


Results. One hundred and thirty five (N=135) subjects (different from the

pretest) from a southeastern university were randomly assigned to experimental

treatments. For each social identity made salient, the percentage was calculated of

instances in which subjects did or did not mentioned a connection to either their "family"

or "friends" in their self-descriptions. The results (collapsed across type of salience

induced conditions) are presented in figure 4. A logistic regression was performed

defining the dependent variable as whether or not a subject explicitly mentioned some

aspect of their "friends" or "family" in how they described themselves. The type of

salience (concept vs. idea) and which social identity was salient (friend, vs. son vs.

daughter) were the explanatory variables in a model including main effects and all

possible higher order interactions.

An analysis examined whether or not subjects explicitly mentioned some aspect

of their "friend" identity in their spontaneous self-descriptions. In an omnibus analysis,

there were no significant effects for the type of salience induced, which social identity

was salient or any corresponding interaction effects (X2I =n.s). Null effects were also

found in a planned contrast that compared the extent to which subjects mentioned some

aspect of their "friend" identity in the friend made salient condition (33%) vs. the extent

to which subjects mentioned some aspect of their "friend" identity both in the "son" and

"daughter" made salient conditions (22% and 28% respectively, X=n.s). A second

analysis examined whether or not subjects explicitly mentioned some aspect of their

"family" identity in their spontaneous self-descriptions. A main effect of which social

identity was salient was found (X2=13.83 p=0.0010). In a follow up analysis, a planned

contrast showed that more subjects spontaneously mentioned some aspect of their





35


"family" identity when either "son" was salient (46%) or "daughter" was salient (53%) as

compared to when "friend" was salient (16%, X2=13.45, p=0.0002).

Discussion. Consistent with hypothesis 2a, subjects who had their family identity

salient by induction of either the "son" or "daughter" social identity salience

manipulation were more likely to mention some aspect of their "family" in their

subsequent spontaneous self-descriptions. The interpretation is that salience of the social

identity made subjects more likely to see themselves in terms of that social identity

indicating that the social identity was at least a temporary part of their working self-

concepts. This finding is consistent with the pattern of results obtained in the pretest that

showed that the salience inductions also altered the extent to which subjects were able to

identify the focal words "son" and "daughter" out of randomly arrayed letters. However,

the attempt to affect the dependent measures by making "friend" salient was again

unsuccessful in experiment la, suggesting the difficulty (at least with these manipulations

for this sample) of pushing subjects' self-description around on the friend dimension.



Experiment lb

Purpose and overview of Experiment Ib. The evidence in experiment la suggests

that the salience manipulation for the "family" social identity was effective at altering

subjects' temporary self-conceptions. Experiment lb will examine the extent to which

this momentary state of self-conception mediates attitude formation when an attitude is

generated toward an identity relevant object. More specifically, experiment lb assesses

the necessity of both salience and object relevance (Hlb) in forming a favorable

consumer attitudes (based on social identity when adopting such an identity would in fact





36


provide a favorable basis to respond to the product). In experiment Ib, subjects' "family"

social identity was made salient. In an ostensibly unrelated study, subjects then evaluated

a product that was either relevant or irrelevant to their family social identity.

Procedure. Subjects first participated in the same visual imagery study used in the

pretest and experiment la. In this experiment, however, only the "family" social identity

was made salient (between subjects) by having subjects evaluate the word "son" or

"daughter" seven different times in the seven different typefaces. Additionally, only the

concept salience manipulation was used in experiment lb.7

After subjects completed the first task, a different experimenter administered an

ostensibly unrelated "new product assessment study." Subjects were asked to evaluate

three different "new product concepts." The first product was the focal stimulus. It was

a palm held interpersonal telecommunication product. The second and third products

were filler products. One of the filler products was an electronic portable head-band

worn in the summer to keep cool and the other filler product was a flat tire re-filler for

automobiles.

For each of the three products, subjects were presented with a picture of the

product and a product description. The focal product (interpersonal telecommunication

product) was manipulated to be either relevant or not relevant to the social identity

primed in the first task by framing the interpersonal telecommunication product (e.g., as

in a product positioning strategy) in terms of the "family" social identity or not. For


7 The "friend" salience manipulation was not used in experiment Ib because of the data
that suggested that the "friend" salience manipulation was not strong enough to heighten
the salience of the friend focal word (pretest) or to alter subjects' working self concepts
(experiment la). Similar logic led to usage of only the concept salience manipulation in
experiment Ib because the supporting data detected no meaningful differences across
concept and idea salience conditions in the pretest and experiment la.






37


example, in the object relevant to the "family" (sons and daughters) social identity

condition, the focal product description included utilitarian features (e.g., calendar,

address book, email) but with an emphasis on the product as a basis for staying connected

to parents and family:

ELECTRONIC COMPANION: Model (SN) DTR-3

Calendar, address book, to-do list, expenses, e-mail. More importantly, exchange information
and well wishes with your loved ones at the touch of a button, keeping you connected with your
parents and loved ones.

This new product allows you to never miss important family events and/ or occasions that are
dear to your heart. In fact, it will automatically remind you of them, and if you like,
automatically send messages, flowers and e-cards. Custom messages can be programmed ahead
of time and sent automatically.

Enjoy up to a month's operation on the road. Compatible with Windows 3.1, 95, NT 4.0
(optional adapter required for Mac). Modem operates on 2 AAA batteries (included). Measures
4.5" x 3.1" x .4".

No such emphasis was made in the object irrelevant condition, where the focal product

description stressed only the presence of the utilitarian features:

ELECTRONIC COMPANION: Model BUS-3

Calendar, address book, to-do work list, expenses, business e-mail. Exchange synchronized data
with your PC at the touch of a button, eliminating dual data-entry and ensuring the safety and
accuracy of information.

This new product allows you to beam data to other Palm platform devices using the infrared port.
In fact, its unique user-expandable design lets you add a modem, memory, software and more.
Custom messages can be programmed ahead of time and sent automatically.

Enjoy up to a month's operation on the road. Compatible with Windows 3.1, 95, NT 4.0
(optional adapter required for Mac). Modem operates on 2 AAA batteries (included). Measures
4.5" x 3.1" x.4".


Dependent measures. The dependent measures were gathered as part of the

second task. After reading the product descriptions, subjects indicated the likelihood that

they would purchase each of the three products, ranging from (1) unlikely to (7) highly





38


likely. They also indicated the likely market success of the products ranging from (1)

unlikely to (7) highly likely. Subjects also wrote down in open-ended responses, what

they thought were the best reason(s) for buying the product. They were subsequently

probed in an additional open-ended question to write down any more reasons that they

could think of. The complete new product assessment study is included in Appendix D

of this dissertation.

Predictions. For experiment lb, the framework predicts that more favorable

attitudes should be observed in the following condition: when a social identity is salient

and the consumer evaluates an identity relevant product. If the "family" social identity is

made salient (as was evidenced in the pretest) and consumers are thinking about

themselves in terms of their "family" social identity (as was evidenced in experiment la),

then they should like the focal product that is framed in terms of their family identity

more than the focal product that is not framed in terms of their salient family identity.

Salient social identification connects a consumer to the identity relevant product, leading

to more favorable attitudes as evidenced in higher purchase likelihood for the identity

relevant focal product. No such interaction of design factors should emerge for purchase

likelihood of the filler products because the filler products (held constant across

conditions) should have no apparent linkage to the family social identity.

Experimental design Subjects in experiment two were assigned to the cells of a 2

(Two focal words evaluated: Son vs. Daughter8) X 2 (Identity relevance: Relevant -

framed in terms of family or Not relevant framed in terms of features) between subjects




8 As in the pretest and experiment la, subjects in the "son" condition were all male, subjects in
the "daughter" condition were all female.






39


factorial design with two control groups.9 The dependent measure was purchase

likelihood for the focal product and purchase likelihood averaged across the two filler

products.

Results. Sixty-two (N=62) subjects from a southeastern university were assigned

to experimental treatments. An analysis of variance tested the main and interaction

effects of salient social identity (sons vs. daughters) and object relevance (relevant:

framed in terms of features + family or irrelevant: framed in terms of just features) on

mean purchase likelihood for the focal product and the mean purchase likelihood

averaged across the two filler products. The omnibus ANOVA of effects of identity and

relevance on purchase likelihood of the focal product and filler products are reported in

Appendices H and I respectively. The means and standard deviations are presented in

Table 1. Consistent with predictions, an analysis on the average purchase likelihood of

the two filler products revealed no significant effects (F < 1). Additionally, there was no

significant social identity (son vs. daughter) by relevance interaction on purchase

likelihood for the focal product indicating no differential effect of gender (son vs.

daughter) within the "family" social identity. In other words, male subjects who had

"son" identity made salient did not respond differently than female subjects who had

"daughter" identity made salient. More importantly, for the focal product, the analysis

yielded a significant main effect of relevance, (F(1,58)=26.67, p.< .0001). Subject's

purchase likelihood for the focal product depended on whether the product was relevant

9 In the absence of any salience manipulation, sixty (N=60) separate males and females
participated in the new product assessment task that included between subjects, the focal product
framed in terms of family plus utilitarian features (identity relevant conditions) or utilitarian
features only (identity irrelevant conditions). No interaction of gender with identity relevance
was found (F < 1) so mean purchase likelihood of the focal product and the two filler products in
the object relevant and object irrelevant conditions was calculated included in the main analysis
as controls.





40


or irrelevant to the social identity that was made salient. Consistent with predictions,

identity relevance resulted in higher purchase likelihood of the focal product when the

family social identity was salient (3.35 vs. 5.31 and 3.20 vs. 5.12) in the "son" and

"daughter" made salient conditions, respectively. To further examine these results, a

follow up analysis was run that compared treatment means to their respective control

groups using Dunnett's (1955) procedure (as reported in Keppel, 1991) controlling for

family wide type one errors. Consistent with predictions, the only treatment means that

differed from their control group were when "son" was salient in the object relevant

condition (Mson/objectrelevant MCONTROL = 1.92, critical difference = 1.63, p < .01 FWa =

.05), and when "daughter" was salient in the object relevant conditions

(Mdaughter/objectrelevant,- MCONTROL = 1.73, critical difference = 1.63, p < .05 FWa = .05).

Discussion and limitation(s) of the experiments. The results of experiment la and

experiment lb provide support that social identification can mediate attitude formation in

ways that are consistent with the framework presented in this research. Experiment la

established that when a social identity was made salient, the likelihood increased that a

person would temporarily "see" themselves in terms of that particular social identity. In

other words, heightening the salience of either "son" or "daughter" at least temporarily

altered subject's self-conceptions such that the social identity was an activated identity in

their working self-concepts (as compared to when "friend" was salient). Experiment lb

showed that differences in the salience of a particular social identity led to predictable

differences in purchase likelihood (assumed to be guided by attitude formation). When a

social identity was salient, subjects more favorably responded to an object relevant focal

stimulus (product framed in terms of that social identity) as compared to when no such





41


social identity was made salient (relevant control group comparison). This critical

difference in attitude generation was found for the focal product, but not for the two filler

products predicted to be unrelated to heightening the salience of the social identities in

question. The combination of these two experiments suggest that salience of a particular

social identity can lead to differences in how a consumer views herself. These

differences in a person's self-conception appeared to mediate attitude generation but only

when the attitude object was perceived as being relevant to the salient social identity.

Some limitations of these studies are of note. In experiment lb, no manipulation

check on the relevance of the focal product to the social identity were taken. To some

extent, this analysis assumes that subjects' perceptions of object relevance was the

critical moderating factor in purchase likelihood assessments. Second, no analysis of

subjects' self-descriptions was taken in the absence of a salience manipulation for the

social identity "son" or "daughter." Therefore, it might be some what difficult to assume

that the effects are truly being mediated by appreciable differences in how the salience

induction technique was able to alter people's momentary spontaneous self-concepts.

However, the self-description procedure is a demanding dependent variable since people

are free to bring everything in their backgrounds to bear, and there is much less

opportunity for the experimental context to affect the outcome. Given the fact that the

predicted differences were found potentially suggests that the assessment of people's

working self-concepts may have actually been underestimated given the open ended

nature of the response task. These two limitations notwithstanding, the amalgamation of

evidence across the pretest and experiments la and lb do provide consistent support for

the prediction that a consumer's shifting social identification can affect their evaluative






42


response if the social identity is brought to mind and the object that they are responding

to is linked to that identity.






43





We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a
reasonably careful manner. In order to get you to thoughtfully evaluate each typeface, we used
a lead in presented in a neutral font and we varied it slightly to slow you down just a little bit.

Now, please evaluate the meaning of the word "friend" in terms of "how it comes across to you"
in the particular type-face presented. To provide us with the most accurate immediate
response, read the phrase each time and try to rehearse it in your mind before providing your
evaluations:

Evaluate the Type face on the
following dimensions provided in the
box

THINK ABOUTTHE FRIEND Warm Q0 0 0 0 0 Impersonal
IMPORTANCE OF Memorable L0 0 0 0 0L Bland
BEING A Fun 0 0 0 [ L Boring
Common L0 0 L0 L0 Unique

CONSIDER THE CF]iECx( Warm L0 LI L0 L Impersonal
VALUE OF BEING A Memorable L l LI l L Bland
Fun L 0 0 0 0 Boring
Common L0 0 0 0 0LI Unique

PONDER THE FRIEND Warm L0 0 0 0 0L Impersonal
SIGNIFICANCE OF Memorable L0 0 0 0 0L Bland
BEING A Fun L0 0 0 0 Boring
Common L0 L0 L0 L Unique

Figure 2.
Pretest: example of the idea salience condition.






44





0.9

0.8

0.7

S0.6

c 0.5
0.o Friend
a 0.4 S
0. Son

0.2

0.1


Friend Family
Salient Social Identity


Figure 3.
Pretest: percentage of response category frequency as a function of which social
category was salient. Family identity was made salient by having male subjects
rehearse the word "son" and having female subjects rehearse the word "daughter."
(N=110)






45





0.9

0.8

0.7

S0.6
S[ Frinend
C 0.5
U* *Family
( 0.4

0.3

0.2

0.1


Friend Family
Social Identity Made Salient



Figure 4.
Experiment la: self descriptions that referenced relationship with social identity as a
function of which social identity was salient. Family identity was made salient by
having male subjects rehearse the word "son" and having female subjects rehearse the
word "daughter."
(N=135)






46




Table 1.
Experiment lb: means and standard deviations of purchase likelihood as a function of
salience and object relevance for focal product and filler products.
Object Relevant Focal Product Object Irrelevant Focal Product
Focal Filler Focal Filler
Product Products Product Products
Son 5.31b 3.46 3.35 3.91
Salient 1.11 1.75 1.41 1.42
Daughter 5.12a 3.91 3.20 3.37
Salient 1.36 1.30 1.86 1.76
Control Group Control Group
Focal Product Filler Products
Purchase Likelihood Relevant 3.39 Purchase Likelihood Relevant 3.93
Irrelevant 3.43 Irrelevant 4.13


Note: a= Different from control at p < .05, b=Different from control at p < .01.













CHAPTER 6
EXPERIMENT 2: THE ROLE OF SALIENCE,
SELF-IMPORTANCE and OBJECT RELEVANCE

Purpose of the Study

The framework argues that there are four key moderating factors that determine

whether the evaluative content of a social identity will lead to attitude generation.

Experiments la and lb assessed the first two factors: salience and object relevance.

Experiment two goes beyond experiments la and lb by examining the third factor; i.e.,

the moderating impact of a social identity's self-importance on the formation of favorable

consumer attitudes (based on social identification processes).

Theoretical Premise of the Study

Marketers can attempt to link consumer social identities to their product

offerings. For example, a soft drink (e.g., Sprite ) may attempt to position itself as the

beverage of choice for some social identity of interest (e.g., Generation-X non-

conformists). However, according to the framework in this research, consumers are more

likely to adopt social identities that they consider to be "self-important" in terms of a

basis for self-definition. Therefore, marketing communications that attempt to connect a

brand to some social identity of interest must consider the extent to which a social

identity is "valued" by consumers in its particular target market. Otherwise, marketing

efforts that try to induce social identity based attitude formation toward brands and

products will be ineffective. Building on experiment lb, experiment two tests the

following hypothesis:



47





48


H2(a): Heightening the salience of an identity leads to higher (lower)
evaluations of an object relevant (object irrelevant) focal stimulus,
particularly if the identity is self-important (not self-important) to the
consumer.

Therefore, the more that a consumer sees a social identity (e.g., "working mother") as

being self-important, the more likely she is to base a response toward an object relevant

product (e.g., attitudes toward mini-vans) on the evaluative content linked to that social

identity.



Experiment 2

Overview of Experiment 2. In experiment two, subjects formed attitudes under

three conditions. First, when a social identity was perceived as being relatively self-

important or not. Second, when a social identity was salient or not. Third, when the

social identity was either relevant or not to the object that was to be evaluated.

Procedure. In the first task, subjects participated in a "handwriting assessment

study" that investigated the link between consumer characteristics (e.g., being frugal) and

handwriting style. Therefore, as part of the research project, subjects were asked to give

samples of their handwriting. As part of the cover-story, in order to provide a "baseline,"

subjects first wrote three neutral sentences in their natural handwriting style. The cover

story then indicated that part of the research looked at "whether or not WHAT people

wrote about was related to their handwriting style." Therefore, subjects were instructed

to write five independent statements concerning a particular topic.

When subjects provided their handwriting samples, they wrote about their self-

conception of their "family" (Baldwin and Holmes, 1987) social identity. This was

intended to manipulate the salience of the "family" social identity. Self-importance was






49


manipulated as high (low) by having subjects write about a particular self-important

experience or event that emphasized the interdependent (independent) nature of their

connectedness with their family. In the high self-importance conditions, subjects read the

following directions:


PLEASE FOLLOW THE INSTRUCTIONS BELOW:

We'd like you to write 4 independent statements (complete sentences in your
usual handwriting style, i.e., cursive or not) each with about 10 to 15 words.
The statements should convey some positive, deeply moving, emotionally
involving thoughts and or sentiments that describe the commitment to your
relationship with one or both of your parents. Concentrate on how you have
maintained your family ties while also STRENGHTHENING the sense of
connectedness as a member of your family.


Subjects in the low self-importance conditions read the following instructions:

PLEASE FOLLOW THE INSTRUCTIONS BELOW:

We'd like you to write 5 independent statements (complete sentences in your
usual handwriting style, i.e., cursive or not) each with about 10 to 15 words.
The statements should convey some positive, deeply moving, emotionally
involving thoughts and or sentiments that describe the commitment to your
relationship with one or both of your parents. Concentrate on how you have
maintained your family ties while also STRENGTHENING a sense of independence as
an individual young adult.


Subjects then completed a battery of consumer personality measures as well as other

cover story-consistent items. The full stimulus materials for the handwriting study are

presented in Appendix E.

Subjects then participated in the same ostensibly unrelated "new product

assessment study" used in experiment lb including the same dependent measures'. To

manipulate salience, a delay condition was added between subjects. In the high salience

conditions, subjects formed attitudes toward identity relevant or irrelevant products



'This was done to allows comparisons of purchase likelihood levels across experiment
Ib and experiment 2 as well as a comparison between the control groups from experiment
Ib who had no salience or self-importance manipulation prior to evaluating the three
products.





50


immediately after generating and writing down the self-importance related thoughts. In

the low salience conditions, subjects formed attitudes toward identity relevant or

irrelevant products following a fifteen-minute delay (filler task). The same object

relevance manipulations used in experiment lb (relevant to the "family" social identity

vs. not relevant to the "family" social identity) was used in Experiment 2.

Predictions. For experiment two, the framework predicts a significant three-way

interaction of design factors on purchase likelihood of the focal product. Favorability of

attitudes (toward the focal product) as evidenced in purchase likelihood measures should

depend on whether or not a social identity is object relevant and self-important. Object

relevance (i.e., when the particular social identity is relevant to the stimulus to be

evaluated) combined with the heightened self-importance of that social identity should

increase the likelihood that the attitude object (new product concept) will be considered

in terms of the particular social identity. Hence, there should be a higher likelihood that

the evaluative content of the social identity should be the basis for attitude formation.

This should lead to more favorable evaluations of the focal product and hence, higher

purchase likelihood. The effect should be attenuated in delay conditions where the

salience induced by the self-importance manipulation should decay relative to the effect

of self-importance induced by generating interdependent, self-important thoughts

regarding the social identity. Null effects of design factors were expected for the object

irrelevant focal product and all filler products.

Experimental design. Subjects were randomly assigned to the cells of a 2 (Self-

importance: High or low) X 2 (Salience: High or low) X 2 (Object relevance: Relevant

or Not relevant) between subjects factorial design. The dependent measure was purchase





51


likelihood for the focal product and the average purchase likelihood for the two filler

products.

Manipulation checks. To check the self-importance manipulation, a separate

pretest was run. Twenty-seven (N=27) subjects first completed the handwriting

assessment study. Subjects then answered fifteen manipulation check items intended to

assess the effectiveness of the self-importance manipulation (e.g., I strongly identify with

my Family, Being a member of my family often affects how I tend to view the world

around me, etc). Items ranged from 1= Strongly disagree to 5=Strongly agree and four of

the items were reverse coded.

After completing the handwriting task, subjects completed an ostensibly unrelated

consumer emotion and mood study" where they responded to the two 10-item mood

scales (using "momentary" instructions) that comprise the Positive and Negative Affect

Schedule (PANAS) Watson, Clark and Tellegen (1988) included in Appendix F of this

dissertation. This was included to assess and or rule out the alternative explanation that

mood changes induced by the self-importance manipulation may have led to any

observed effects.

The fifteen manipulation check items were collapsed into one measure of self-

importance (a =.783) and the means of the self-importance measure were calculated as a

function of the low and high self-importance conditions. The averages in the low and

high self-importance conditions were Mlowselfimp=3.57 Mhighselfmp=4.19, (F(1,25) = 29.58

p<.001) respectively, indicating that the self-importance manipulation altered subjects

perceptions of the self-importance of their "family" identity. As a function of low and





52


high self-importance conditions, no differences in mood were found on either the positive

affect (PA) or negative affect (NA) dimensions of the PANAS scale (F's < 1).

Results. One hundred twenty-one (N=121) subjects from a southeastern university

were randomly assigned to experimental treatments. An analysis of variance tested the

main and interaction effects of self-importance of the "family" identity (low vs. high),

salience (low vs. high) and object relevance (relevant: framed in terms of features +

family or irrelevant: framed in terms of just features) on mean purchase likelihood for the

focal product and the two filler products. The omnibus ANOVA effects of salience,

relevance and self-importance on purchase likelihood of the focal product and the filler

products are reported in Appendices J and K respectively. The average purchase

likelihood as a function of salience, self-importance and object relevance for the focal

product and filler product 1 and filler product 2 are presented in figures 5, 7 and 9

respectively.

For the focal product, although the means were in the predicted direction, the

three way interaction of self-importance, salience and object relevance was not

significant, (F < 1). However, consistent with predictions, main effects of self-

importance, (F(1, 113)=11.30, p < .01, and object relevance, (F(1,113)=14.28, p <.01

were qualified by a significant two way interaction of self-importance by object relevance

(F(l,113)=4.33, p < .05). Subjects' purchase likelihood of the focal product depended on

whether or not the self-importance of their family identity was low or high and whether

or not the focal product was framed in an identity object relevant or irrelevant fashion.

To examine the nature of this result, figure 6 shows mean purchase likelihood of the focal

product as a function of self importance and object relevance (collapsed across salience





53


conditions). Consistent with predictions, the mean purchase likelihood was highest in the

high self-importance / identity object relevant condition, M=5.33 than in the high self-

importance / identity object irrelevant condition, M=3.64, the low self-importance /

identity object relevant condition, M=3.76 or the low self-importance / identity object

irrelevant condition, M=3.27. As a secondary analysis, Dunnett's procedure was again

used to test differences between treatment means of purchase likelihood against the base-

line control groups (from experiment lb) who received no manipulation prior to the

evaluation of the new product concepts. Again consistent with predictions, the only

treatment mean that differed from its control group was when "family" identity was

highly self-important in the object relevant condition (Mselfimphigh/objectrelevant -

MCONTROL = 1.95, critical difference = 1.51, p < .01 FWa- = .05). Curiously, for filler

product 1, a significant main effect of self-importance (F(1,113)= 6.53, p < .05) was

found. Collapsing across self-importance and salience conditions, revealed that high self-

importance of the "family" identity lead to greater purchase likelihood for the second

filler product (Mlowselfimp = 2.32, Mhighselfimp= 3.15) which was a high-tech electronically

air-cooled head band. However, a follow up analysis comparing these treatment means

further broken down by identity relevance conditions revealed no significant differences

as compared to relevant control groups from experiment lb. Finally, consistent with

predictions, an analysis revealed no systematic treatment effects on purchase likelihood

for the third filler product (F's < 1) and no significant differences between corresponding

control groups and treatment means of purchase likelihood for the third filler product.

Discussion and limitation(s) of the experiment. The results of experiment two

provide support for hypothesis 2(a). The pattern of data is consistent with the assertion





54


that self-importance is an important moderating variable in the process of social identity

based attitude formation. Results of this experiment suggest that when people see a

particular aspect of their social identification as being relatively self-important, they are

more likely to favorably evaluate an identity relevant object. Based on relevant control

group comparisons, this pattern was found for the focal product and not for filler products

predicted to be deemed irrelevant to the particular social identity in question. These

effects appear not to be confounded with any differential levels of positive affect that

might have been associated with generating high versus low self-importance thoughts.

Several limitations of this experiment are of note, however. First, an effect in the

omnibus test for purchase likelihood of filler product 1 was unexpectedly found.

Subjects who perceived their "family" identity as being self-important, on averaged

reported higher purchase likelihood for the high-tech electronically air-cooled head band.

This effect seems somewhat odd. A review of the description of filler product 1 was

conducted and is shown in figure 9. Two things are of note here. Firstly, and purely

based on speculation, the description of filler product 1 starts with a lead-in rhetorical

question "When it's blazingly hot, don't you dread leaving the air-conditioned comfort of

your home?" It might be possible that subjects in the high self-importance conditions

(who generated thoughts concerning their interdependence with their parents) somehow

interpreted "home" as having some connection to their familial identity. This might have

lead to slightly higher purchase likelihood. Secondly, an examination of the mean levels

for filler product 1 shows that mean purchase likelihood did not reach the midpoint of the

scale for any of the treatment groups. This means that no one really "liked" the product

that much (to the extent that favorability maps onto purchase likelihood), rather in the





55


high self-importance conditions, subjects disliked it less than subjects in the low self-

importance conditions. This notion of "statistical significance" versus "practical

importance" combined with the fact that the follow up control group analyses suggested

no real differences, is taken for evidence that this effect may be spurious and or trivial.

Additionally, in Experiment 2, salience did not interact with design factors as

predicted by the theory. There are several reasons why this might have occurred.

Although the manipulation check on self-importance revealed a statistically significant

difference on the fifteen item self-importance measure for the hold-out sample, inspection

of the actual mean levels suggests that the psychological difference in subject's minds

might have been minimal. This makes sense given the reasonable expected difficulty in

heightening the self-importance of a person's familial identity. However, because the

mean differences in the low and high self-importance manipulation check appear to be

minimal, the fifteen minute intervening task intended to alter the salience induced by

generating and writing down self-important related thoughts, may have been at an

inappropriate delay interval. Another speculative explanation might suggest that the

salience of this kind of social identity (i.e., "family") may be highly correlated with self-

importance. The operationalization of salience in experiment two is based on the

assumption that the salience of social identification induced by the self-importance

manipulation decays at a relatively different rate than the salience of the self-importance

related thoughts per se. Research has suggested that certain bases for self-conception

tend to be chronically salient for some individuals (Bem, 1981). An interesting follow up

study might attempt to determine the precise self-importance thresholds (for a particular

social identity) beyond which the social identity tends to be chronically activated.





56


Unfortunately, no manipulation check proxy for salience in this study was taken.

Nonetheless, support was found for the assertion that self-importance (as operationalized

in this study) is a moderating factor in determining the extent to which the evaluative

content of a social identity will impact on attitude generation.






57






7 7 Relevant (Family
+ Features)

6 6 Irrelevant
(Features)





3 3

2 2



High Self- Low Self- High Self- Low Self-
importance Importance Importance Importance
Self-Importance Self-importance



A B
Figure 5.
Experiment 2: purchase likelihood of Focal product as a function of object relevance,
self-importance and low salience (A) and high salience (B) of the social identity




7 ---

6

0
o 5
S[ Identity Object Irrelevant
Focal Product
_J4
Identity Object Relevant
(U Focal Product






Low High
Self-Importance




Figure 6.
Experiment 2: purchase likelihood of Focal product as a function of object relevance and
self-importance (collapsed across salience conditions)






58





7i Relevant (Family
7 7 + Features)
.Irrelevant
6 6 (Features)

0o5




3. -
22

1
High Self- Low Self-
hel- Lo High Self- Low Self-
Importance Importancef- Low Se
Importance Importance
Self-im portance S
Self-Importance


A B
Figure 7.
Experiment 2: purchase likelihood of Filler Product 1 as a function of object relevance,
self-importance and low salience (A) and high salience (B) of the social identity



PERSONAL COOLING SYSTEM

When it's blazingly hot, don't you dread leaving the air-conditioned comfort of your home? Maybe you've
tried an ice-soaked towel around your neck but that little bit of evaporative cooling is nothing
compared to the long-lasting, total-body comfort of our all-new 2.0 version of the Personal Cooling
System. It's the best way to stay comfortable and avoid heat stress when walking, mowing,
gardening, watching sports or concerts...just about any activity, anywhere.

This new and improved 2.0 model offers two levels of cooling: "regular" helps you feel up to 20F cooler
than the hot, dry air around you; and "high" now offers 20 percent greater cooling for even the hottest
days. What's more, the tiny fan inside version 2.0 adds even more cooling by directing a gentle breeze
over the back of your neck.

The 2.0 model also is adjustable to fit a wider range of neck sizes comfortably and securely. It's trimmer
and lighter, too: the sleek, die-cast aluminum body weighs just 10 oz. empty; 14 oz. filled. Runs on one
AA battery (order separately).

Figure 8.
Experiment 2: description of Filler Product 1






59






7 7 i Relevant (Family
+ Features)
= Irrelevant
6-: 6- (Features)






3 3
I=


2 2


High Self- Low Self-
krportance Importance High Self- Low Self-
Importance Importance
Sef-Self-Iportanmportancece
-J












A B
Figure 9.
Experiment 2: purchase likelihood of Filler product 2 as a function of object relevance,
self-importance and low salielien(A) and high salience (B) of the social identity
self-importance and low salience (A) and high salience (B) of the social identity














CHAPTER 7
EXPERIMENT 3: THE ROLE OF SELF-IMPORTANCE
AND EVALUATIVE DIAGNOSTICITY

Purpose of the Study and Hypotheses to be Tested

Suppose a consumer's political identity (e.g., "Democrat") is highly salient (e.g.,

the consumer is attending a political rally) at the time an attitude toward an identity

relevant object is called for (e.g., the consumer is asked to assess and sign a petition

against estate taxes). Suppose that the consumer is asked to either giver her attitude

toward the estate tax, or judge the readability of the real estate tax petition itself. The

first judgment (attitude toward estate taxes) may be highly diagnostic to her Democrat

identity (i.e., a clear norm might exist Kallgren, et. al 2000) in which case her political

identity is likely to be the basis for the response. However, the second judgment

(readability of the petition) may be less diagnostic to her political identity, minimizing

the effect of her political identity's salience and self-importance on her subsequent

judgment.

Experiment three is a partial replication of experiments Ib and 2, but goes beyond

those by demonstrating that social identity will impact consumer attitude formation only

when the social identity is evaluatively diagnostic to the judgment (Feldman and Lynch,

1988). In experiment three, this notion was tested by examining the relationship between

self-importance of a social identity and the formation of response judgments in instances

where the identity does (does not) provide and evaluative basis to respond:




60





61


H3(a): The self-importance of a social identity is more likely to be positively
related (unrelated) to evaluative dimensions that are diagnostic (non
diagnostic) to the social identity.

Overview of the Experiment

During the experiment, subjects generated evaluative responses under two

different conditions. When the social identity was salient (not salient), and when the

attitude object was relevant (not relevant) to the social identity. The key distinction in

this experiment was that the evaluative responses tapped dimensions in which the social

identity provided a basis (no basis) to make a judgment. More details follow.

Experiment 3

Participants and experimental design. Subjects were taken from a sample of 140

freshman and sophomore students, ages fourteen to sixteen at a southeastern high school.

They participated as part of course credit. The students were about 50% White, 30%

Black, 15% Hispanic, and 5% other. The sample had an equal number of males and

females. In experiment three, the students were randomly assigned to the cells of a 2

(Salience: High or moderate) X 2 (Object relevance: Relevant or Not relevant) between

subjects factorial design. In addition, participants were classified as either high or low in

identification (self-importance) with the social identity.

Procedure. At time one, a pretest was conducted to determine various self-

important social identities for this particular sample. In this pretest, subjects

spontaneously generated various social identities in response to the prompt "describe

yourself." Based on their open-ended responses, a list of various social identities was

generated. Several weeks later at time two, a series of "social identity" scales was

administered. These measured the "self-importance" (i.e., whether participants identified





62


with, admired and characterized themselves as being a person who holds the identity) of a

subset of social identities generated at time one. Based on a preliminary analysis of the

data generated at time to and to avoid ceiling or floor effects, a social identity was chosen

(for use in the main experiment) in which the distribution of self-importance scores was

roughly uniformly distributed amongst participants.

Approximately two weeks later at time three, each subject was given a series of

various surveys and research tasks. Buried within the tasks was a "consumer web-page

assessment study." Subjects were told that marketers often want to understand the

potential persuasiveness of web page content. In order to understand how to create web

page media that have the most impact, message content and web page design is often pre-

tested on relevant populations. Subjects were told that this study would assess their

perceptions regarding types of message content, types of web page content, web page ads

and or products/services.

Independent variables. The social identity chosen from the pretest was "Future

College Educated Leaders of America." The background information subjects read prior

to the main experiment manipulated the salience of the social identity. Subjects were told

that research pre-testing on the consumer web-page assessment study had already begun

on three "market segment clusters" that were subsequently described.


In the high salience conditions, the last of three "segments" read as follows:

This research has also been conducted on a segment conveniently entitled
"Towns and Gowns." This segment makes up about 1.2% of U.S.
households. The towns and gowns segment earns on average about
$17,862. Members of the Towns and Gowns segment typically range
from ages 18 to 34. They have been described as college educated, and
highly likely to be the future leaders of this country. They are likely to
read magazines like Natural History and watch television shows like Good





63


Morning America. Sample ZIP codes include College Station, Texas
77840; Bloomington, Indiana 47401; Ithaca, New York 14850 and Ft.
Lauderdale, Florida 33324.

In the low salience conditions, the last of three "segments" read as follows:

This research has also been conducted on a segment conveniently entitled
"Blue Blood Estates." This segment makes up about 1.2% of U.S.
households. The blue blood estates segment earns on average about
$17,862. Members of the blue blood estates segment typically range from
ages 18 to 34. They have been described as high school educated, and
highly likely to possess reasonable spending power. They are likely to
read magazines like Natural History and watch television shows like Good
Morning America. Sample ZIP codes include College Station, Texas
77840; Bloomington, Indiana 47401; Ithaca, New York 14850 and Ft.
Lauderdale, Florida 33324.

In the main experiment, subjects examined the target product/service of interest which

was a Smithsonian Magazine Association membership. The subjects saw an example of

the Web page. Subjects were then asked to take several seconds to inspect the web page.

They were told to read all information carefully. Afterwards, subjects read a description

of the service that was depicted on the web page.

In the object not relevant condition, the Smithsonian Association membership was

described in terms of strict, utilitarian benefits (attributes):


Why would you want to be a Smithsonian National Associate Member? Just
think about the pros and cons! Being a Smithsonian National Associate Member
has many advantages. For example, each member receives twelve total issues of
the Smithsonian Magazine at a special rate that has been discounted specifically
for Association members. Twelve issues sent directly to your home or office. In
addition to book, record and video discounts as well as other discounts at various
museum shops, each member is also kept up to date on various notices of
important events and intellectual activities going on in your area. Also, members
of the Smithsonian National Association receive eligibility for specially
designated travel programs, as well as free admission to a number of the country's
best museums and other cultural events. All of these amazing features for just
$26.00 for the entire year! Why don't you just weigh the costs and benefits! How
can you lose?





64

In the object relevant condition, the Smithsonian Association membership was

described in terms of the reference group in question:


Why would you want to be a Smithsonian National Associate Member? Just
think about whom might be likely to possess such a membership? Think about
the type of people who would receive the twelve issues of the Smithsonian
Magazine that is included in the membership. What type of person? Individuals
who always present themselves with the highest level of intelligence and future
success. For example, college educated individuals of this country might keep up
to date on various notices of important events and activities going on in their area.
These Smithsonian National Association members are the future leaders of this
country. The magazines that they read tell others that they appreciate the more
culturally involving things in life. Can't you imagine yourself as this type of
person, i.e., a college educated, future leader of this country and Smithsonian
Associate Member? All of these amazing features for just $26.00 for the entire
year! How can you lose?

After reviewing the above information, subjects rated the web page on various

evaluative dimensions and then reported their attitude toward the Smithsonian Associate

membership.

Dependent measures. Subjects were asked to rate the example web page on the

following six dimensions: interestingness, readability, visual appeal, persuasiveness,

complexity, and design quality. The main dependent measure asked subjects to rate the

extent to which they felt favorable or unfavorable toward the Smithsonian Association

membership, which ranged from unfavorable (1) to very favorable (7). The full set of

stimulus materials for this experiment is presented in Appendix G.

Predictions. Of the seven dependent measures that were chosen, three are

expected to be diagnostic to the salient social identity: interestingness and persuasiveness

of the web page as well as overall attitude toward the product depicted on the web page.

The other four dimensions: readability, visual appeal, complexity and design quality of

the web page are expected to be non-diagnostic. Therefore, the self-importance of a





65


salient (not salient) social identity--Future college leaders of America--to an object

relevant (object irrelevant) stimulus--Smithsonian National Associate Membership

framed in terms of the social identity--should be more positively correlated (uncorrelated)

with dimensions that are diagnostic-- interestingness, persuasiveness, attitude than

dimensions that are (nondiagnostic)--readability, visual appeal, complexity and design

quality) to the social identity.

Manipulation checks. A hold out sample was used to check both the salience and

the object relevance treatment manipulations. To check the salience manipulation,

subjects were asked to read the intro and background information to the Web page

assessment study. One of the questions asked the subjects to indicate at that particular

moment, to what extent did the information make them think about their identity as a

future college educated leader of America? The scale ranged from Did not make me think

about it (1) to Made me really think about it (7). The difference across the high salience

and low salience conditions was significant (N=34, Mhigh=4.47, Miow=3.53, F(1,

32)=4.59, p=.04). To check the object relevance manipulation, subjects were asked to

read the descriptions of the service. One of the questions asked the subjects to indicate to

what extent did the service seem relevant to what it means to be a Future college

educated leader of America ranging from Not relevant (1) to Very relevant. The

difference across the object relevant and object not relevant condition was highly

significant (N=32, Mrelevant=5.25, MNotreevant=2.44, F(1, 30)=65.60, p <.001).

Results. The key effects tested in this experiment were the relationships between

evaluative dimensions and the self-importance of the social identity in question.





66


Evaluative dimensions. Separate analyses were run on each of the six evaluative

dimensions (interestingness, readability, visual appeal, persuasiveness, complexity, and

design quality) and the subjects reported attitude toward joining the Smithsonian

Associate membership. For these analyses, self-importance of the social identity that was

measured two weeks prior to the main experiment was included in the analysis as an

independent continuous variable. Self-importance of the social identity consisted of the

average of three items (a = .853). In a prior survey, subjects were asked to indicate the

extent to which "future college educated leaders of America" described them ranging

from Does not describe (1) to Describes me perfectly (7), the extent to which they

identified with that group ranging from Do not identify w/ group in any way (1) to

Strongly identify with the group (7), and whether or not they admired the group ranging

from Do not admire the group (1) to Really admire the group (7). Predictions were tested

in an ANCOVA including main effects of salience and object relevance and all possible 2

and 3 way interactions of self-importance, salience and object relevance. The omnibus

ANCOVA effects for each dimension is reported in Appendix L.

The analysis revealed a significant three-way interaction of self-importance x

object relevance x salience (F(1, 90) = 18.84, p < .0001). The effect of self-importance

of the social identity on subjects perceptions of the interestingness of the web page,

depended on two things: whether or not the subject's social identity (future college

educated leaders of America) was made salient or not and whether or not the object was

relevant (message was framed in terms of the social identity or object features). A

separate model was run to investigate the nature of this result. Figure 10 reports a





67


separate, slope estimate of the self-importance continuous variable nested in each

treatment condition. Corresponding t value(s) are in parenthesis.

Consistent with predictions, the key result here is that the effect of self-

importance of the social identity is related to this relevant dimension only when the social

identity is salient and the attitude object is relevant to the social identity (i.e., when the

message has been framed in terms of the social identity). So in the high salience and

object relevant condition, every one-unit change in self-importance of the social identity

results in a statistically significant, .62 positive unit change in "interestingness" (t=6.95, p

<.0001) of the web page. This is indirect evidence that self-importance of the social

identity impacts the formation of this relevant evaluative response. As self-importance of

social identification increases, so do perceptions of the webpage's interestingness.

The same analysis was run on the "readability" and the "visual appeal"

dimensions. According to the framework, the self-importance should be related only to

evaluatively diagnostic dimensions. It should have no effect (should not interact with the

treatment conditions) in determining perceptions of "readability" or "visual appeal."

Consistent with predictions, no effects were found to be significant (F's < 1).

An analysis examined the effects of self-importance, salience and object

relevance on the "persuasiveness" dimension. This analysis revealed a significant two-

way interaction of self-importance x salience for the persuasiveness dimension (F(1, 90)

= 6.53, p < .0123). The effect of self-importance of the social identity on subjects

perceptions of the persuasiveness of the web page, depended on whether or not the

subject's social identity was made salient or not. A separate model was again run to

investigate the nature of this result. Figure 11 reports a separate slope estimate of the





68


self-importance variable nested in each class variable treatment condition.

Corresponding t value(s) are in parenthesis.

Very similar to the "interestingness" dimension and consistent with predictions,

the key result here is that the effect of self-importance is positively related to this relevant

dimension only when a social identity is salient. More specifically, in the low salience

condition, every one unit change of self-importance results in a .483 unit change of

persuasiveness (t=5.34, p<.0001) while in the high salience condition, every one unit

change of self-importance results in a .729 unit change of persuasiveness (t=7.23,

p<.0001). This is again, consistent with the notion that self-importance impacts the

formation of this relevant evaluative response. As self-importance of the social identity

increases, so do perceptions of persuasiveness of the web page. Subjects in these

conditions may be relying on social identification as a basis for generating perceptions of

message persuasiveness.

The same analysis was run on the "complexity" dimension. Surprisingly, this

analysis revealed a marginally significant three-way interaction of strength of self-

importance x object relevance x salience interaction for the complexity dimension (F(1,

90) = 4.19, p < .0435). The effect of self-importance of the social identity on subjects

perceptions of the complexity of the web page depended on whether or not the subject's

social identity was made salient or not and the attitude object's relevance (i.e., the type of

message that they were exposed to). A separate model was again run to investigate the

nature of this unexpected result. Figure 12 reports a separate slope estimate of the self-

importance variable nested in each treatment condition. Corresponding t value(s) are in

parenthesis.





69


These results are similar to the "interestingness" and "persuasiveness"

dimensions. However, the self-importance of the social identity impacts this dimension

only when a social identity is salient and the message is framed in terms of the social

identity. In the high salience conditions, perceptions of complexity are correlated with

self-importance of the social identity. More specifically, in the high salience condition,

every one unit change of self-importance results in a weak, albeit significant .155 unit

change of perceptions of complexity (t=1.81, p<.10).

The same analysis was run on the "design quality" dimension. Consistent with

predictions, this analysis revealed no significant interactions between self-importance of

the social identity and the other conceptual variables of interest (F's < 1).

After evaluating the web page on the six dimensions previously outlined, subjects

were asked to give their attitude (very unfavorable (1) to very favorable (7)) toward the

Smithsonian Associate membership. This analysis revealed a significant two-way

interaction of self-importance x object relevance for the attitude dimension (F(1, 90) =

30.84, p < .0001). The effect of self-importance on subjects' attitude toward the

Smithsonian membership depended on whether or not the message was framed in terms

of the social identity (object relevant) or object attributes (object not relevant). A

separate model was once again run to investigate the nature of this result. Figure 13

reports a separate slope estimate of self-importance nested in each treatment condition.

Corresponding t value(s) are in parenthesis.

Consistent with predictions, the key result is that the effect of self-importance of

the social identity is related to attitudes toward the Smithsonian Associate membership

only when the message is framed in terms of the social identity. In the moderate salience





70


conditions, every one unit change of self-importance results in a .543 unit change in

attitudes toward the Smithsonian membership (t=4.73, p<.0001) while in the high

salience condition, every one unit change of self-importance results in a .6947 unit

increase in attitudes toward the Smithsonian membership (t=5.43, p<.0001). The effect

of self-importance of the social identity in these two treatment conditions closely mirrors

the pattern of results found for the persuasiveness dimension. This is intuitively logical,

because out of all six dimensions, one could argue that the "persuasiveness" dimension

should probably be most correlated with overall attitude. Nonetheless, this result is again

consistent with the notion that self-importance of the social identity impacts the

formation of attitudes toward the object when the product is framed in terms of social

identification based concerns.

Discussion and limitations. It was hypothesized that the self-importance of a

social identity is more likely to impact attitude formation when a salient social identity

provides a basis for an attitude. The results of this experiment provide indirect evidence

for this assertion. Strong support was found for hypotheses H3a. For example, self-

importance of the social identity positively co-varied with interestingness of the web page

for participants who were led to form an attitude based on social identity and for whom

the social identity was made salient. This effect was somewhat stronger for the

persuasiveness dimension where a stronger effect of self-importance of the social identity

was found when social identity salience was high versus low. Although it was a much

weaker effect statistically, the pattern of results for the complexity dimension somewhat





71


mirrored the results for the interestingness dimension. This was not expected.' However,

a very strong effect of self-importance of the social identity was also found for the

attitude dimension as a function of object relevance. This pattern of slope effects was

very similar to the results found for the persuasiveness dimension. In general, this

experiment provides consistent evidence that social identification (more specifically, the

potency of group identification as captured by the self-importance variable can be related

to the formation of responses in conditions where conceptually, a consumer is more likely

to be drawn to a salient social self-identity which provides a basis for a response

(message framed in terms of social identity) on measures that are evaluatively diagnostic

to the social identity.

Several limitations of this study are of note, however. Firstly, although taken

approximately two weeks earlier, the self-importance of the social identity was a

measured variable. Of course, this weakens the ability to assert a causal relationship

between self-importance of the social identity and attitude formation in this experiment.

Therefore, a more stringent test of the theory would attempt to manipulate the self-

importance of a social identity (as was attempted in experiment two). Secondly, no direct

behavioral measure was taken.2 Therefore, one could argue that the attitudes expressed

by participants in this experiment would not actually be diagnostic guides for action.


1 Subjects rated the web page's complexity immediately following the persuasiveness
dimension. Although purely post hoc speculation, it might be argued that the
idiosyncratic nature of the sample of participants (high school students) and the close
proximity of these two dependent measures may have led them to infer some kind of
relationship between persuasiveness and complexity.

2 Philosophically, it can be argued that expressing an attitude is a "behavior." However, a
more complete test of the framework will include additional measures that represent
action toward the object in question.





72


This criticism is quite valid. But the focus of this research is on how the evaluative

content of a social identity might affect attitude formation. Whether or not the formed

attitude will guide actual behavior is a separate issue.

Thirdly, the manipulation check taken on the salience manipulation was quite

weak both in its statistical significance and the mean levels across conditions. However,

the results show that in the relevant treatment conditions, the patterns of results produced

the expected variation in the measure of the relationship between the conceptual

independent and dependent measures. Therefore, a specific alternative explanation must

be offered that assumes no relationship between the conceptual independent and

dependent variables described by the framework being tested. If no such alternative

explanation exists, then having a "stronger" independent variable check does not add

anything necessary, and even a failure to have one does not constitute a flaw (Sigall and

Mills, 1998). A related issue is that the operationalization of the object relevance

manipulation probably made the social identity salient too. Therefore, the relative role on

the dependent measures of pure salience induced by the introduction of the experiment's

cover story and salience induced by the object relevance manipulation is difficult to

distinguish in this experiment. Certainly, an issue for future research.

Finally, there is a different general theoretical account for these findings.3 This

experiment provides no direct evidence for the process described in the overall model

(figure one). The results presented in experiment one do not speak directly to



3 A different general theoretical account is not the same as an alternative explanation.
The former "reinterprets" the proposed causal relationship in the context of another
framework. However, the latter assumes no relationship between the conceptual
variables that the theory asserts to be related and may propose that totally different
conceptual independent variables resulted in dependent measure variation.





73


differentiating the process of attitudes formed on the basis of a social identity from other

processes. For example, take a multi-attribute explanation of attitude formation. It can

be argued that subjects in the object relevant (framing of the message in terms of social

identity) conditions, were simply led to create an additional "belief" about the object, and

the salience of the social identity simply altered the evaluative implication of that

additional belief. Of course it is possible to post hoc explain these or any other results in

terms of a multi-attribute model (i.e., shifting importance weights of newly created belief

elements in a multi-attribute premise). However, the purpose of this framework is to

describe the process of how a consumers' social identification is likely to affect her

attitude. Nonetheless, future experiments must also incorporate other process measures

that would be expected to differentiate other theoretical accounts.





74




OBJECT
RELEVANCE
NOT
RELEVANT
RELEVANT




MODERATE


-.004 .114
SALIENCE (t = -.05) (t 1.31)



*^--
HIGH .__.



.62 .003
(**t = 6.95) (t =.04)



X = SELF-IMPORTANCE ; Y =INTERESTINGNESS




Figure 10.
Experiment 3: interestingness as a function of salience, self-importance and object
relevance





75




OBJECT
RELEVANCE
NOT
RELEVANT
RELEVANT



MODERATE


.483 -.152
SALIENCE (* t = 5.34) (t =- 1.54)



HIGH .



.729 .091
S(**t= 7.23) (t=.99)



X = SELF-IMPORTANCE ; Y =PERSUASIVENESS




Figure 11.
Experiment 3: persuasiveness as a function of salience, self-importance and object
relevance





76




OBJECT
RELEVANCE
NOT
RELEVANT
RELEVANT



MODERATE


-.037 .0984
SALIENCE (t=-.48) (t=- 1.18)



HIGH -*--



.1545 -.043
(t = 1.81) (t = -.54)



X = SELF-IMPORTANCE ; Y =COMPLEXITY


Figure 12.
Experiment 3: complexity as a function of salience, self-importance and object
relevance





77







OBJECT
RELEVANCE
NOT
RELEVANT N
RELEVANT



MODERATE


.543 .039
SALIENCE (**t =4.73) (t= -.31)



HIGH



~,./ 6947 -.0724
t (**t= 5.43) ( t=-.62)



X = SELF-IMPORTANCE ; Y =ATTITUDE

Figure 13.
Experiment 3: attitude as a function of salience, self-importance and object relevance














CHAPTER 8:
GENERAL DISCUSSION AND CONCLUSIONS

The research presented here draws upon research in social cognition, functional

attitude theory, and social identity theory to organize a unifying framework for how a

consumer's social identification is likely to lead to the formation of her attitude. The

framework of this research identified four key factors that increase the likelihood that a

consumer will form an attitude based on social identification processes. Several

experiments investigated this assertion and aggregate results across studies suggest that a

social identity pathway to attitude formation has both verifiable theoretical import and

well as useful descriptive value.

For example, part of the current research examined the role of a social identity's

salience on people's spontaneous self-concepts (McGuire et al., 1979. 1981) and how this

impacts the formation of their attitudes. It was found that the salience of a particular

basis for self-definition ("sons and daughters") mediated people's momentary self-

conceptions of themselves. This impacted people's judgments (purchase likelihood)

when they were evaluating an object (new product concept) that was relevant to that

activated identity. This is consistent with social cognitions' view that "activated

concepts" can affect judgments, as well as consistent with more recent findings in the

literature concerning the dynamic nature of constructed judgments and retrieved attitudes

(Schwarz and Bohner, 2000; Wilson and Hodges, 1992). For example, researchers

working in the area of "biased information scanning" found that compared to white



78





79


females, black females who had recently had their ethnic (female) identity made salient,

had more favorable (unfavorable) perceptions of OJ Simpson's innocence (Newman,

Duff, Schnopp-Wyatt, Brock and Hoffman, 1997). In the stereotype literature, another

recent study demonstrated that when a particular social identity was made salient at an

implicit level, performance was altered in the direction predicted by the stereotype

associated with the identity. In this study, Shinh, Pittinsky and Ambady (1999) found

that Asian American women performed better on a math test when their ethnic identity

was activated but worse when their gender identity was activated as compared with a

control group who had neither identity activated. Therefore, a consumer's response to a

brand, product or company can be markedly different depending on how they are

thinking about themselves at the moment they generate an attitude.

In the current research, the data also suggest that the extent to which a consumer

identifies with a particular social identity (i.e., the self-importance of the social identity)

moderates the extent to which the evaluative content of an adopted social identity will

impact attitude formation. When people saw their "family" identity (Baldwin and

Holmes, 1987) as relatively more self-important than not, they were more likely to

favorably evaluate a product that was relevant to their "family" identity. This finding is

in concert with early discussions of "identification" (Kelman 1958) as a basis for

adopting an attitude and the rich literature in social psychology that describes

motivational underpinnings of social identification in terms of how objects can be seen as

symbolically linked to social identities people value (Katz, 1960; Smith, Bruner and

White, 1956; Shavitt, 1990). The results presented here are also consistent with recent

work in which researchers have argued that psychological "group membership" can





80


influence people's attitudes. For example, Terry and Hogg (1996) presented evidence

that norms linked to a behaviorally relevant reference group ("regular exercisers")

influenced intentions to engage in exercise but only for people who identified with the

group (Terry and Hogg, 1996; see also Ybarra and Trafimow, 1998).

Lastly, evidence presented here suggests that a person's salient and self-important

social identity will only impact evaluative responses that are diagnostic to the social

identity (c.f. Feldman and Lynch, 1988). In other words, adopting a social identity as

one's social identity must provide a person with a meaningful basis to respond to

something that they want to judge. This notion is consistent with recent findings in the

attitude recruitment literature (c.f. Wilson, Lindsey and Schooler, 2000). For example, in

a very complex hybrid lab/field study, Cohen and Reed (2000) showed that, outside of

any obvious pressure to conform, heightening the salience of a particular social identity

("women in business") affected the way that people recruited attitudes toward the object

("affirmative action"). The results of their study suggest that when identity was salient

(not salient) attitudes were more aligned with group membership norms (personal

attitudes) when the membership group had a clear (unclear) position on the issue (see

also, Kallgren et al., 2000). Therefore, social identification may be salient, self-important

and relevant to a product or brand; but may also leave the consumer unable to generate an

attitude if that identity does not provide some direction with regard to how that "type" of

person will or should respond.

Moving beyond a multi-attribute centered focus. The point was made earlier that

the framework of this dissertation may provide added value above and beyond an object

centered multi-attribute approach. This begs the question as to whether the effects and





81

phenomena described here can be equally formulated, predicted, described and explained

by multi-attribute models in general or more sophisticated attitude frameworks such as

the classic behavioral intentions model.

A multi-attribute model defines an attitude in terms of beliefs and weights. The

point that has been made by some researchers is that perhaps the best way to think about

a multi-attribute model is not as a psychological process but as a very simple and useful

measurement method that can derive a result.' For example, in past discussions and

debates, (c.f. Miniard and Cohen, 1981) researchers have argued that having a model that

incorporates everything precludes a recognition of potential differences among the kinds

of components that may make up the person's summary evaluation. For example, it does

not matter in a multi-attribute model whether the beliefs have to do with product

attributes, or social identity, personal values or some other aspect. This is potentially

important for the following reason: If a researcher focuses on product attributes and the

salient beliefs linked to them and it turns out that the driving force behind the consumer's

attitude has more to do with symbolic associations such as social identity, then it is

difficult if not impossible to understand what the process is that led to the attitude. Of

course, the response to this line of reasoning is that a multi-attribute model can

adequately measure the social identity based beliefs associated with an attitude based on

social identification processes. For example, if a consumer believes that some product is

for freedom loving people, and the consumer values freedom loving people, then the

consumer will have a more favorable attitude toward the product. The caveat is that if


'This is not a criticism of multi-attribute models because their purpose is not to describe
a psychological process. That might confuse some researchers because multi-attribute
models seem to have a process involved in the multiplication of evaluative implications
times beliefs summed across salient beliefs.





82


one does not care about the key internal and external factors that may lead a consumer to

form their attitude based on a self-conception of themselves as desiring to be or appear to

be a freedom loving person, then it is not a problem to use a multi-attribute approach to

simply ask people what their attitudes are based on specifying the set of social identity

related beliefs. But, if one wants to know when a particular kind of attitudinal influence

like social identification is likely to lead to an attitude and to understand how this

influence occurs, then one is interested in understanding the process that underlies it. In

line with the latter emphasis, this dissertation starts with an interest in social identity as

one potential driving mechanism of an attitude formation process. In this dissertation

research there is no quarrel with the original multi-attribute model. It is designed as a

measurement tool, it is definitionally correct and it is certainly acknowledged that the

dissertation's empirical work cannot be distinguished from what can be post hoc

explained by any multi-attribute model. Nonetheless, research is more and more coming

to understand that non-cognitive influences might be particularly important in certain

instances, and it is doubtful whether that is nearly as well represented in various kinds of

multi-attribute models (c.f. Terry and Hogg, 1996). To the extent that social identity

based influences do not rise to the level of beliefs or that they really should be

characterized somewhat differently, then these kinds of qualitatively different influence--

such as social identity--might not translate into beliefs very well.

An expanded view of social influence. With respect to the behavioral intentions

model, a very brief discussion is necessary in which the point will be made that the major

point of difference with the classic behavioral intentions model is its view of social

influence i.e., in treating all social influences as being normative in nature. Heavy use of





83


the behavioral intentions model has led many consumer researchers to characterize social

influence on attitudes as some outside force that stops a person from willingly doing what

they would otherwise do. For example, imagine a mother whose son is doing poorly in

school because he is spending too much time playing basketball. He does not like school.

Besides, he says that he has to practice because he really wants to be like his favorite

basketball players in the NBA. However, mom is now angry. She wants her son to do

better and to be more favorable toward his studies. Consider two interventions: 1) The

mother threatens to punish her child (withhold allowance) if he does not do well in school

by maintaining balanced academic and athletic interests. 2) The mother shows her son an

article where his favorite basketball players describe the value of maintaining balanced

academic and athletic interests. Both interventions may lead to behavioral change, i.e.,

the son might study more. However, these two types of influence may be thought of

quite differently. Normative influence (as might be described by the first intervention

and heavily reflected in the behavioral intentions model) affects public expressions

(Deutsch and Gerard, 1955). Social identification based influence (as described by the

second intervention and reflected in this dissertation's framework) impacts on attitude

generation. The only part of the behavioral intentions model that is challenged directly in

the implications of this dissertation is the treatment of social influence where social

influence is pulled outside of individual attitudes. The kind of social influence described

in this dissertation contests the view that a person forms their individual attitude without

ever considering social influence and that social influence can only modify the

relationship between attitudes and intentions. Such a conceptualization might be

considered a particularly narrow view of the importance of a social self. This is an





84


important point for the following reason: A model that defines all social influences as

lying outside of a person's individual attitude will only account for social influence

processes that are associated with the first example above. That is fine if the focus is on a

normative kind of intervention. However, such a conceptualization will be less useful to

a researcher whose primary goal is to understand the attitude formation process in order

to affect/modify/change an attitude that is based on something other than the evaluative

implications of motivational pressures to conform to some referent other. There may be

many instances in which it is not enough to know that people tend to superficially modify

their behavior in light of complying to some external contingency. Rather, in some

instances, a more effective behavioral change strategy might be to lead someone to

generate an attitude based on identification with a social identity. Therefore, if one has

an a priori theoretical framework that specifies when this type of social influence is likely

to come into play, then there may be instances where one may affect attitudes in a more

enduring fashion by a clearer understanding of how identification may help to form a

person's attitude in the first place.

Contributions

At any given point in time consumers have a subset of social identities that they

may adopt in order to guide their thoughts and actions. These social identities can be

very important bases for self-definition and in some instances, be the basis by which a

consumer forms her attitude. In fact, academic research has consistently shown that

categorizing consumers in terms of various social identities can lead to differences in

how that consumer is likely to behave. Intriguingly, marketing practitioners certainly

understand the value of targeting consumers, segmenting markets and positioning





85


products based on the evaluative implications of these social identities. However, the

field is in its infancy in understanding the precise nature of how consumers may generate

their attitude on the basis of identification with a social identity. Therefore, the current

framework presented here may be an important contribution to both attitude theory and

marketing practice.

Contribution to Attitude Theory. Consumer research has been heavily influenced

by multi-attribute conceptualizations of attitude, and rightly so. This very parsimonious

way to represent a consumer's evaluation of a product or brand is very useful in

determining which product features should be emphasized in marketing strategy.

However, some consumer decisions may not be based on a reasoned consideration of

product features. Sometimes, a consumer may feel connected with a brand, product or

behavior simply because it is "who they are." In fact, the notion that a consumer's sense

of who they are should relate to their purchase decisions has been an important idea

discussed by several scholars (e.g., Levy, 1959); whether in terms of overall

"congruency" between the self and a brand (Sirgy, 1982), in terms of precise roles that

the consumer wishes to enact (Solomon, 1983) or in terms of particular personality

associations embodied within the brand itself (Aaker, 1997). This extremely simple

premise is not new. Though a few researchers have acknowledged that identification

may lead a person to willingly adopt an attitudinal position to "maintain a desired

relationship and the self-definition that is anchored in the relationship" (Kelman, 1958,

page 62), and functional theorists have described social identity motives as reasons why a

person might hold a particular attitude, almost no research has systematically examined a

social identity based attitude formation process including identifying when and how it is





86

likely to occur. This research attempts to make two contributions to attitude theory: 1) to

provide a better understanding of the role of social identity as an alternative mechanism

leading to attitude formation and change, and 2) to provide a more precise understanding

of the moderating factors that are likely to determine when social identity will affect

consumer attitudes.

Contribution to Marketing Practice. In terms of marketing strategy, brands and

products are often created or positioned to embody a particular social identity oriented

lifestyle. Often it appears that marketers are hoping people will look upon such lifestyle

presentations favorably and connect them to the advertised brand. The model described

in this research should be a useful guide for developing various kinds of social identity

based marketing strategy in several relevant substantive domains. The model uses an

expansive, but precise definition of self (i.e., social identity). This allows marketers the

flexibility to work with different kinds of social identities in formulating a particular

strategy. The model should also contribute to marketing practice by identifying the

precise roles likely to be played by salience, self-importance, object relevance and

diagnosticity. The model introduced here and its subsequent validation in future research

programs should be useful to marketers who wish to formulate or evaluate a social

identity-based advertisement, positioning strategy, persuasive communication or market

segmentation scheme because the model specifies each of the key components of a

successful execution. Finally, the model presented in this dissertation research should

allow marketers to emphasize the most cost effective and easily alterable aspects of the

proposed process-given a particular social identity that the marketer chooses to work

with. In the future, it is hoped that the model can lead to improved insights--some






87


discussed here--for marketers who wish to connect consumers to their brands or products

through the use of social identification processes.








APPENDIX A
VISUAL IMAGERY STUDY




UNIVERSITY OF


FLORIDA
Department of Marketing
VISUAL IMAGERY STUDY:
PRETEST OF INTERNET TYPEFACE FORMAT(S)


INTRODUCTION:
As you probably well know, the Internet has revolutionized information transmission. A much greater percentage of
communication is taking place in "Cyber space."
This research project deals with visual imagery and the effectiveness of Internet communication formats. One kind of
imagery is connected to both the presentations of format and style. That is, not just what we say, but how we say it.
This is particularly important both in mass communication (e.g., Web page design) and specialized communications
(e.g., interpersonal emails, etc.)
One active area of research is the effectiveness of typeface used in electronic communications. For example, look at
the following types of presentation formats:
HAVE A NICE DAY HAVE A NICE DAY HAVE A NICE DAY

These presentation formats might convey different meanings. One may be perceived as more "warm & outgoing,"
another might be perceived as more "cold & impersonal."

Your task in this study will be to review a set of typeface formats. Then you will be asked to evaluate each typeface
format. Your responses are completely anonymous.



To begin, please provide us with the following Administrative information:

Last four digits of your student number:



Please turn to the Next page >










88






89


We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful
manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and
we varied it slightly to slow you down just a little bit.

Now, please evaluate the meaning of the word "friend" in terms of "how it comes across to you" in the particular type
face presented. To provide us with the most accurate immediate response, read the phrase each time and try to
rehearse it in your mind before providing your evaluations:

Evaluate the Type face on the following dimensions
provided in the box

THINK ABOUT THE WORD FRIEND Warm E 13 0 0 Q Impersonal
Memorable 013 1 3 0 Bland
Fun O C 130 U Boring
Common Q 0 0 U0 Unique

PONDER THE WORD FRIEND Warm Q 03 0 Q Impersonal
Memorable 1313 0 3 0 Bland
Fun 00003 0 ] Boring
Common 13 0 U 13 0 Unique

REFLECT ON THE WORD FPrEnK Warm 0 13 0 13 0 Impersonal
Memorable 13 13 300 Bland
Fun 73 00000 Boring
Common U0 0 U U Unique

CONSIDER THE WORD FRIEND Warm O13 0 Q Impersonal
Memorable 0 1 000 Bland
Fun 013 03 0 Boring
Common 0 0U 0 ] ] Unique

THINK ABOUT THE WORD FRIEND Warm 0 13 0 03 Impersonal
Memorable C O 0 0 00 Bland
Fun 0000013 0 Boring
Common 0Q 0 0 0 Unique
PONDER THE WORD FRIEND Warm 0 0 13 D0 Impersonal
Memorable O3 E3 Q ] Bland
Fun 00000 Boring
Common ] L3 [Q Q Q Unique

REFLECT ON THE WORD FRIEND Warm 13 0 D 13 Impersonal
Memorable O Q0 13 0 Bland
Fun 00000 Boring
Common 3 Q[ 13 Q 3 Unique
Please turn to the Next page >






90


We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful
manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and
we varied it slightly to slow you down just a little bit.

Now, please evaluate the meaning of the word "friend" in terms of "how it comes across to you" in the particular type
face presented. To provide us with the most accurate immediate response, read the phrase each time and try to
rehearse it in your mind before providing your evaluations:

Evaluate the Type face on the following dimensions
provided in the box

THINK ABOUT THE FRIEND Warm ]0 DO ] 0 Impersonal
IMPORTANCE OF BEING A Memorable 0 0 0 0 0 Bland
Fun Q 0 0 0 0 Boring
Common 0 0 0 0 0 Unique

CONSIDER THE VALUE OF FRIEND Warm 10 013 0 Impersonal
BEING A Memorable 0 0 0 00 Bland
Fun 0000 0 Boring
Common 0 0 Q 0 0 Unique

PONDER THE SIGNIFICANCE F UEJWD Warm 0 0 00 0 Impersonal
OF BEING A Memorable 0 [] O 0 0 Bland
Fun 0 000 Boring
Common E] 00 0 n Q Unique

THINK ABOUT THE VALUE OF FRIEND Warm [] 0 0 0 0 Impersonal
BEING A Memorable 0 10 0 0 0 Bland
Fun 00000 Boring
Common U 0 ] 0 U Unique

PONDER THE IMPORTANCE OF FRIEND Warm 0 0 0 0 0 Impersonal
BEING A Memorable 01 0 3 0 0 Bland
Fun 00000 Boring
Common 013 0 0 Q] 0 Unique

THINK ABOUTTHE FRIEND Warm 0 0 D 00 Impersonal
IMPORTANCE OF BEING A Memorable 0 0 0 003 Bland
Fun 00 0 0 Boring
Common 0] 0 0 Q0 Q Unique

CONSIDER THE VALUE OF FRIEND Warm 0 0 0 0 L0 Impersonal
BEING A Memorable 03 0 0 0 Bland
Fun 00000 Boring
Common 0O Q 0 Q0 Unique
Please turn to the Next page >




Full Text

PAGE 1

WHEN WHAT I THINK DEPENDS ON WHO I AM: THE ROLE OF SOCIAL IDENTITY IN CONSUMER ATTITUDE FORMATION By AMERICUS REED II A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2000

PAGE 2

Like an iron fist in a velvet glove, you don't know what I'm made of. Neil Peart

PAGE 3

ACKNOWLEDGMENTS The completion of this dissertation would not have been possible without the assistance of several key people. First, I would like to thank my mother, father and sister for their ongoing love and familial support. I would also like to thank my beautiful wife Karen—whom I love dearly—for her timely encouragement, love and patience to put up with me. Of course, my dissertation chair Joel Cohen has been both an inspiration and an irreplaceable mentor, patiently guiding this work with a steady hand, spending meticulous time and effort to facilitate my ongoing development as an aspiring scholar. In that regard, my committee members must also be acknowledged for their copious and insightfiil feedback: I thank Rich Lutz, Alan Sawyer, and Dolores Albarracin for their commitment to this project and especially Barry Schlenker, who is both an incredible human being and a scholar. A Special thanks go to Karl Aquino-long time friend, mentor and colleague-for taking me under his wing and exposing me to life in academia. I also appreciate the encouragement and intellectual stimulation provided in my undergraduate consumer behavior class taught by Lois Mohr, who was the first to introduce me to the interesting concepts and ideas in consumer research and who was the first to suggest the University of Florida as a good place to pursue my interests. Speaking of the University of Florida, I would be remiss if I did not acknowledge my friendship with two of the greatest people that I know: my life-long companions and colleagues Lisa Bolton and Kevin Bradford. I should also mention three key people in iii

PAGE 4

the department, Julie Ritter, Jennifer Maynard, and Cathy Koenig for being so kind in assisting me with all of the "devil in the details." I would like to thank all of my other colleagues at the University of Florida for the numerous get-togethers, informal discussions, round-robin basketball tournaments, fireside chats about research and life in general, and for an academic environment that was both intellectually stimulating and fun. Last but not least, thanks to Jonesy D. Cat—who does not really care about the "significance of a value added contribution to the literature," but only that his food dish remains full— for keeping the apartment animated and stress levels attenuated. iv

PAGE 5

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS iii ABSTRACT viii CHAPTERS 1 INTRODUCTION AND OVERVIEW 1 2. FACTORS THAT AFFECT A CONSUMER'S ATTITUDE 4 Direct Experience with the Object 4 Inferential Attitude Formation Based on Past Behavior 5 Categorization Based Attitude Formation 5 Analytical Attitude Construction 6 Personal Values 5 What About Social Identity? 7 3. THE DEFINITION OF SOCIAL IDENTITY 9 The Notion of Identity 9 Social Identification Processes 10 Sociological Models j 1 Social Identity Theory 12 Social Categorization Theory I3 Definition of Social Identity and the Current Research 14 4. SOCIAL IDENTITY BASED ATTITUDE FORMATION 16 Is the Social Identity Salient? I5 Is the Social Identity Self-important? I7 Is the Social Identity Object-Relevant? I9 Does the Social Identity Provide a Basis to Respond? 19 Summary of the Framework 20 5. EXPERIMENT(S) lA AND IB: THE INTERACTIVE ROLE OF SALIENCE AND OBJECT RELEVANCE 25 Overview of the Studies 25 Theoretical Premise of the Studies 25 V

PAGE 6

Pretest 26 Experiment la 32 Experiment lb 35 6. EXPERIMENT 2: THE ROLE OF SALIENCE, SELF-IMPORTANCE AND OBJECT RELEVANCE 47 Purpose of the Study 47 Theoretical Premise of the Study 47 Experiment 2 43 7. EXPERIMENT 3 : THE ROLE OF SELF-IMPORTANCE AND EVALUATIVE DIAGNOSTICITY 60 Purpose of the Study and Hypotheses to be Tested 60 Overview of the Experiment 61 Experiment 3 61 8. GENERAL DISCUSSION AND CONCLUSIONS 78 Contributions g4 APPENDICES A. VISUAL IMAGERY STUDY gg B. WORD SEARCH STUDY 9g C. CONSUMER SELF DESCRIPTION STUDY 100 D. NEW PRODUCT ASSESSMENT STUDY 101 E. HANDWRITING STUDY lOg F. THE PANAS MOOD SCALE II9 G. WEB PAGE DESIGN STUDY 120 H. EXPERIMENT IB: OMINBUS ANOVA OF EFFECTS OF IDENTITY AND RELEVANCE ON PURCHASE LIKELIHOOD OF FOCAL PRODUCT 128 L EXPERIMENT 1 B : OMINBUS ANOVA OF EFFECTS OF IDENTITY AND RELEVANCE ON PURCHASE LIKELIHOOD OF FILLER PRODUCTS nn vi

PAGE 7

J. EXPERIMENT 2: OMINBUS ANOVA OF EFFECTS OF SALIENCE 130 RELEVANCE AND SELF-IMPORTANCE ON PURCHASE LIKELIHOOD OF FOCAL PRODUCT K. EXPERIMENT 2: OMINBUS ANOVA OF EFFECTS OF SALIENCE RELEVANCE AND SELF-IMPORTANCE ON PURCHASE LIKELIHOOD OF FILLER PRODUCTS 131 L. EXPERIMENT 3: OMINBUS ANCOVA OF EFFECTS OF SALIENCE RELEVANCE AND SELF-IMPORTANCE ON EVALUATIVE DIMENSIONS 132 REFERENCES I39 BIOGRAPHICAL SKETCH I49 vii

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements of the Degree of Doctor of Philosophy WHEN WHAT I THINK DEPENDS ON WHO I AM: THE ROLE OF SOCIAL IDENTITY IN CONSUMER ATTITUDE FORMATION By Americus Reed II December 2000 Chairperson: Joel B. Cohen Major Department: Marketing Attitudes are viewed as a particularly important intervening variable between a consumer's awareness of a brand and the intention to purchase that brand. Since consumers do not buy brands they do not like, the importance of creating favorable attitudes toward products, brands and companies can hardly be over-emphasized. However, subsequent to the introduction of multi-attribute models in the late 1960s, practically all textbooks and journal articles relied on a componential view of consumer attitudes. These models define attitudes as a weighted assembly of beliefs about the features or benefits of the product. It is worth noting that practitioners-who are actually responsible for creating favorable attitudes toward their product offerings-appear to believe that consumers often form attitudes in a very different way. Consider the prevalence of social identity oriented advertising themes (e.g., a brand is held up to be the embodiment of a particular viii

PAGE 9

"lifestyle" associated with consumers grouped into some type of social category). Moreover, consumers' social identification often becomes the focus of market segmentation and product positioning strategies. Yet, a favorable attitude toward a socially symbolized brand that is created using this approach is not well described via a multi-attribute, feature/benefit-based process. hi some respects, the situation described above might be thought of as a "disconnect" between scholarly research on consumer attitudes and practitioners' frequently selected means of creating favorable attitudes. This research attempts to address this gap. It takes the view that there are multiple bases for attitude formation. Further, it is important to understand the precise conditions under which a particular basis for attitude formation is likely to lead to an attitude. Social identity is hypothesized to serve as a basis for attitude formation when a social identity is 1) salient, 2) self-important, 3) object relevant, and provides 4) an evaluative basis to respond. Three laboratory experiments examine the theoretical underpinnings of this assertion. Evidence from experiment 1(a) suggests that a social identity's salience can mediate consumers' momentary self-conceptions which when combined with object relevance (experiment lb) leads to more favorable consumer attitudes (when adopting a social identity would in fact provide a favorable basis to respond to the product). Results from experiment 2 demonstrate the moderating role of self-importance in forming an attitude based on social identification processes and data from experiment 3 suggest that a social identity's evaluative content is more likely to impact attitude formation when it is diagnostic for an evaluative response. The totality of the empirical results, a comparison of the dissertation framework to other frameworks ix

PAGE 10

that describe similar phenomena, and the concluding implications of the framework for attitude theory and marketing practice are then discussed.

PAGE 11

CHAPTER 1 INTRODUCTION AND OVERVIEW The "psychological tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor" (Eagly and Chaiken, 1993, page 1) is a critical intervening variable between a consumer's awareness of a brand and the decision to buy it. However, attitude research has been dominated by early multi-attribute frameworks developed by Fishbein and Ajzen (1975) and others, structural models emphasizing the cognitive components of attitude (c.f Lutz, 1975; Rosenberg and Hovland, 1960), information integration models and economic utility approaches (c.f Kahn and Meyer, 1991; Lynch, 1985). These models imply a computational process of attitude formation where in the most general sense, each of the attitude object's components is given a rating (i.e., on level of performance, strength of association) and a weight (whose meaning and measurement varies with the model). Of course there are situations where consumers might form attitudes in this way. For example, a consumer might purchase a car because she believes the car gets good gas mileage, is modestly priced and is also safe. Additionally, she thinks it is important for a car to have those product features. However, this conceptualization does not adequately describe other processes that might lead to the formation of a consumer attitude. For instance, marketing practitioners seem to believe that a consumer's social identification with a particular subculture, ethnic identity, family, gender, set of peers or even spokesperson, can be the basis for a favorable consumer attitude. This is often reflected 1

PAGE 12

in "life style" advertising and certainly in various product positioning and market segmentation strategies. This intuition demonstrated by marketing practitioners has not totally gone unnoticed in academic research. Various studies looking at the persuasiveness of spokespersons, (Deshpande and Stayman 1994) gender and ethnic differences in advertisements, (Meyers-Levy, 1988; Jaffe, 1991; Forehand and Deshpande, in press) food consumption, (Wooten, 1995; Stayman and Deshpande 1989; O' Guinn and Meyer, 1984; Wallendorf and Reilly, 1983; Hirschman, 1981) media usage, (Saegert, Hoover and Hilger, 1985) brand loyalty / organizational patronage, (Deshpande, Hoyer and Donthu, 1986) and even information search behavior have demonstrated that classifying consumers based on a particular social categorization can lead to differences in the descriptions of how consumers behave. But merely knowing that consumers linked to different social categories sometimes respond differently sheds very little light on when and how this is likely to occur. Despite the early work by functional theorists who pointed out that social identity matters in motivating people to hold certain attitudes, (Shavitt, 1990; Katz, 1960; Smith, Bruner and White, 1956)--numerous discussions by social scientists regarding reference group influence in terms of people's willingness and ability to "adopt a standpoint," (Turner, 1956)-and more recent interests in social identity and inter-group relations, (Tajfel, 1981; van Knippenberg and Ellemers, 1990) relatively little work has been done

PAGE 13

3 to examine a process by which social identification is hkely to lead to the formation of an attitude' (see Eagly and Chaiken 1993, page 660 on this point). This research attempts to build on prior work by incorporating social identity within an attitude framework i.e., outlining the conditions of when and how a consumer's social identification is likely to be the basis for attitude generation. The rest of this dissertation is organized as follows. In chapter 2, a brief review is given of the different factors that affect how a consumer might form an attitude. Chapter 2 concludes with a discussion of social identity as an under researched alternative basis by which a consumer may form an attitude. Chapter 3 is a discussion and review of the concept of "social identity," i.e., how it has been defined across past literatures and how the term social identity will be used in this dissertation. Chapter 4 is the theoretical background for this dissertation. It describes a proposed process by which a consumer may form an attitude based on social identification processes including the precise factors that are likely to impact the process. Chapters 5-7 discuss the overview, design and results of empirical work intended to test the theoretical assertions made in Chapter 4. Chapter 8 concludes with a general discussion of how the empirical results conducted in this dissertation relate to previous conceptual frameworks that describe similar phenomena and the implications of this research to attitude theory and marketing practice. It is always possible to redefine any factor or variable (e.g., social identification) as an attribute or benefit and then move it into the attitude portion of a multi-attribute model Each input is simply reduced to the status of a belief element. However, the model would not attempt to distinguish between these different inputs and would provide no insight into the process or when this type of influence is likely to occur.

PAGE 14

CHAPTER 2 FACTORS THAT AFFECT A CONSUMER'S ATTITUDE The consumer attitude literature is replete with various factors that are thought to impact a consumer's attitude toward an object or issue. These different factors may reflect different processes by a which a consumer may form an attitude and are briefly reviewed here to motivate and provide context for the subsequent dissertation work. Direct Experience with the Object A consumer might form an attitude on the basis of direct experience with the attitude object (Regan and Fazio, 1977; Songer-Nocks, 1976; Fazio and Zanna, 1978a, 1978b; Fazio, Chen, McDonel and Sherman, 1982) or by an imagined attempt to experience the attitude object (Keller and McGill, 1994; Pham, 1998; Schwartz, 1990). Physical or imagined interaction with the attitude object results in a confluence of cognitive, affective and perhaps behavioral information. This information can be used as inputs to object evaluation. For instance, consider a consumer at a Thai restaurant trying to decide on what dish to consume for dinner. S/he might actually try a sample of some particular dish on the menu and generate an attitude toward that meal based on direct experience. As another example, consider the husband and wife who are trying to decide on a particular locale for their impending vacation. They might imagine themselves in various situations (sipping drinks on a beach in Jamaica, cozying up to a fire in the snowcapped mountains of Colorado, etc.). By asking the question "how do I feel about it?," 4

PAGE 15

5 the couple can generate attitudes toward each alternative based on the unique affective information derived from "trying on the consumer episode" (Pham, 1998). Inferential Attitude Formation Based on Past Behavior The work by Bern (1967, 1972) on "self-perception" theory suggests that there may be instances in which a consumer will infer her attitude based on thinking about past behavior. For example, consider the consumer who has purchased a brand of dishwasher repeatedly. Her behavior might have become habitualized in that she no longer thinks about the purchase decision. She just routinely buys the brand. Yet, when queried as to her attitude toward the particular brand of dishwasher detergent, she indicates a favorable attitude because she infers one based on the implications of her consistent behavior. In these instances, attitudes are a function of some retrospective consideration of sequences of past behavioral choices. Categorization Based Attitude Formation The categorization and consumer learning literatures suggests that a person may generate an attitude by considering its similarity to other liked or disliked objects (Lingle and Ostrom, 1979; Fiske and Pavelchak, 1986; Gregan-Paxton and Roedder John 1997). In these instances evaluative information associated with an attitude object similar to the attitude object in question is brought to mind as a basis for generating an attitude toward the specific attitude object. So in the restaurant example, the consumer might already possess an attitude toward a similar meal (e.g., Chinese food), and based on an anchoring and adjustment type of categorization mechanism, come up with an attitude toward some Thai dish on the menu. Or in the other example, the couple may generate evaluative information toward the Jamaica trip by considering the similarity of an attitude toward a

PAGE 16

previous trip taken to the Bahamas last summer. In general, the process of classifying an object and/or thinking about how it relates to some analogous entity can bring to mind attitudes based on its perceived similarity to other objects for which the consumer holds preexisting attitudes. Analytical Attitude Construction Of course, much of the research in consumer behavior and attitude theory suggests that an attitude may be analytically constructed. In other words, a process whereby some set of cognitions and beliefs weighted by their evaluations leads to an attitude and subsequent behavioral tendency (Fishbein and Ajzen, 1975; Ajzen and Fishbein, 1980). The combinatorial attitude models imply that evaluative information is generated when a subset of the universe of beliefs about the object becomes salient. Depending upon what beliefs become accessible, the attitude can be computed vis a vis the assessed implications of those beliefs. For example, the person in the restaurant can generate an attitude toward the Thai dish by considering specific features of the attitude object (e.g., this dish contains fish, cilantro, a bit of sodium, curry etc.) and the evaluative implications of those features (I like fish and cilantro, but I am watching my sodium intake, etc.) Personal Values An attitude may also come from one's personal belief system (Rokeach, 1968, 1973, 1980). In these instances, evaluative information comes to mind as a ftmction of referencing one's personal values or beliefs about the consequences of an emergent attimde object for one's central value system (Stem, Deitz, Kalof and Guagnano, 1995). For example, a consumer might possess a higher order value structure that leads to the

PAGE 17

7 belief that protecting the environment is an important act because it is "the right thing to do." Therefore, attitudes can be generated by thinking about the convictions that are Hnked to central and basic values espoused by the consumer (Rosenberg, 1956). What About Social Identity? In classic attitude research, a number of researchers focused on the motivational antecedents of attitudes. Ironically, social identity was identified as one important factor as to why some people hold certain opinions (Katz, 1960; Smith, Bruner and White, 1956). The term social adjustment function which was later termed more generally as the "social identity function," (Shavitt, 1990, 1989) denoted how some attitudes can mediate self-other relations and establish one's social identity: the function of social adjustment served by holding an opinion is at once more subtle and more complex. For it is by holding certain views that one identifies with, or indeed differentiates oneself from various 'reference groups' those groups in terms of whose standards the individual judges himself and with which he identifies with or feels kinship, (page 42) To the extent that an attitude embodies a unique social classification or reference group, the attitude serves a social idenfity fiinction (Shavitt, 1990). For example, an attitude regarding the American flag may symbolize the nations' values and represent a social classification with which consumers may identify with. The functional approach is a direct lead in to the current research because the fiinctional approach (e.g., identifying and measuring different motivational underpinnings of attitude) made the point that social identity may motivate people to hold certain attitudes. This work established why social identity matters in attitude formation. But if consumers hold certain attitudes in order to establish or maintain a social identity, then it

PAGE 18

s makes sense to examine a process whereby consumers form an attitude based on social identity. The current research builds on the work by functional theorists by spelling out when and how the motivational factor of social identity is likely to lead to attitude formation, including the precise factors that are likely to lead a person to rely on social identification as the basis for their attitude. However, in order to succinctly talk about a process that involves "social identity," as its driving mechanism, one must first establish a clear and concise definition. In the next chapter, social identity is briefly reviewed in order to set forth a consistent definition of social identity that will be used throughout this dissertation.

PAGE 19

CHAPTER 3 THE DEFINITION OF SOCIAL IDENTITY The Notion of Identity The concept of "identity" was bom within the psychoanalytic tradition. NeoFreudian psychoanalyst Erik Erikson (1950, 1956a, 1956b) investigated the struggle of children to synthesize the continual bodily changes of youth into a meaningful and integrated personality structure. He introduced the term "ego identity" into the literature as "the awareness of the fact that there is a self sameness and continuity to the ego's synthesizing methods and that these methods are effective in safeguarding the sameness and continuity of one's meaning for others" (1946 page 23). In his later work, Erikson (1974, 1978) goes on to write about issues related to the development of identity over the life course. The idea of identity seemed to be a useful concept that cut across research disciplines. For example, in the early 1960s, two influential statemems on identity were introduced within the symbolic interaction tradition. "Appearances and the self," by Gregory P. Stone (1990), offered one of the first concise definition of identity widely used within sociology. Identity, according to Stone, is a meaning that a self acquires when "situated." In other words, a person's identity is cast in the shape of a social object by the acknowledgment of his participation or membership in social relations. In this regard, identity is dynamic in that it is continuously changing. It is "intrinsically associated with all the joinings and departures of social life (page 94)." This dynamic 9

PAGE 20

10 nature of identity lead theorists to conceptualize how the search for identity could lead to conflict. Heinz Lichtenstein (1977) published a compilation of work on identity that was organized under the idea of identity as an existential "dilemma." The premise here is that conflict associated with different identities emerged from a simultaneously and irreducibly private and public reality. Within the symbolic interaction tradition, Goffman (1959, 1961a) was perhaps one of the first scholars to use the term "social identity" as distinguished from personal identity and ego identity. His dramaturgical analysis begins by describing the entrance of a stranger and the process of interpreting appearances that "enable us to anticipate his category and attributes," i.e., his social identity. Other theorists like Berger and Luckmann (1966, 1971), conceptualized identity as a social meaning constructed like other meanings, but with the uniquely existential dimension of being anchored in an individual's body (Weigert, 1983). In a more positivist framework. Burke (1980) conceptualized identity as the subjective component of a role. In his analysis, interrelated multiple role identities constitute an individual's self-concept. Social Identification Processes To socially identify implies a psychological connection with some other person or group (Deaux 1997). This notion of identification has developed in a wide variety of different research traditions. For example, psychoanalytic models define identification as a process by which the individual develops ties with another person or group. Freud (1955, 1985) stated that idenfification can "arise with any new percepUon of a common quality shared with some other person" (page 137). Early in life, identification is considered an unconscious process of imitating referent others who serve as models for

PAGE 21

11 beliefs, values and behaviors. In later stages of maturation, the identification process involves conscious choice and discrimination among possible identities (Higgins, Loeb and Moretti, 1995). For example, Kelman's (1958, 1961, 1974b) typology of social influence uses the term 'identification' to describe situations where an individual willingly adopts an attitudinal position recommended by some referent other. The motivation for adopting the position is that it establishes or maintains a positive selfdefining relationship with the referent other. Conceptually similar to French and Raven's (1959) idea of referent power, this type of social relationship may take the form of the target desiring to adopt the role of the referent other. It may also take the form of a reciprocal role relationship in which the target desires to participate in activities vis-a-vis the referent other (Eagly and Chaiken, 1993).' Identification processes are invoked only in the context of the role relationship with the referent other that is the basis of the identification. Therefore, once the relationship no longer becomes important to the individual, identificafion ceases to exist. Foote (1951) expanded this analysis beyond dyads when he interpreted human motivation as a consequence of identification with an important reference group. He conceptualized identification as the "appropriation of and commitment to a particular identity or series of identities" (page 14). Sociological Models Whereas psychological development theories of identification focus on members of the family unit as identification targets, sociological models emphasize a broader set of • An identification process (Kelman, 1958) is very different from a compliance based process m that it is not dependent on the referent other's direct surveillance of the individual's behavior. Neither is it contingent on the mediation of punishments associated with not adopting the referent other's particular attitudinal position However Kelman's notion of identification is a function of the salience of the self-defining relationship with the referent other.

PAGE 22

12 social categories that are subsets of the social system (Stryker and Statham, 1985). These categories are large and the members are not necessarily known to one another; yet every occupant of a particular role category shares behavior and expectations with other occupants of the category. In this way, roles exist in the broader social system and the outcome of conscious role selection is a defined position relative to others. These models conceptualize identification as an interdependent process, whereby the relationships are carried out via cooperating participants (Deaux, 1997). Symbolic interaction models (Mead, 1934) including Stryker's (1980) identity theory and the role-identity model of McCall and Simmons (1978) offer a similar perspecfive on social identification. These models posit that social interaction operates through the enactment of roles that the individual chooses to play (Deaux, 1997). However, these theories not only take into account standard roles like mother and child (Stryker and Statham, 1985; Thoits and Virshup, 1995), but also identities not so readily defined in terms of specific systems, (e.g., mentor and student). Social Identity Theory In another developing research stream, social psychology has expanded its conceptualization of the self concept to capture aspects of inter-group and intra-group processes that are linked to important, collective social identities (Hogg, 1996; Hogg and Abrams, 1988; Tajfel, 1978, 1982; Tajfel and Turner, 1979). Tajfel's (1959, 1969b) work has generated empirical findings suggesting that an individual can defme him or herself in various ways. Accordingly, membership within a self-inclusive social category provides a category-congruent self-definition that constitutes an element of the self concept (see also Turner, 1982; Turner and Giles, 1981). In an elaboration of social

PAGE 23

13 identity theory Tajfel (1978, 1981; Tajfel and Turner, 1979) presumes a process of self definition impacted by specific membership groups. The essence of self-definition is "that part of the individual's self concept which derives from knowledge of his membership of a social group (or groups) together with the value and emotional significance attached to that membership" (Tajfel, 1981, page 255). Social identity theory explicitly recognizes the importance of considering both basic motivational and cognitive processes in order to account for inter-group perceptions and behavior. As Tajfel and Turner (1979) put it, "social categorization entails much more than cognitive classification of events, objects or people. It is a process impregnated by values, culture and social representations" (page 114). Social identity theory was originally developed to capture aspects of inter-group categorization and group dynamics associated with large scale social identities such as races, nations and so forth. Social Categorization Theory Later, Turner (1985) and his colleagues developed social categorization theory to conceptualize aspects of intra-group dynamics related to more minimal groups. These groups are primarily characterized by frequent face to face interaction among all members. Self-categorization essentially expands the operation of the categorization process as the cognitive basis of intra-group behavior. This theory suggests that people can categorize themselves and others at a number of different levels of abstraction that can define one's social identity. For self-categorization theory, the depersonalization of self-perception is the basic process underlying group phenomena.'^ This includes various The term depersonalization is not meant to arouse a negative connotation. It does not contain any of the implications of 'dehumanization' or de-individuation, but simply refers to a contextual change in the level of identity and not a loss of identity.

PAGE 24

14 elements of collective behavior, shared group attitudes and mutual influence processes (Turner 1985, pages 99-100). Therefore, self categorizations are defined as cognitive groupings of oneself and some class of stimuli as the same, in contrast to other classes of stimuli that are different (Turner, Hogg, Oaks, Reicher and Wetherell, 1987). The cognitive aspects of identification can be considered from two perspectives: The process of categorization itself, and an analysis of the beliefs associated with a self designated category (Deaux, 1997). Turner (1982, 1984) increases the weight given to cognitive aspects of identification. He frames his basic premise as "a cognitive elaboration of the nature of social identity as a higher order level of abstraction in the perception of self and others" (1987 page 42). The definition of social identity by Tajfel makes it possible to conceptualize a social identity without reference to an out-group. In contrast, however; Turner's definition requires a we-they distinction (Deaux, 1997). In any event, social identity and self-categorization theory emphasize the independence and in the case of social categorization, the possibility of competition between in group and out groups in the identification process. Definition of Social Identity in the Current Research It appears clear that throughout the life course, socialization within a culture causes a person to become aware of an infinite number of social categories in the external environment. Some of these potential bases for self-definition are more permanent (e.g., mother, daughter, friend, etc.) while others may be more transitory (e.g., Republican, athlete, graduate student etc.) This research starts with the idea that consumers may perceive themselves in terms of these various levels of abstraction (Tajfel, 1959; Tajfel and Turner, 1979; Turner and Oakes, 1986) and at any given point in fime will have

PAGE 25

15 available a subset of social categories that can become a part of their "working" or "spontaneous" self-concept (Markus and Kunda, 1986; McGuire, McGuire and Winton, 1979). These social categories are "cognitive structures" whose meaning comes about over time through socialization and many of the other processes described in this chapter. A consumer may adopt a social category as one of her social identities in order to think about various actions or judgments. For clarity of exposition, in this research, the term "social category" refers to the infinite number of potential social constructions that may come from culture, society, marketers, peer groups, etc. The term "social identity" refers to the actuated perspective or frame of reference that a consumer possesses as part of the repertoire of who they are or want to appear to be.^ The next chapter describes a process of how an attitude may come to be formed on the basis of "social identity." The current research does not seek to determine the different social identities available to a person. Instead, it takes the view that for any social identity, a specified set of factors determines whether it is likely to generate an attitude. The emphasis is on those factors. A social identity needs a social referent, but that could be a group, an abstracted ideal, an individual--or any social construction. For instance, for a child, a comic book hero may well be a social identity. As long as the child attempts to see the world through the eyes of that entity, he has adopted that social identity to make some type of assessment.

PAGE 26

CHAPTER 4 SOCIAL IDENTITY BASED ATTITUDE FORMATION ^ Consider the social identity of "athlete." What does it mean to "be" an athlete? Thinking about this social identity might trigger a mental representation or distinct picture of what an athlete thinks (attitudes) and does (behaviors). A consumer might adopt this social identity and use its associated evaluative content as the basis to form an attitude (e.g., attitude toward Nike sports shoes or attitudes toward exercise). This is likely to result in a collectively anchored attitude that is formed via identification processes (Kelman, 1958) and is held, expressed or used as a guide for behavior in order to establish, maintain or even communicate a social identity to others (Shavitt, 1990). This chapter describes this process and the key factors that increase the likelihood of its occurrence. A consumer's social identity is more likely to affect her attitude when a social identity is 1) salient, 2) self-important 3) relevant to the object and 4) provides a basis to evaluatively respond. The framework is graphically pictured in figure 1. To motivate the forthcoming empirical work, each of the moderating factors is now briefly described. Is the Social Identity Salient? Social categories are internal mental representations that can become a basic part of how consumers view themselves. For example, a consumer may see herself as a "democrat" a "professor," "tomboy" or "working mother." However, no matter how 16

PAGE 27

17" extensive is knowledge about them, social identities can have little impact on consumer attitudes and behaviors unless social identity information is accessed: Postulate 1 : If a social identity is salient, then there is an increased likelihood that the evaluative content linked to that social identity will impact the formation of a consumer attitude. This model is using the term salience to refer simply to the extent that a social identity is an "activated" conceptual structure in the consumer's working self-concept. If a social identity is salient, it can bring to mind attitudes and behaviors consistent with the social identity (e.g., the social identity "Democrat" might be linked to perceptions of favorableness toward government sponsored social programs). Consider an African American consumer in a grocery store foil of Caucasian consumers. His ethnic identity might be salient. If the same African American consumer were in a grocery store foil of African American women his gender (male) identity might be more salient (see McGuire et al., 1979). The salience of his social identity might trigger attitudes usefol for consumer decision-making. For example, if standing in front of the grocery store magazine rack, the African American consumer might be more likely to peruse Ebony Magazine in the first case and GQ in the second. In these cases, the consumer perceives a stronger link between the social identity (e.g., ethnic identity vs. gender identity) and the object (Ebony vs. GQ). According to the framework, salience of a social identity moderates the relationship between evaluative content linked to a social identity and attitudes generated on the basis of it (see Figure 1). Is the Social Identity Self-important? Not all social identities are alike. Consider the following example. Two consumers Alan and Rich both consider themselves "athletes." Alan is a former high

PAGE 28

18 school track star who has moved on to other things. He is now a weekend-warrior-type, playing the occasional game of tennis to stay in shape. In general, he tries to eat healthy. Rich on the other hand is also a former high school track star, but is a former Olympic silver medal winner in the 100 meters. He is a fierce competitor who anxiously awaits the opportunity to reclaim his prior days of glory. Both of these consumers probably possess "athlete" as part of their sense of who they are. But because of past experience and future aspirations. Rich's identity as an athlete might carry more personal consumer value to him. The fact that this identity is much more engulfing to Rich may lead to a higher likelihood that many of Rich's attitudes will be based on the "athlete" aspect of his social identity. Additionally, holding all else constant, Rich is probably more likely to be more favorable toward and object (e.g., Nike brand shoes vs. Keds) that is linked to his athlete identity. Therefore, even if a social identity is salient, it might not be the basis by which a consumer attitude is formed. The consumer must identify with that social identity (i.e. the social identity must be selfimportant): Postulate 2: As a social identity becomes more self-important, then there is an increased likelihood that the evaluative content linked to that social identity will impact the formation of a consumer attitude. A consumer may be drawn to a social identity for various motivations. At one extreme, the self-importance of a social identity may manifest as an impression motivated and temporary public standpoint (Schlenker, 1985). At another extreme, the social identity may serve as a "phenomenological lens" that deeply engulfs the individual as a powerful basis for self-definition (as it does for Rich in the example above).

PAGE 29

19 Is the Social Identity Object-Relevant? Even if a social identity is salient and self-important to the consumer, it might not be a basis for attitude formation. The social identity must be relevant to the object that is to be evaluated. The object-relevance of a social identity increases the likelihood that the attitude object will be thought of in terms of the social identity: Postulates: As a social identity becomes more object relevant, then there is an increased likelihood that the evaluative content linked to that social identity will impact the formation of a consumer attitude. For example, a consumer who perceives herself as a "working mother" may be more favorable to an automobile that emphasizes safety and practicality. As another example, an attitude toward affirmative action may be more relevant to a consumer's ethnic social identity if the person perceives affirmative action as a means of achieving an important identity based goal such as insuring equitable employment (Kravitz, 1995). The perception of relevance may be momentarily heightened if the person is concerned about his or her immediate, upcoming opportunities on the job market. Therefore, active goals and current concerns may impact the relevance of the attitude object to the social identity and thereby motivate the consumer to consider the attitude object in relationship to some social identity. Does the Social Identity Provide a Basis to Respond ? Even if a social identity is salient, self-important and object relevant, it might not be the basis by which a consumer forms an attitude. This model uses the term "evaluatively diagnostic" to refer to the extent to which the evaluative content of the social identity has sufficient clarity and specificity to inform the consumer's evaluation of

PAGE 30

20 the object and guide a behavioral response. In other words, does the social identity point a person in a meaningful direction? Postulate 4: If a social identity is evaluatively diagnostic, there is an increased likelihood that the evaluative content linked to that social identity will impact the formation of a consumer attitude. For example, suppose a consumer is evaluating a consideration set of shoe brands. Hence, an attitude is called for. Further suppose that the person's social identity (e.g., "Urban Youth") is highly salient at the time an attitude is generated (e.g., the person is watching a program on Black Entertainment Television). Even if the consumer is evaluating a set of brands (e.g., sketcher's, FUBU) that is clearly relevant to her selfimportant social identity, she may not be clear on which brand best embodies her "Urban Youth" identity. In the current terminology, her social identity in this case would be evaluatively non-diagnostic relative to evaluating the brands of shoes (c.f., Feldman and Lynch, 1988). Her social identity as an urban youth and the fact that there is no clear identity related norm (Kallgren, Reno and Cialdini, 2000) provides her with an inadequate basis to respond to the object (i.e., in this example, choose among the two brands). Summary of the Framework The framework of this dissertation research suggests that a consumer can think about herself in terms of various social identities that can connect her to a brand, product or behavior. These social identities may lead to the formation of an attitude, but only under specific conditions. There are four key moderating factors that determine if social identity, (defined as the evaluative content of a social identity) will impact consumer attitude formation. If when an attitude is called for, a social identity is salient (the

PAGE 31

21 consumer is thinking about herself as being that kind of person), self-important (the person strongly identifies with that social identity), relevant to the object to be evaluated (is functionally linked to the object), and provides a basis to respond (gives meaningfiil evaluative direction for the consumer), then there is an increased likelihood that an attitude will be generated on the basis of social identification processes. Intended contribution of the framework The framework integrates past research traditions in social identity theory and functional attitude theory to develop a theory of social identity based attitude formation. For example, recent theorizing in social psychology assumes that much of the working self-concept's content is socially embedded. Specific social self-categorizations are brought into play in the social field (Turner, Hogg, Oakes, Reicher and Wetherell, 1987). These social identities become the basis of perception conduct and behavior (e.g., Giles and Johnson, 1987; Hinkle and Brown, 1990; Hogg, 1992). Therefore, this framework argues that the salience of a social identity is important in bringing to mind evaluative content (i.e., attitudes and behaviors) that is linked to that social identity. However, the salience of a social identity is a necessary but not sufficient condition for social identity based attitude formation. A consumer must also be drawn to a social identity as a basis for self-definition. The more a consumer identifies with a social identity, the more likely that social identity will impact her attitudes. However, a salient and self-important social identity will not lead to an attitude unless it is linked to the object that is being evaluated. The object must be "relevant" to the social identity in that an attitude generated on the basis of the social identity serves a psychological function of self-expressiveness (Katz, 1960; Smith, Bruner and White, 1956) or what has been more recently called a "social identity"

PAGE 32

22 ftinction (Shavitt, 1990). Consequently, the object relevance of the social identity moderates the impact of the social identity's evaluative content on forming an attitude toward the object. Lastly, the social identity must provide a clear basis to respond to the object. Even if a salient, self-important and object relevant identity is not diagnostic (Feldman and Lynch, 1988) to a judgment, it is unlikely that social identity will be the basis for attitude formation. Each of the above theoretical assertions have been discussed in past literatures in different theoretical domains. But taken together, the current model attempts to make a contribution by integrating these conceptual underpinnings in a way that describes a process of how and when social identity is likely to impact consumer attitude formation. Importance of the framework. Past research has often defined the process of attitude formation as specific belief elements and their evaluative implications. The result has been a focus on features / aspects of the attitude object and how these attributes should be measured and represented in some algebraic form. Consumers might form attitudes in this way. However, there might be situations where the driving mechanism behind an attitude has less to do with product features or benefits and more to do with social identity. If being of some social identity is salient, and a consumer is thinking about herself in relation to that particular social identity, then it is possible for that frame of reference to dominate how she interprets the world around her. If one assumes that it is important to distinguish the different bases by which a consumer forms an attitude in order to understand why consumers react to a product offering in a certain way or to identify the most effective ways to modify a consumer's response, then conceptualizing a social identity based process of attitude formation and defining the contributing factors

PAGE 33

23 that impact the process should lead to important implications for attitude theory, marketing practice and public policy.

PAGE 34

24 Consumer Needs an — j attimde EVALUATIVE IMPLICATIONS OF A SOCIAL IDENTITY SALIENCE SELFIMPORTANCE MODERATING FACTORS OBJECT RELEVANCE DIRECTION FACTORS THAT AFFECT ATTITUDE: 1) Direct Experience / Imagined Experience 2) Similarity to Other Objects 3) Past Behavior 4) Object Features 5) Personal Values ATTITUDE Figure 1. Conceptual framework of social identity based attitude formation.

PAGE 35

CHAPTERS EXPERIMENTS 1 A AND IB: THE INTERACTIVE ROLE OF SALIENCE AND OBJECT RELEVANCE Overview of the Studies This chapter describes the results from a pretest that assesses the effectiveness of a salience induction manipulation. The salience manipulation was used in two experiments~also described in this chapter-that carefully examine attitude formation when a social identity is both salient and relevant to an attitude object. Theoretical Premise of the Studies Consumers are exposed daily to various marketing stimuli and consumer situations. Any of these can heighten a particular basis for social identification. For example, a female consumer in a store full of males might be likely to have her "gender" identity salient (c.f McGuire, McGuire and Winton 1979; McGuire, McGuire, Child and Fujioka 1978). Or a commercial that shows a concentration of ethnically homogenous individuals may make ethnic identity salient to consumers who view that ad and share that identity (Forehand and Deshpande, in press). Marketers can in a sense, alter how a consumer momentarily thinks about herself by making a particular social identity salient. In experiment la, the following hypothesis will be tested: Hl(a): Heightening the salience of an identity increases the likelihood that a consumer will see herself in terms of that identity. However, according to the framework, the salience of a social identity alone is not enough to ensure that an attitude will be based on the evaluative content of a social 25

PAGE 36

26 identity. If the social identity is salient and relevant to the brand or product, then social identification is more likely to connect the consumer to the brand or product. This increases the likelihood of social identity leading to the formation of a consumer attitude. In experiment lb, the following hypothesis will be tested: Hl(b): Heightening the salience of an identity leads to higher (lower) evaluations of an identity relevant (identity irrelevant) product when adopting such an identity would provide a favorable basis to respond to the product. When a social identity is made salient, it is at least temporarily a part of a person's "working" or "spontaneous" self-concept (Markus and Kunda 1986; McGuire and Padawer-Singer 1978). This heightened salience combined with object relevance (i.e., when the particular social identity is relevant to the stimulus to be evaluated) should increase the likelihood that the object will be thought of in terms of the particular social identity. This should increase the likelihood of attitude generation based on social identity. Pretest Purpose and overview of pretest The subsequently described pretest examines the effectiveness of the salience manipulation used in experiments la and lb. In the pretest, one of two different social identities was made salient between subjects. Subjects had these social identities made salient in one of two different ways. Subjects then participated in an ostensibly unrelated task intended to assess the effectiveness of the salience induction technique. More details follow.

PAGE 37

27 Procedure In the first task, subjects participated in a "visual imagery study" that involved evaluating "Internet typeface formats" for use in electronic communications (e.g., Web page designs). Subjects evaluated different multiple typeface formats on several dimensions (e.g., warm, memorable, fun and coiimion) in terms of how a "random" printed word came across to them. Unbeknownst to the subjects, the task made either their friend (family) social identity salient by requiring the subjects to evaluate the word "friend" ("son" / "daughter") seven different times in seven different typefaces'. This is a pure salience manipulation intended to prime the particular social identity of interest. Additionally, two different types of salience were induced between subjects. These two different types of salience have no substantive significance in the theoretical framework, rather, these two different types of salience were tested to assess possible differences in the extent to which a particular depth of processing was needed to invoke the salience of the social identity in question. Therefore, in the concept salience condition, subjects evaluated the focal word (friend, son or daughter) in no particular evaluative context. In other words, the preamble to the focal word was evaluatively neutral (e.g., think about the word "friend."). In the idea salience condition, the evaluative context was manipulated in the lead in (e.g., think about the importance of being a "friend.") priming not only the social identity, but also a valenced idea linked to that social identity. A portion of an example idea salience manipulation is shown in This pretest involves two social identities that have been established in prior work. They are the "friend" and "family" social identities that are associated with peer versus familial role contexts (cf Baldwin and Holmes, 1987; Solomon, 1983). Because males and females participated in the study, the family social identity was further broken down into the precise social identities of "son" and "daughter" depending on the gender of the subject.

PAGE 38

28 figure T. Subjects rated the type faces and to bolster the cover story, answered several questions regarding their experience with Internet communications (e.g., time spent surfing web sites, etc.) Immediately afterwards, subjects participated in an ostensibly unrelated exercise whose cover story concerned assessing familiarity with "word search games". Subjects were instructed to peruse a word search grid on a subsequent page that contained eighteen words (including the focal word "friend", "son" and "daughter") amidst randomly arrayed letters in a 20 x 20 word search grid. Subjects were given approximately ten minutes to find as many words as possible.'' Subjects were then debriefed and dismissed. Predictions This pretest was designed to assess the extent to which the salience manipulation in the visual imagery study was able to heighten the salience of a particular social identity. In a figure ground sense, it was expected that heightening the salience of either the "friend," or "family" ("son" / "daughter") social identity would increase the likelihood that subjects would be able to locate the focal word (friend, son or daughter) associated with that salient social identity. This might be particularly the case if the salience induction contained an evaluatively valenced context (idea salience conditions) because of the additional depth of processing presumably evoked by the idea salience conditions. ^ The full stimulus materials for these salience manipulafions are included in Appendix A at the end of this dissertation. Subjects were fiirther told that every time they found a word, to 1) circle the word in the grid and then to 2) write down the word in one of eighteen spaces provided. Subjects were told to stop after ten minutes. The instructions encouraged subjects not to cheat. Also, the experimenter verbally indicated that the exercise was intended to be fun and was not a test of their abilities in any way. Subjects were also told that the data would be useful only if instructions were carefully followed.

PAGE 39

29 Experimental design Subjects were assigned to a 3 (focal word that was evaluated; friend, son, daughter) x 2 (type of salience: concept vs. idea salience) between subjects factorial.'* The dependent measure of the pretest was the total number of instances in which the word "friend," "son" or "daughter" was identified (circled) in the word search grid. Results One hundred and ten (N=110) subjects from a southeastern university participated for class credit and were assigned to experimental treatments. For each social identity made salient, the percentage was calculated of instances in which subjects identified any of the focal words "friend," "son" or "daughter." These results (collapsed across type of salience induced conditions) are presented in figure 3. A logistic regression was performed defining the dependent variable as whether or not a subject identified one of the focal words. The type of salience (concept vs. idea) and which social identity was salient (friend, vs. family (son / daughter)) were the explanatory variables in a model including main effects and all possible higher order interactions. An analysis examining whether or not subjects identified the word "fi-iend" in the word search grid produced the following results. There was a main effect of type of salience (^^^=4.06, p=0.0439), indicating that the percentage m which subjects picked out the focal word "friend" depended on which type of salience was induced. Collapsing across which social identity was salient showed that 22% vs. 39% of the subjects found Subjects in the "son" conditions were all male, subjects in the "daughter" conditions were all female. However, the son vs. daughter distinction in this analysis has no substantive significance as the theory has no a priori predictions as to differences within the "family" social identity conditions. However, the results will be described in terms of all three different social identities.

PAGE 40

30 the focal word "friend" in the concept and idea saUence conditions, respectively. When subjects in the first task examined any of the three focal words in an evaluatively valenced context, they identified the focal word "friend" more often in the subsequent word search grid. All other effects were not significant. An analysis examined whether or not subjects identified the word "son" in the word search grid. A main effect of which focal word was evaluated was found (X^==6.62, p=0.0366). The percentage in which subjects found the focal word "son" in the word search grid depended on which social identity was salient (focal word was evaluated). In a follow up analysis, a planned contrast showed that more subjects found the focal word "son" when son was salient (57%) as compared to when "friend" was salient (29%) or when "daughter" was salient (48%), {X^=3A9, p=0.0616). No other effects were significant. Finally, an analysis examined whether or not subjects identified the word "daughter" in the word search grid. A marginal main effect of type of which focal word was evaluated was found (X^=5.47, p=0.0651). The percentage in which subjects found the focal word "daughter" in the word search grid depended on which social identity was salient (focal word was evaluated). A planned contrast showed that more subjects found the focal word "daughter" when daughter was salient (39%) as compared to both when "friend" was salient (19%) and when "son" was salient (17%), (X^^5A9, p-0.0194). Again, no other effects were significant. Discussion The results of this pretest are consistent with the idea that heightening the salience of a social identity (having subjects rehearse the focal word) made the focal word stand out. In other words, the salience manipulations increased the

PAGE 41

31 likelihood that subjects would locate the focal word arrayed in the subsequent word search grid. These effects were found for the focal words associated with the "family" social identity (i.e., "sons" and "daughter") but not for the "friend" social identity.^ Heightening the salience of "friend" had no significant effect on subjects finding the focal word "friend." Given the subtlety of the salience manipulation and that subjects were college students participating in a task located in their school environment, it might have been more difficult to heighten the salience of their "friend" social identity because it may already be salient, minimizing the impact of the "friend" salience manipulation on subjects ability to identify "friend" in the word search grid. In contrast, in the aforementioned experimental situation, it might be easier to heighten subjects' familial ("son" and "daughter") social identity because subjects were certainly in a context in which that identity might not likely be already activated (c.f Bern, 1980). It also appears that the concept and idea salience conditions had no impact on the likelihood of subjects identifying the focal words "son" and "daughter." However, the concept and idea salience inductions produced differential effects on the extent to which the focal word "friend" was found. When salience of "friend," "son" or "daughter" included an evaluatively valenced context, subjects were more likely to spot the word friend. Although purely speculative, this could be the result of the fact that when subjects rehearsed the word "son" and "daughter" in an evaluatively valenced context, (e.g., think ^ This pretest is based on the assumption that the more salient a focal word, the more likely it will be identified in the word search grid. Of course, there are other stimulus factors (e.g., word usage frequency, number of letters in the focal word, where the word is arrayed in the grid, etc.) that may increase or decrease subjects ability to identify a focal word. Hence, the analysis presented contrasts across a particular focal word.

PAGE 42

32 about the importance of being a daughter) they were more likely to bring to mind associations that also tapped the "friend" identity. For example, consider a person who is so close to her/his parents that s/he sees them as friends rather than in the role of parent vs. daughter/son. Nonetheless, the overall results of this pretest do suggest that the visual task manipulations were particularly effective at heightening the salience of the "family" (son and daughter) social identity. Experiment la Purpose and overview of Experiment 1 a Experiment la tests hypothesis la by investigating the impact of salient social identities on altering consumers' momentary sense of self-defmition. Exactly as in the pretest, one of two different social identities (friend vs. family) was again made salient between subjects. As in the pretest, subjects had these social identities made salient in one of two different ways. Subjects then participated in an ostensibly unrelated task intended to assess the extent to which the salience induction manipulations shifted subjects' spontaneous self-concepts. More details follow. Procedure In the first task, subjects participated in the same visual imagery study described in the pretest. Unbeknownst to the subjects, the task made either their friend (family) social identity salient by requiring the subjects to evaluate the word "friend" ("son" / "daughter") seven different times in the seven different typefaces. The two different types of salience used in the pretest were also used in experiment la^. ^These two manipulations are included to again asses the level of processing associated with making a particular social identity salient, and to shed possible light on the results from the pretest that suggested differential effects of the concept vs. idea salience manipulations on locating the focal word "friend."

PAGE 43

33 Immediately after completing the visual imagery study, subjects completed an ostensibly unrelated consumer "self-description study" (c.f. Mcguire et al., 1979) where subjects were asked to "please take a minute to describe yourself and to provide as much detail as you wish." This open ended probe was the only instruction given. Subjects completed the one page questionnaire and were debriefed and dismissed. This questionnaire is included in Appendix C. Predictions Experiment la predicts that when a particular social identity (friend vs. family) is made salient, then the person is more likely to "see themselves" in terms of that social identity. These differences should be evidenced in the content of a person's spontaneous self-descriptions. It is expected that when subjects have their "friend" social identity made salient (by rehearsing and evaluating the word "friend" in the first task), they will be more likely to spontaneously describe themselves in terms of their "friend" social identity. Conversely, it was also expected that when subjects have their "family" social identity made salient (by rehearsing and evaluating either the focal word "son" or "daughter"), they would be more likely to spontaneously describe themselves in terms of their "family" social identity. The effects described above might be particularly likely if subjects rehearsed the focal word in an evaluatively valenced context (idea salience conditions). Experimental design Subjects in experiment la were assigned to a 3 (focal word that was evaluated: friend, son daughter) x 2 (type of salience: concept vs. idea salience) between subjects factorial. The dependent measure was the number of instances in which subjects did or did not explicitly mentioned some aspect of their relationship/connection to either their "family" or "friends" in their spontaneous self-descriptions.

PAGE 44

34 Results One hundred and thirty five (N=135) subjects (different from the pretest) firom a southeastern university were randomly assigned to experimental treatments. For each social identity made salient, the percentage was calculated of instances in which subjects did or did not mentioned a connection to either their "family" or "friends" in their self-descriptions. The results (collapsed across type of salience induced conditions) are presented in figure 4. A logistic regression was performed defining the dependent variable as whether or not a subject explicitly mentioned some aspect of their "friends" or "family" in how they described themselves. The type of salience (concept vs. idea) and which social identity was salient (friend, vs. son vs. daughter) were the explanatory variables in a model including main effects and all possible higher order interactions. An analysis examined whether or not subjects explicitly mentioned some aspect of their "friend" identity m their spontaneous self-descriptions. In an omnibus analysis, there were no significant effects for the type of salience induced, which social identity was salient or any corresponding interaction effects (A'^'^^n-s). Null effects were also found in a planned contrast that compared the extent to which subjects mentioned some aspect of their "friend" identity in the friend made salient condition (33%) vs. the extent to which subjects mentioned some aspect of their "friend" identity both in the "son" and "daughter" made salient conditions (22% and 28% respectively, X^=n.s). A second analysis examined whether or not subjects explicitly mentioned some aspect of their "family" identity in their spontaneous self-descriptions. A main effect of which social identity was salient was found p^O.OOlO). In a follow up analysis, a planned conti-ast showed that more subjects spontaneously mentioned some aspect of their

PAGE 45

35 "family" identity when either "son" was salient (46%) or "daughter" was salient (53%) as compared to when "friend" was salient (16%, X^=UA5, p=0.0002). Discussion Consistent with hypothesis 2a, subjects who had their family identity salient by induction of either the "son" or "daughter" social identity salience manipulation were more likely to mention some aspect of their "family" in their subsequent spontaneous self-descriptions. The interpretation is that salience of the social identity made subjects more likely to see themselves in terms of that social identity indicating that the social identity was at least a temporary part of their working selfconcepts. This finding is consistent with the pattern of results obtained in the pretest that showed that the salience inductions also altered the extent to which subjects were able to identify the focal words "son" and "daughter" out of randomly arrayed letters. However, the attempt to affect the dependent measures by making "fiiend" salient was again unsuccessful in experiment la, suggesting the difficulty (at least with these manipulations for this sample) of pushing subjects' self-description around on the friend dimension. Experiment lb Purpose and overview of Experiment lb. The evidence in experiment la suggests that the salience manipulation for the "family" social identity was effective at altering subjects' temporary self-conceptions. Experiment lb will examine the extent to which this momentary state of self-conception mediates attitude formation when an attitude is generated toward an identity relevant object. More specifically, experiment lb assesses the necessity of both salience and object relevance (Hlb) in forming a favorable consumer attitudes (based on social identity when adopting such an identity would in fact

PAGE 46

36 provide a favorable basis to respond to the product). In experiment lb, subjects' "family" social identity was made salient. In an ostensibly unrelated study, subjects then evaluated a product that was either relevant or irrelevant to their family social identity. Procedure Subjects first participated in the same visual imagery study used in the pretest and experiment la. In this experiment, however, only the "family" social identity was made salient (between subjects) by having subjects evaluate the word "son" or "daughter" seven different times in the seven different typefaces. Additionally, only the concept salience manipulation was used in experiment lb. ^ After subjects completed the first task, a different experimenter administered an ostensibly unrelated "new product assessment study." Subjects were asked to evaluate three different "new product concepts." The first product was the focal stimulus. It was a palm held interpersonal telecommunication product. The second and third products were filler products. One of the filler products was an electronic portable head-band worn in the summer to keep cool and the other filler product was a flat tire re-filler for automobiles. For each of the three products, subjects were presented with a picture of the product and a product description. The focal product (interpersonal telecommunication product) was manipulated to be either relevant or not relevant to the social identity primed in the first task by framing the interpersonal telecommunication product (e.g., as in a product positioning strategy) in terms of the "family" social identity or not. For The "friend" salience manipulation was not used in experiment lb because of the data that suggested that the "friend" salience manipulation was not strong enough to heighten the salience of the friend focal word (pretest) or to alter subjects' working self concepts (experiment la). Similar logic led to usage of only the concept salience manipulation in experiment lb because the supporting data detected no meaningftil differences across concept and idea salience conditions in the pretest and experiment la.

PAGE 47

37 example, in the object relevant to the "family" (sons and daughters) social identity condition the focal product description included utilitarian features (e.g., calendar, address book, email) but with an emphasis on the product as a basis for staying connected to parents and family: ELECTRONIC COMPANION: Model (SN) DTR-3 Calendar, address book, to-do list, expenses, e-mail. More importantly, exchange information and well wishes with your loved ones at the touch of a button, keeping you connected with your parents and loved ones. This new product allows you to never miss important family events and/ or occasions that are dear to your heart. In fact, it will automatically remind you of them, and if you like, automatically send messages, flowers and e-cards. Custom messages can be programmed ahead of time and sent automatically. Enjoy up to a month's operation on the road. Compatible with Windows 3. 1, 95, NT 4.0 (optional adapter required for Mac). Modem operates on 2 AAA baUeries (included). Measures 4.5"x3.1"x.4". No such emphasis was made in the object irrelevant condition where the focal product description stressed only the presence of the utilitarian features: ELECTRONIC COMPANION: Model BUS-3 Calendar, address book, to-do work list, expenses, business e-mail. Exchange synchronized data with your PC at the touch of a button, eliminating dual data-entry and ensuring the safety and accuracy of information. This new product allows you to beam data to other Palm platform devices using the infrared port. In fact, its unique user-expandable design lets you add a modem, memory, software and more. Custom messages can be programmed ahead of time and sent automatically. Enjoy up to a month's operation on the road. Compatible with Windows 3. 1 95, NT 4.0 (optional adapter required for Mac). Modem operates on 2 AAA batteries (included). Measures 4.5"x3.1"x.4". Dependent measures The dependent measures were gathered as part of the second task. After reading the product descriptions, subjects indicated the likelihood that they would purchase each of the three products, ranging from (1) unlikely to (7) highly

PAGE 48

38 likely. They also indicated the likely market success of the products ranging from (1) unlikely to (7) highly likely. Subjects also wrote down in open-ended responses, what they thought were the best reason(s) for buying the product. They were subsequently probed in an additional open-ended question to write down any more reasons that they could think of The complete new product assessment study is included in Appendix D of this dissertation. Predictions For experiment lb, the framework predicts that more favorable attitudes should be observed in the following condition: when a social identity is salient and the consumer evaluates an identity relevant product. If the "family" social identity is made salient (as was evidenced in the pretest) and consumers are thinking about themselves in terms of their "family" social identity (as was evidenced in experiment la), then they should like the focal product that is framed in terms of their family identity more than the focal product that is not framed in terms of their salient family identity. Salient social identification connects a consumer to the identity relevant product, leading to more favorable attitudes as evidenced in higher purchase likelihood for the identity relevant focal product. No such interaction of design factors should emerge for purchase likelihood of the filler products because the filler products (held constant across conditions) should have no apparent linkage to the family social identity. Experimental design Subjects in experiment two were assigned to the cells of a 2 (Two focal words evaluated: Son vs. Daughter^) X 2 (Identity relevance: Relevant framed in terms of family or Not relevant framed in terms of features) between subjects As in the pretest and experiment la, subjects in the "son" condition were all male, subjects in the "daughter" condition were all female.

PAGE 49

39 factorial design with two control groups.^ The dependent measure was purchase Ukelihood for the focal product and purchase likelihood averaged across the two filler products. Results Sixty-two (N=62) subjects from a southeastern university were assigned to experimental treatments. An analysis of variance tested the main and interaction effects of salient social identity (sons vs. daughters) and object relevance (relevant: framed in terms of features + family or irrelevant: framed in terms of just features) on mean purchase likelihood for the focal product and the mean purchase likelihood averaged across the two filler products. The omnibus ANOVA of effects of identity and relevance on purchase likelihood of the focal product and filler products are reported in Appendices H and I respectively. The means and standard deviations are presented in Table 1 Consistent with predictions, an analysis on the average purchase likelihood of the two filler products revealed no significant effects (F < 1). Additionally, there was no significant social identity (son vs. daughter) by relevance interaction on purchase likelihood for the focal product indicating no differential effect of gender (son vs. daughter) within the "family" social identity. In other words, male subjects who had "son" identity made salient did not respond differently than female subjects who had "daughter" identity made salient. More importantly, for the focal product, the analysis yielded a significant main effect of relevance, (F(l,58)=26.67, p.< .0001). Subject's purchase likelihood for the focal product depended on whether the product was relevant In the absence of any salience manipulation, sixty (N=60) separate males and females participated in the new product assessment task that included between subjects, the focal product framed in terms of family plus utilitarian features (identity relevant conditions) or utilitarian features only (identity irrelevant conditions). No interaction of gender with identity relevance was found (F < 1) so mean purchase likelihood of the focal product and the two filler products in the object relevant and object irrelevant conditions was calculated included in the main analysis as controls.

PAGE 50

40 or irrelevant to the social identity that was made salient. Consistent with predictions, identity relevance resulted in higher purchase likelihood of the focal product when the family social identity was salient (3.35 vs. 5.31 and 3.20 vs. 5.12) in the "son" and "daughter" made salient conditions, respectively. To further exattiine these results, a follow up analysis was run that compared treatment means to their respective control groups using Dunnett's (1955) procedure (as reported in Keppel, 1991) controlling for family wide type one errors. Consistent with predictions, the ody treatment means that differed from their control group were when "son" was salient in the object relevant condition {Mson/objcctreievam McoNTROL = 1-92, crftical difference = 1.63, p < .01 FWa = .05), and when "daughter" was salient in the object relevant conditions ( ^daughler/objectre levant ^CONTROL = 1.73, critical difference = 1.63, p < .05 FWa = .05). Discussion and limitation(s) of the experiments The results of experiment la and experiment lb provide support that social identification can mediate attitude formation in ways that are consistent with the framework presented in this research. Experiment la established that when a social identity was made salient, the likelihood increased that a person would temporarily "see" themselves in terms of that particular social identity. In other words, heightening the salience of either "son" or "daughter" at least temporarilv altered subject's self-conceptions such that the social identity was an activated identity in their working self-concepts (as compared to when "friend" was salient). Experiment lb showed that differences in the salience of a particular social identity led to predictable differences in purchase likelihood (assumed to be guided by attitude formation). When a social identity was salient, subjects more favorably responded to an object relevant focal stimulus (product framed in terms of that social identity) as compared to when no such

PAGE 51

41 social identity was made salient (relevant control group comparison). This critical difference in attitude generation was found for the focal product, but not for the two filler products predicted to be unrelated to heightening the salience of the social identities in question. The combination of these two experiments suggest that salience of a particular social identity can lead to differences in how a consumer views herself These differences in a person's self-conception appeared to mediate attitude generation but only when the attitude object was perceived as being relevant to the salient social identity. Some limitations of these studies are of note. In experiment lb, no manipulation check on the relevance of the focal product to the social identity were taken. To some extent, this analysis assumes that subjects' perceptions of object relevance was the critical moderating factor in purchase likelihood assessments. Second, no analysis of subjects' self-descriptions was taken in the absence of a salience manipulation for the social identity "son" or "daughter." Therefore, it might be some what difficult to assume that the effects are truly being mediated by appreciable differences in how the salience induction technique was able to alter people's momentary spontaneous self-concepts. However, the self-description procedure is a demanding dependent variable since people are free to bring everything in their backgrounds to bear, and there is much less opportunity for the experimental context to affect the outcome. Given the fact that the predicted differences were found potentially suggests that the assessment of people's working self-concepts may have actually been underestimated given the open ended nature of the response task. These two limitations notwithstanding, the amalgamation of evidence across the pretest and experiments la and lb do provide consistent support for the prediction that a consumer's shifting social identification can affect their evaluative

PAGE 52

42 response if the social identity is brought to mind and the object that they are responding to is Hnked to that identity.

PAGE 53

43 We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and we varied it slightly to slow you down just a little bit. Now, please evaluate the meaning of the word "friend" in terms of "how it comes across to you" in the particular type-face presented. To provide us with the most accurate immediate response, read the phrase each time and try to rehearse it in your mind before providing your evaluations: Evaluate the Type face on the following dimensions provided in the box THINK ABOUT THE FRIEND Warm Impersonal IMPORTANCE OF Memorable Bland BEING A Fun Boring Common Unique CONSIDER THE Warm impersonal VALUE OF BEING A Memorable a Bland Fun a Boring Common Unique PONDER THE FRIEMD Warm a a Impersonal SIGNIFICANCE OF Memorable a a a Bland BEING A Fun a a Bonng Common Unique Figure 2. Pretest: example of the idea salience eondition.

PAGE 54

o 0.6 Friend Salient Social Identity Figure 3. Pretest: percentage of response category frequency as a function of which social category was salient. Family identity was made salient by having male subjects rehearse the word "son" and having female subjects rehearse the word "daughter: (N=110)

PAGE 55

45 Figure 4. Experiment la: self descriptions that referenced relationship with social identity as a function of which social identity was salient. Family identity was made salient by having male subjects rehearse the word "son" and having female subjects rehearse the word ''daughter." (N-135)

PAGE 56

46 Table 1. Experiment lb: means and standard deviations of purchase likelihood as a function of salience and object relevance for focal product and filler products. Object Relevant Focal Product Object Irrelevant Focal Product Focal Filler Focal Filler Product Products Product Products Son 5.31" 3.46 3.35 3.91 Salient 1.11 1.75 1.41 1.42 Daughter 5.12' 3.91 3.20 3.37 Salient 1.36 1.30 1.86 1.76 Control Group Control Group Focal Product Filler Products Purchase Likelihood Relevant 3.39 Purchase Likelihood Relevant 3.93 Irrelevant 3.43 Irrelevant 4.13 Note: a= Different from control at p < .05, b^Different from control at p < .01.

PAGE 57

CHAPTER 6 EXPERIMENT 2: THE ROLE OF SALIENCE, SELF-IMPORTANCE and OBJECT RELEVANCE Purpose of the Study The framework argues that there are four key moderating factors that determine whether the evaluative content of a social identity will lead to attitude generation. Experiments la and lb assessed the first two factors: salience and object relevance. Experiment two goes beyond experiments la and lb by examining the third factor; i.e., the moderating impact of a social identity's self-importance on the formation of favorable consumer attitudes (based on social identification processes). Theoretical Premise of the Study Marketers can attempt to link consumer social identities to their product offerings. For example, a soft drink (e.g.. Sprite ) may attempt to position itself as the beverage of choice for some social identity of interest (e.g., Generation-X nonconformists). However, according to the framework in this research, consumers are more likely to adopt social identities that they consider to be "self-important" in terms of a basis for self-definition. Therefore, marketing communications that attempt to connect a brand to some social identity of interest must consider the extent to which a social identity is "valued" by consumers in its particular target market. Otherwise, marketing efforts that try to induce social identity based attitude formation toward brands and products will be ineffective. Building on experiment lb, experiment two tests the following hypothesis: 47

PAGE 58

48 H2(a): Heightening the salience of an identity leads to higher (lower) evaluations of an object relevant (object irrelevant) focal stimulus, particularly if the identity is self-important (not self-important) to the consumer. Therefore, the more that a consumer sees a social identity (e.g., "working mother") as being self-important, the more likely she is to base a response toward an object relevant product (e.g., attitudes toward mini-vans) on the evaluative content linked to that social identity. Experiment 2 Overview of Experiment 2 In experiment two, subjects formed attitudes under three conditions. First, when a social identity was perceived as being relatively selfimportant or not. Second, when a social identity was salient or not. Third, when the social identity was either relevant or not to the object that was to be evaluated. Procedure In the first task, subjects participated in a "handwriting assessment study" that investigated the link between consumer characteristics (e.g., being frugal) and handwriting style. Therefore, as part of the research project, subjects were asked to give samples of their handwriting. As part of the cover-story, in order to provide a "baseline," subjects first wrote three neutral sentences in their natural handwriting style. The cover story then indicated that part of the research looked at "whether or not WHAT people wrote about was related to their handwriting style." Therefore, subjects were instructed to write five independent statements concerning a particular topic. When subjects provided their handwriting samples, they wrote about their selfconception of their "family" (Baldwin and Holmes, 1987) social identity. This was intended to manipulate the salience of the "family" social identity. Self-importance was

PAGE 59

49 manipulated as high (low) by having subjects write about a particular self-important experience or event that emphasized the interdependent (independent) nature of their connectedness with their family. In the high self-importance conditions, subjects read the following directions: PLEASE FOLLOW THE INSTRUCTIONS BELOW: We'd like you to write 4 independent statements (complete sentences in your usual handwriting style, i.e., cursive or not) each with about 10 to 15 words The statements should convey some positive, deeply moving, emotionally involving thoughts and or sentiments that describe the commitment to your relationship with one or both of your parents Concentrate on how you have maintained your family ties while also STRENGHTHENING the sense of connectedness as a member of your family. Subjects in the low self-importance conditions read the following instructions: PLEASE FOLLOW THE INSTRUCTIONS BELOW: We'd like you to write 5 independent statements (complete sentences in your usual handwriting style, i.e., cursive or not) each with about 10 to 15 words The statements should convey some positive, deeply moving, emotionally involving thoughts and or sentiments that describe the commitment to your relationship with one or both of your parents. Concentrate on how you have maintained your family ties while also STRENGTHENING a sense of independence as an individual young adult. Subjects then completed a battery of consumer personality measures as well as other cover story-consistent items. The full stimulus materials for the handwriting study are presented in Appendix E. Subjects then participated in the same ostensibly unrelated "new product assessment study" used in experiment lb including the same dependent measures'. To manipulate salience, a delay condition was added between subjects. In the high salience conditions subjects formed attitudes toward identity relevant or irrelevant products This was done to allows comparisons of purchase likelihood levels across experiment lb and experiment 2 as well as a comparison between the control groups from experiment lb who had no salience or self-importance manipulation prior to evaluating the three products.

PAGE 60

50 immediately after generating and writing down the self-importance related thoughts. In the low salience conditions subjects formed attitudes toward identity relevant or irrelevant products following a fifteen-minute delay (filler task). The same object relevance manipulations used in experiment lb (relevant to the "family" social identity vs. not relevant to the "family" social identity) was used in Experiment 2. Predictions For experiment two, the framework predicts a significant three-way interaction of design factors on purchase likelihood of the focal product. Favorability of attitudes (toward the focal product) as evidenced in purchase likelihood measures should depend on whether or not a social identity is object relevant and self-important. Object relevance (i.e., when the particular social identity is relevant to the stimulus to be evaluated) combined with the heightened self-importance of that social identity should increase the likelihood that the attitude object (new product concept) will be considered in terms of the particular social identity. Hence, there should be a higher likelihood that the evaluative content of the social identity should be the basis for attitude formation. This should lead to more favorable evaluations of the focal product and hence, higher purchase likelihood. The effect should be attenuated in delay conditions where the salience induced by the self-importance manipulation should decay relative to the effect of self-importance induced by generating interdependent, self-important thoughts regarding the social identity. Null effects of design factors were expected for the object irrelevant focal product and all filler products. Experimental design Subjects were randomly assigned to the cells of a 2 (Selfimportance: High or low) X 2 (Salience: High or low) X 2 (Object relevance: Relevant or Not relevant) between subjects factorial design. The dependent measure was purchase

PAGE 61

51 likelihood for the focal product and the average purchase likelihood for the two filler products. Manipulation checks To check the self-importance manipulation, a separate pretest was run. Twenty-seven (N=27) subjects first completed the handwriting assessment study. Subjects then answered fifteen manipulation check items intended to assess the effectiveness of the self-importance manipulation (e.g., I strongly identify with my Family, Being a member of my family often affects how I tend to view the world around me, etc). Items ranged from 1= Strongly disagree to 5=Strongly agree and four of the items were reverse coded. After completing the handwriting task, subjects completed an ostensibly unrelated consumer emotion and mood study" where they responded to the two lO-ltem mood scales (using "momentary" instructions) that comprise the Positive and Negative Affect Schedule (PANAS) Watson, Clark and Tellegen (1988) included in Appendix F of this dissertation. This was included to assess and or rule out the alternative explanation that mood changes induced by the self-importance manipulation may have led to any observed effects. The fifteen manipulation check items were collapsed into one measure of selfimportance (a =.783) and the means of the self-importance measure were calculated as a fiinction of the low and high self-importance conditions. The averages in the low and high self-importance conditions were M,owseifimp=3.57 Mhighseif,mp=4.19, (F(l,25) = 29.58 p<.001) respectively, indicating that the self-importance manipulation altered subjects perceptions of the self-importance of their "family" identity. As a fiinction of low and

PAGE 62

52 high self-importance conditions, no differences in mood were found on either the positive affect (PA) or negative affect (NA) dimensions of the PANAS scale (F's < 1). Results One hundred twenty-one (N=121) subjects from a southeastern university were randomly assigned to experimental treatments. An analysis of variance tested the main and interaction effects of self-importance of the "family" identity (low vs. high), salience (low vs. high) and object relevance (relevant: framed in terms of features + family or irrelevant: framed in terms of just features) on mean purchase likelihood for the focal product and the two filler products. The omnibus ANOVA effects of salience, relevance and self-importance on purchase likelihood of the focal product and the filler products are reported in Appendices J and K respectively. The average purchase likelihood as a function of salience, self-importance and object relevance for the focal product and filler product 1 and filler product 2 are presented in figures 5, 7 and 9 respectively. For the focal product, although the means were in the predicted direction, the three way interaction of self-importance, salience and object relevance was not significant, (F < 1). However, consistent with predictions, main effects of selfimportance, (F(l, 113)-11.30, p < .01, and object relevance, (F(l,l 13)=14.28, p <.01 were qualified by a significant two way interaction of self-importance by object relevance (F(l,l 13)=4.33, p < .05). Subjects' purchase likelihood of the focal product depended on whether or not the self-importance of their family identity was low or high and whether or not the focal product was framed in an identity object relevant or irrelevant fashion. To examine the nature of this result, figure 6 shows mean purchase likelihood of the focal product as a function of self importance and object relevance (collapsed across sahence

PAGE 63

53 conditions). Consistent with predictions, the mean purchase likelihood was highest in the high self-importance / identity object relevant condition, M=5.33 than in the high selfimportance / identity object irrelevant condition, M=3.64, the low self-importance / identity object relevant condition, M=3.76 or the low self-importance / identity object irrelevant condition, M==3.27. As a secondary analysis, Dunnett's procedure was again used to test differences betweeri treatment means of purchase likelihood against the baseline control groups (from experiment lb) who received no manipulation prior to the evaluation of the new product concepts. Again consistent with predictions, the only treatment mean that differed from its control group was when "family" identity was highly self-important in the object relevant condition {Mseifia,pnigh/objectreievant ^CONTROL = 1.95, critical difference = 1.51, p < .01 FWa = .05). Curiously, for filler product 1, a significant main effect of self-importance (F(l,113)= 6.53, p < .05) was found. Collapsing across self-importance and salience conditions, revealed that high selfimportance of the "family" identity lead to greater purchase likelihood for the second filler product (Miowseifimp = 2.32, Mhighseifimp= 3.15) which was a high-tech electronically air-cooled head band. However, a follow up analysis comparing these treatment means further broken down by identity relevance conditions revealed no significant differences as compared to relevant control groups from experiment IK Finally, consistent with predictions, an analysis revealed no systematic treatment effects on purchase likelihood for the third filler product (F's < 1) and no significant differences between corresponding control groups and treatment means of purchase likelihood for the third filler product. Discussion and limitation(s) of the experiment The results of experiment two provide support for hypothesis 2(a). The pattern of data is consistent with the assertion

PAGE 64

54 that self-importance is an important moderating variable in the process of social identity based attitude formation. Results of this experiment suggest that when people see a particular aspect of their social identification as being relatively self-important, they are more likely to favorably evaluate an identity relevant object. Based on relevant control group comparisons, this pattern was found for the focal product and not for filler products predicted to be deemed irrelevant to the particular social identity in question. These effects appear not to be confounded with any differential levels of positive affect that might have been associated with generating high versus low self-importance thoughts. Several limitations of this experiment are of note, however. First, an effect in the omnibus test for purchase likelihood of filler product 1 was unexpectedly found. Subjects who perceived their "family" identity as being self-important, on averaged reported higher purchase likelihood for the high-tech electronically air-cooled head band. This effect seems somewhat odd. A review of the description of filler product 1 was conducted and is shown in figure 9. Two things are of note here. Firstly, and purely based on speculation, the description of filler product 1 starts with a lead-in rhetorical question "When it's blazingly hot, don't you dread leaving the air-conditioned comfort of your home?" It might be possible that subjects in the high self-importance conditions (who generated thoughts concerning their interdependence with their parents) somehow interpreted "home" as having some connection to their familial identity. This might have lead to slightly higher purchase likelihood. Secondly, an examinatioR of the mean levels for filler product 1 shows that mean purchase likelihood did not reach the midpoint of the scale for any of the treatment groups. This means that no one really "liked" the product that much (to the extent that favorability maps onto purchase likelihood), rather in the

PAGE 65

55 high self-importance conditions, subjects disliked it less than subjects in the low selfimportance conditions. This notion of "statistical significance" versus "practical importance" combined with the fact that the follow up control group analyses suggested no real differences, is taken for evidence that this effect may be spurious and or trivial. Additionally, in Experiment 2, salience did not interact with design factors as predicted by the theory. There are several reasons why this might have occurred. Although the manipulation check on self-importance revealed a statistically significant difference on the fifteen item self-importance measure for the hold-out sample, inspection of the actual mean levels suggests that the psychological difference in subject's minds might have been minimal. This makes sense given the reasonable expected difficulty in heightening the self-importance of a person's familial identity. However, because the mean differences in the low and high self-importance manipulation check appear to be minimal, the fifteen minute intervening task intended to alter the salience induced by generating and writing down self-important related thoughts, may have been at an inappropriate delay interval. Another speculative explanation might suggest that the saHence of this kind of social identity (i.e., "family") may be highly correlated with selfimportance. The operationalization of salience in experiment two is based on the assumption that the salience of social identification induced by the self-importance manipulation decays at a relatively different rate than the salience of the self-importance related thoughts per se. Research has suggested that certain bases for self-conception tend to be chronically salient for some individuals (Bem, 1981). An interesting follow up study might attempt to determine the precise self-importance thresholds (for a particular social identity) beyond which the social identity tends to be chronically activated.

PAGE 66

56 Unfortunately, no manipulation check proxy for salience in this study was taken. Nonetheless, support was found for the assertion that self-importance (as operationalized in this study) is a moderating factor in determining the extent to which the evaluative content of a social identity will impact on attitude generation.

PAGE 67

57 High SelfLow SelfImportance Importance Self-lmoortance : [T] Rele\ant (Family + Features) High SelfLow Selfimportance Importance Self-importance A B Figure 5. Experiment 2: purchase likelihood of Focal product as a function of object relevance, self-importance and low salience (A) and high salience (B) of the social identity Low High Self-importance Identity Object Irrelevant i Focal Product Identity Object Relevant | Focal Product Experiment 2: Figure 6. purchase likelihood of Focal product as a function of object relevance and self-importance (collapsed across salience conditions)

PAGE 68

D O O 4 o w (S 3 3 a. Hgh SelfLow Selfhportance Importance Self-importance 6 3 Rele\Qnt (Family + Features) I Irrelevant (Features) High SelfLow Selfimportance Inrportance Self-importance A B Figure 7. Experiment 2: purchase likelihood of Filler Product 1 as a function of object relevance, self-importance and low salience (A) and high salience (B) of the social identity PERSONAL COOLING SYSTEM When it's blazingly hot, don't you dread leaving the air-conditioned comfort of your home? Maybe you've tried an ice-soaiced towel around your neck — but that little bit of evaporative cooling is nothing compared to the long-lasting, total-body comfort of our all-new 2.0 version of the Personal Cooling System. It's the best way to stay comfortable — and avoid heat stress — when walking, mowing, gardening, watching sports or concerts...just about any activity, anywhere. This new and improved 2.0 model offers two levels of cooling: "regular" helps you feel up to 20F cooler than the hot, dry air around you; and "high" now offers 20 percent greater cooling — for even the hottest days. What's more, the tiny fan inside version 2.0 adds even more cooling by directuig a gentle breeze over the back of your neck. The 2.0 model also is adjustable to fit a wider range of neck sizes comfortably and securely. It's trimmer and lighter, too: the sleek, die-cast aluminum body weighs just 10 oz. empty; 14 oz. filled. Runs on one AA battery (order separately). Figure 8. Experiment 2: description of Filler Product 1

PAGE 69

59 •a o o M (S o 3 Q. High SelfLow SelfImportance Importance Se If-lmportance o a o JZ [Tj Relevant (Family + Features) g irrelevant (Features) High Selfimportance Low Selfimportance Seif-importance A B Figure 9. Experiment 2: purchase likelihood of Filler product 2 as a function of object relevance, self-importance and low salience (A) and high salience (B) of the social identity

PAGE 70

CHAPTER 7 EXPERIMENT 3: THE ROLE OF SELF-IMPORTANCE AND EVALUATIVE DIAGNOSTICITY Purpose of the Study and Hypotheses to be Tested Suppose a consumer's political identity (e.g., "Democrat") is highly salient (e.g., the consumer is attending a political rally) at the time an attitude toward an identity relevant object is called for (e.g., the consumer is asked to assess and sign a petition against estate taxes). Suppose that the consumer is asked to either giver her attitude toward the estate tax, or judge the readability of the real estate tax petition itself The first judgment (attitude toward estate taxes) may be highly diagnostic to her Democrat identity (i.e., a clear norm might exist Kallgren, et. al 2000) in which case her political identity is likely to be the basis for the response. However, the second judgment (readability of the petition) may be less diagnostic to her political identity, minimizing the effect of her political identity's salience and self-importance on her subsequent judgment. Experiment three is a partial replication of experiments lb and 2, but goes beyond those by demonstrating that social identity will impact consumer attitude formation only when the social identity is evaluatively diagnostic to the judgment (Feldman and Lynch, 1988). In experiment three, this notion was tested by examining the relationship between self-importance of a social identity and the formation of response judgments in instances where the identity does (does not) provide and evaluative basis to respond: 60

PAGE 71

61 H3(a): The self-importance of a social identity is more likely to be positively related (unrelated) to evaluative dimensions that are diagnostic (non diagnostic) to the social identity. Overview of the Experiment During the experiment, subjects generated evaluative responses under tv^o different conditions. When the social identity was salient (not salient), and when the attitude object was relevant (not relevant) to the social identity. The key distinction in this experiment was that the evaluative responses tapped dimensions in which the social identity provided a basis (no basis) to make a judgment. More details follow. Experiment 3 Participants and experimental design Subjects were taken from a sample of 140 freshman and sophomore students, ages fourteen to sixteen at a southeastern high school, They participated as part of course credit. The students were about 50% White, 30% Black, 15% Hispanic, and 5% other. The sample had an equal number of males and females. In experiment three, the students were randomly assigned to the cells of a 2 (Salience: High or moderate) X 2 (Object relevance: Relevant or Not relevant) between subjects factorial design. In addition, participants were classified as either high or low in identification (self-importance) with the social identity. Procedure. At time one, a pretest was conducted to determine various selfimportant social identities for this particular sample. In this pretest, subjects spontaneously generated various social identities in response to the prompt "describe yourself" Based on their open-ended responses, a list of various social identities was generated. Several weeks later at time two, a series of "social identity" scales was administered. These measured the "self-importance" (i.e., whether participants identified

PAGE 72

62 with, admired and characterized themselves as being a person who holds the identity) of a subset of social identities generated at time one. Based on a preliminary analysis of the data generated at time to and to avoid ceiling or floor effects, a social identity was chosen (for use in the main experiment) in which the distribution of self-importanee scores was roughly uniformly distributed amongst participants. Approximately two weeks later at time three, each subject was given a series of various surveys and research tasks. Buried within the tasks was a "consumer web-page assessment study." Subjects were told that marketers often want to understand the potential persuasiveness of web page content. In order to understand how to create web page media that have the most impact, message content and web page design is often pretested on relevant populations. Subjects were told that this study would assess their perceptions regarding types of message content, types of web page content, web page ads and or products/services. Independent variables The social identity chosen from the pretest was "Future College Educated Leaders of America." The background information subjects read prior to the main experiment manipulated the salience of the social identity. Subjects were told that research pre-testing on the consumer web-page assessment study had already begun on three "market segment clusters" that were subsequently described. In the high salience conditions, the last of three "segments" read as follows: This research has also been conducted on a segment conveniently entitled "Towns and Gowns." This segment makes up about 1.2% of U.S. households. The towns and gowns segment earns on average about $17,862. Members of the Towns and Gowns segment typically range from ages 18 to 34. They have been described as college educated, and highly likely to be the future leaders of this country. They are likely to read magazines like Natural History and watch television shows like Good

PAGE 73

63 Morning America. Sample ZIP codes include College Station, Texas 77840; Bloomington, Indiana 47401; Ithaca, New York 14850 and Ft. Lauderdale, Florida 33324. In the low salience conditions, the last of three "segments" read as follows: This research has also been conducted on a segment conveniently entitled "Blue Blood Estates." This segment makes up about 1.2% of U.S. households. The blue blood estates segment earns on average about $17,862. Members of the blue blood estates segment typically range from ages 1 8 to 34. They have been described as high school educated, and highly likely to possess reasonable spending power. They are likely to read magazines like Natural History and watch television shows like Good Morning America. Sample ZIP codes include College Station, Texas 77840; Bloomington, Indiana 47401; Ithaca, New York 14850 and Ft. Lauderdale, Florida 33324. In the main experiment, subjects examined the target product/service of interest which was a Smithsonian Magazine Association membership. The subjects saw an example of the Web page. Subjects were then asked to take several seconds to inspect the web page. They were told to read all information carefully. Afterwards, subjects read a description of the service that was depicted on the web page. In the object not relevant condition, the Smithsonian Association membership was described in terms of strict, utilitarian benefits (attributes): Why would you want to be a Smithsonian National Associate Member? Just think about the pros and cons! Being a Smithsonian National Associate Member has many advantages. For example, each member receives twelve total issues of the Smithsonian Magazine at a special rate that has been discounted specifically for Association members. Twelve issues sent directly to your home or office. In addition to book, record and video discounts as well as other discounts at various museum shops, each member is also kept up to date on various notices of important events and intellectual activities going on in your area. Also, members of the Smithsonian National Association receive eligibility for specially designated travel programs, as well as free admission to a number of the country's best museums and other cultural events. All of these amazing features for just $26.00 for the entire year! Why don't you just weigh the costs and benefits! How can you lose?

PAGE 74

64 In the object relevant condition the Smithsonian Association membership was described in terms of the reference group in question: Why would you want to be a Smithsonian National Associate Member? Just think about whom might be likely to possess such a membership? Think about the type of people who would receive the twelve issues of the Smithsonian Magazine that is included in the membership. What type of person? Individuals who always present themselves with the highest level of intelligence and future success. For example, college educated individuals of this country might keep up to date on various notices of important events and activities going on in their area. These Smithsonian National Association members are the future leaders of this country. The magazines that they read tell others that they appreciate the more culturally involving things in life. Can't you imagine yourself as this type of person, i.e., a college educated, future leader of this country and Smithsonian Associate Member? All of these amazing features for just $26.00 for the entire year! How can you lose? After reviewing the above information, subjects rated the web page on various evaluative dimensions and then reported their attitude toward the Smithsonian Associate membership. Dependent measures Subjects were asked to rate the example web page on the following six dimensions: interestingness, readability, visual appeal, persuasiveness, complexity, and design quality. The main dependent measure asked subjects to rate the extent to which they felt favorable or unfavorable toward the Smithsonian Association membership, which ranged from unfavorable (1) to very favorable (7). The full set of stimulus materials for this experiment is presented in Appendix G. Predictions Of the seven dependent measures that were chosen, three are expected to be diagnostic to the salient social identity: interestingness and persuasiveness of the web page as well as overall attitude toward the product depicted on the web page. The other four dimensions: readability, visual appeal, complexity and design quality of the web page are expected to be non-diagnostic. Therefore, the self-importance of a

PAGE 75

65 salient (not salient) social identity— Future college leaders of America—to an object relevant (object irrelevant) stimulus-Smithsonian National Associate Membership framed in terms of the social identity-should be more positively correlated (uncorrelated) with dimensions that are diagnostic— interestingness, persuasiveness, attitude than dimensions that are (nondiagnostic)-readability, visual appeal, complexity and design quality) to the social identity. Manipulation checks A hold out sample was used to check both the salience and the object relevance treatment manipulations. To check the salience manipulation, subjects were asked to read the intro and background information to the Web page assessment study. One of the questions asked the subjects to indicate at that particular moment, to what extent did the information make them think about their identity as a future college educated leader of America? The scale ranged from Did not make me think about it (1) to Made me really think about it (7). The difference across the high salience and low salience conditions was significant (N=34, Mhigh=4.47, Miow=3.53, F(l, 32)=4.59, p=.04). To check the object relevance manipulation, subjects were asked to read the descriptions of the service. One of the questions asked the subjects to indicate to what extent did the service seem relevant to what it means to be a Future college educated leader of America ranging from Not relevant (1) to Very relevant. The difference across the object relevant and object not relevant condition was highly significant (N=32, Mrelevant=5.25, MNotrelevant=2.44, F(l, 30)=65.60, p <.001). Results The key effects tested in this experiment were the relationships between evaluative dimensions and the self-importance of the social identity in question.

PAGE 76

66 Evaluative dimensions Separate analyses were run on each of the six evaluative dimensions (interestingness, readability, visual appeal, persuasiveness, complexity, and design quality) and the subjects reported attitude toward joining the Smithsonian Associate membership. For these analyses, self-importance of the social identity that was measured two weeks prior to the main experiment was included in the analysis as an independent continuous variable. Self-importance of the social identity consisted of the average of three items (a = .853). In a prior survey, subjects were asked to indicate the extent to which "ftjture college educated leaders of America" described them ranging from Does not describe (1) to Describes me perfectly (7), the extent to which they identified with that group ranging from Do not identify w/ group in any way (1) to Strongly identify with the group (7), and whether or not they admired the group ranging from Do not admire the group (1) to Really admire the group (7). Predictions were tested in an ANCOVA including main effects of salience and object relevance and all possible 2 and 3 way interactions of self-importance, salience and object relevance. The omnibus ANCOVA effects for each dimension is reported in Appendix L. The analysis revealed a significant three-way interaction of self-importance x object relevance x saHence (F(l, 90) = 18.84, p < .0001). The effect of self-importance of the social identity on subjects perceptions of the interestingness of the web page, depended on two things: whether or not the subject's social identity (future college educated leaders of America) was made salient or not and whether or not the object was relevant (message was framed in terms of the social identity or object features). A separate model was run to investigate the nature of this result. Figure 10 reports a

PAGE 77

67 separate, slope estimate of the self-importance continuous variable nested in each treatment condition. Corresponding t value(s) are in parenthesis. Consistent with predictions, the key result here is that the effect of selfimportance of the social identity is related to this relevant dimension only when the social identity is salient and the attitude object is relevant to the social identity (i.e., when the message has been framed in terms of the social identity). So in the high salience and object relevant condition, every one-unit change in self-importance of the social identity results in a statistically significant, .62 positive unit change in "interestingness" (t=6.95, p <.0001) of the web page. This is indirect evidence that self-importance of the social identity impacts the formation of this relevant evaluative response. As self-importance of social identification increases, so do perceptions of the webpage's interestingness. The same analysis was run on the "readability" and the "visual appeal" dimensions. According to the framework, the self-importance should be related only to evaluatively diagnostic dimensions. It should have no effect (should not interact with the treatment conditions) in determining perceptions of "readability" or "visual appeal." Consistent with predictions, no effects were found to be significant (F's < 1). An analysis examined the effects of self-importance, salience and object relevance on the "persuasiveness" dimension. This analysis revealed a significant twoway interaction of self-importance x salience for the persuasiveness dimension (F(l, 90) = 6.53, p < .0123). The effect of self-importance of the social identity on subjects perceptions of the persuasiveness of the web page, depended on whether or not the subject's social identity was made salient or not. A separate model was again run to investigate the nature of this result. Figure 1 1 reports a separate slope estimate of the

PAGE 78

68 self-importance variable nested in each class variable treatment condition. Corresponding t value(s) are in parenthesis. Very similar to the "interestingness" dimension and consistent with predictions, the key result here is that the effect of self-importance is positively related to this relevant dimension only when a social identity is salient. More specifically, in the low salience condition, every one unit change of self-importance results in a .483 unit change of persuasiveness (t=5.34, p<.0001) while in the high salience condition, every one unit change of self-importance results in a .729 unit change of persuasiveness (t=7.23, p<.0001). This is again, consistent with the notion that self-importance impacts the formation of this relevant evaluative response. As self-importance of the social identity increases, so do perceptions of persuasiveness of the web page. Subjects in these conditions may be relying on social identification as a basis for generating perceptions of message persuasiveness. The same analysis was run on the "complexity" dimension. Surprisingly, this analysis revealed a marginally significant three-way interaction of strength of selfimportance X object relevance x salience interaction for the complexity dimension (F(l, 90) 4.19, p < .0435). The effect of self-importance of the social identity on subjects perceptions of the complexity of the web page depended on whether or not the subject's social identity was made salient or not and the attitude object's relevance (i.e., the type of message that they were exposed to). A separate model was again run to investigate the nature of this unexpected result. Figure 12 reports a separate slope estimate of the selfimportance variable nested in each treatment condition. Corresponding t value(s) are in parenthesis.

PAGE 79

69 These results are similar to the "interestingness" and "persuasiveness" dimensions. However, the self-importance of the social identity impacts this dimension only when a social identity is salient and the message is framed in terms of the social identity. In the high salience conditions, perceptions of complexity are correlated with self-importance of the social identity. More specifically, in the high salience condition, every one unit change of self-importance results in a weak, albeit significant .155 unit change of perceptions of complexity (t=1.81, p<.10). The same analysis was run on the "design quality" dimension. Consistent with predictions, this analysis revealed no significant interactions between self-importance of the social identity and the other conceptual variables of interest (F's < 1). After evaluating the web page on the six dimensions previously outlined, subjects were asked to give their attitude (very unfavorable (1) to very favorable (7)) toward the Smithsonian Associate membership. This analysis revealed a significant two-way interaction of self-importance x object relevance for the attitude dimension (F(l, 90) = 30.84, p < .0001). The effect of self-importance on subjects' attitude toward the Smithsonian membership depended on whether or not the message was framed in terms of the social identity (object relevant) or object attributes (object not relevant). A separate model was once again run to investigate the nature of this result. Figure 13 reports a separate slope estimate of self-importance nested in each treatment condition. Corresponding t value(s) are in parenthesis. Consistent with predictions, the key result is that the effect of self-importance of the social identity is related to attitudes toward the Smithsonian Associate membership only when the message is framed in terms of the social identity. In the moderate salience

PAGE 80

70 conditions, every one unit change of self-importance results in a .543 unit change in attitudes toward the Smithsonian membership (t=4.73, p<.0001) while in the high salience condition, every one unit change of self-importance results in a .6947 unit increase in attitudes toward the Smithsonian membership (t=5.43, p<.0001). The effect of self-importance of the social identity in these two treatment conditions closely mirrors the pattern of results found for the persuasiveness dimension. This is intuitively logical, because out of all six dimensions, one could argue that the "persuasiveness" dimension should probably be most correlated with overall attitude. Nonetheless, this result is again consistent with the notion that self-importance of the social identity impacts the formation of attitudes toward the object when the product is framed in terms of social identification based concerns. Discussion and limitations It was hypothesized that the self-importance of a social identity is more likely to impact attitude formation when a salient social identity provides a basis for an attitude. The results of this experiment provide indirect evidence for this assertion. Strong support was found for hypotheses H3a. For example, selfimportance of the social identity positively co-varied with interestingness of the web page for participants who were led to form an attitude based on social identity and for whom the social identity was made salient. This effect was somewhat stronger for the persuasiveness dimension where a stronger effect of self-importance of the social identity was found when social identity salience was high versus low. Although it was a much weaker effect statistically, the pattern of results for the complexity dimension somewhat

PAGE 81

71 mirrored the results for the interestingness dimension. This was not expected.' However, a very strong effect of self-importance of the social identity was also found for the attitude dimension as a function of object relevance. This pattern of slope effects was very similar to the results found for the persuasiveness dimension. In general, this experiment provides consistent evidence that social identification (more specifically, the potency of group identification as captured by the self-importance variable can be related to the formation of responses in conditions where conceptually, a consumer is more likely to be drawn to a salient social self-identity which provides a basis for a response (message framed in terms of social identity) on measures that are evaluatively diagnostic to the social identity. Several limitations of this study are of note, however. Firstly, although taken approximately two weeks earlier, the self-importance of the social identity was a measured variable. Of course, this weakens the ability to assert a causal relationship between self-importance of the social identity and attitude formation in this experiment. Therefore, a more stringent test of the theory would attempt to manipulate the selfimportance of a social identity (as was attempted in experiment two). Secondly, no direct behavioral measure was taken.^ Therefore, one could argue that the attitudes expressed by participants in this experiment would not actually be diagnostic guides for action. Subjects rated the web page's complexity immediately following the persuasiveness dimension. Although purely post hoc speculation, it might be argued that the idiosyncratic nature of the sample of participants (high school students) and the close proximity of these two dependent measures may have led them to infer some kind of relationship between persuasiveness and complexity. Philosophically, it can be argued that expressing an attitude is a "behavior." However, a more complete test of the framework will include additional measures that represent action toward the object in question.

PAGE 82

72 This criticism is quite valid. But the focus of this research is on how the evaluative content of a social identity might affect attitude formation Whether or not the formed attitude will guide actual behavior is a separate issue. Thirdly, the manipulation check taken on the salience manipulation was quite weak both in its statistical significance and the mean levels across conditions. However, the results show that in the relevant treatment conditions, the patterns of results produced the expected variation in the measure of the relationship between the conceptual independent and dependent measures. Therefore, a specific alternative explanation must be offered that assumes no relationship between the conceptual independent and dependent variables described by the framework being tested. If no such alternative explanation exists, then having a "stronger" independent variable check does not add anything necessary, and even a failure to have one does not constitute a flaw (Sigall and Mills, 1998). A related issue is that the operationalization of the object relevance manipulation probably made the social identity salient too. Therefore, the relative role on the dependent measures of pure salience induced by the introduction of the experiment's cover story and salience induced by the object relevance manipulation is difficult to distinguish in this experiment. Certainly, an issue for future research. Finally, there is a different general theoretical account for these findings.^ This experiment provides no direct evidence for the process described in the overall model (figure one). The results presented in experiment one do not speak directly to ^ A different general theoretical account is not the same as an alternative explanation. The former "reinterprets" the proposed causal relationship in the context of another framework. However, the latter assumes no relationship between the conceptual variables that the theory asserts to be related and may propose that totally different conceptual independent variables resulted in dependent measure variation.

PAGE 83

73 differentiating the process of attitudes formed on the basis of a social identity from other processes. For example, take a multi-attribute explanation of attitude formation. It can be argued that subjects in the object relevant (framing of the message in terms of social identity) conditions, were simply led to create an additional "belief about the object, and the salience of the social identity simply altered the evaluative implication of that additional belief. Of course it is possible to post hoc explain these or any other results in terms of a multi-attribute model (i.e., shifting importance weights of newly created belief elements in a multi-attribute premise). However, the purpose of this framework is to describe the process of how a consumers' social identification is likely to affect her attitude. Nonetheless, future experiments must also incorporate other process measures that would be expected to differentiate other theoretical accounts.

PAGE 84

RELEVANT MODERATE SALIENCE HIGH OBJECT RELEVANCE NOT RELEVANT .62 (**t = 6.95) .114 (t1.31) .003 (t = .04) X = SELF-IMPORTANCE ; Y =INTERESTINGNESS Figure 10. Experiment 3: interestingness as a function of salience, self-importance and object relevance

PAGE 85

RELEVANT OBJECT RELEVANCE NOT RELEVANT MODERATE SALIENCE -.152 (t = -L54) HIGH .729 (** / = 7.25) .091 (t = .99) X = SELF-IMPORTANCE ; Y =PERSUASIVENESS Figure 1 1 Experiment 3: persuasiveness as a function of salience, self-importance and object relevance

PAGE 86

RELEVANT OBJECT RELEVANCE NOT RELEVANT MODERATE SALIENCE -.037 ( t = -.48) .0984 (t = -L18) fflGH .1545 (t = 1.81) -.043 ( t = -.54) X = SELF-IMPORTANCE ; Y =COMPLEXITY Figure 12. Experiment 3: complexity as a function of salience, self-importance and obj relevance

PAGE 87

77 RELEVANT MODERATE SALIENCE OBJECT RELEVANCE NOT RELEVANT .543 (** t = 4.73) -.039 (t--.31) HIGH 1^' .6947 (**t = 5.43) -.0724 (t = -.62) X = SELF-IMPORTANCE ; Y ^ATTITUDE Figure 13. Experiment 3: attitude as a function of salience, self-importance and object relevance

PAGE 88

CHAPTER 8: GENERAL DISCUSSION AND CONCLUSIONS The research presented here draws upon research in social cognition, functional attitude theory, and social identity theory to organize a unifying framework for how a consumer's social identification is likely to lead to the formation of her attitude. The framework of this research identified four key factors that increase the likelihood that a consumer will form an attitude based on social identification processes. Several experiments investigated this assertion and aggregate results across studies suggest that a social identity pathway to attitude formation has both verifiable theoretical import and well as useful descriptive value. For example, part of the current research examined the role of a social identity's salience on people's spontaneous self-concepts (McGuire et al., 1979. 1981) and how this impacts the formation of their attitudes. It was found that the salience of a particular basis for self-definition ("sons and daughters") mediated people's momentary selfconceptions of themselves. This impacted people's judgments (purchase likelihood) when they were evaluating an object (new product concept) that was relevant to that activated identity. This is consistent with social cognitions' view that "activated concepts" can affect judgments, as well as consistent with more recent findings in the literature concerning the dynamic nature of constructed judgments and retrieved attitudes (Schwarz and Bohner, 2000; Wilson and Hodges, 1992). For example, researchers working in the area of "biased information scanning" found that compared to white 78

PAGE 89

79 females, black females who had recently had their ethnic (female) identity made salient, had more favorable (unfavorable) perceptions of OJ Simpson's innocence (Newman, Duff, Schnopp-Wyatt, Brock and Hoffman, 1997). In the stereotype literature, another recent study demonstrated that when a particular social identity was made salient at an implicit level, performance was altered in the direction predicted by the stereotype associated with the identity. In this study, Shinh, Pittinsky and Ambady (1999) found that Asian American women performed better on a math test when their ethnic identity was activated but worse when their gender identity \was activated as compared with a control group who had neither identity activated. Therefore, a consumer's response to a brand, product or company can be markedly different depending on how they are thinking about themselves at the moment they generate an attitude. In the current research, the data also suggest that the extent to which a consumer identifies with a particular social identity (i.e., the self-importance of the social identity) moderates the extent to which the evaluative content of an adopted social identity will impact attitude formation. When people saw their "family" identity (Baldwin and Holmes, 1987) as relatively more self-important than not, they were more likely to favorably evaluate a product that was relevant to their "family" identity. This finding is in concert with early discussions of "identification" (Kelman 1958) as a basis for adopting an attitude and the rich literature in social psychology that describes motivational underpinnings of social identification in terms of how objects can be seen as symbolically linked to social identities people value (Katz, 1960; Smith, Bruner and White, 1956; Shavitt, 1990). The results presented here are also consistent with recent work in which researchers have argued that psycholoRical "group membership" can

PAGE 90

80 influence people's attitudes. For example, Terry and Hogg (1996) presented evidence that norms linked to a behaviorally relevant reference group ("regular exercisers") influenced intentions to engage in exercise but only for people who identified with the group (Terry and Hogg, 1996; see also Ybarra and Trafimow, 1998). Lastly, evidence presented here suggests that a person's salient and self-important social identity will only impact evaluative responses that are diagnostic to the social identity (c.f. Feldman and Lynch, 1988). In other words, adopting a social identity as one's social identity must provide a person with a meaningful basis to respond to something that they want to judge. This notion is consistent with recent findings in the attitude recruitment literature (c.f Wilson, Lindsey and Schooler, 2000). For example, in a very complex hybrid lab/field study, Cohen and Reed (2000) showed that, outside of any obvious pressure to conform, heightening the salience of a particular social identity ("women in business") affected the way that people recruited attitudes toward the object ("affirmative action"). The results of their study suggest that when identity was salient (not salient) attitudes were more aligned with group membership norms (personal attitudes) when the membership group had a clear (unclear) position on the issue (see also, Kallgren et al., 2000). Therefore, social identification may be salient, self-important and relevant to a product or brand; but may also leave the consumer unable to generate an attitude if that identity does not provide some direction with regard to how that "type" of person will or should respond. Moving beyond a multi-attribute centered focus The point was made earlier that the framework of this dissertation may provide added value above and beyond an object centered multi-attribute approach. This begs the question as to whether the effects and

PAGE 91

81 phenomena described here can be equally formulated, predicted, described and explained by multi-attribute models in general or more sophisticated attitude frameworks such as the classic behavioral intentions model. A multi-attribute model defines an attitude in terms of beliefs and weights. The point that has been made by some researchers is that perhaps the best way to think about a multi-attribute model is not as a psychological process but as a very simple and useful measurement method that can derive a result.' For example, in past discussions and debates, (c.f Miniard and Cohen, 1981) researchers have argued that having a model that incorporates everything precludes a recognition of potential differences among the kinds of components that may make up the person's summary evaluation. For example, it does not matter in a multi-attribute model whether the beliefs have to do with product attributes, or social identity, personal values or some other aspect. This is potentially important for the following reason: If a researcher focuses on product attributes and the salient beliefs linked to them and it turns out that the driving force behind the consumer's attitude has more to do with symbolic associations such as social identity, then it is difficult if not impossible to understand what the process is that led to the attitude. Of course, the response to this line of reasoning is that a multi-attribute model can adequately measure the social identity based beliefs associated with an attitude based on social identification processes. For example, if a consumer believes that some product is for freedom loving people, and the consumer values freedom loving people, then the consumer will have a more favorable attitude toward the product. The caveat is that if This is not a cridcism of multi-attribute models because their purpose is not to describe a psychological process. That might confuse some researchers because multi-attribute models seem to have a process involved in the multiplicafion of evaluative implications times beliefs summed across salient beliefs.

PAGE 92

82 one does not care about the key internal and external factors that may lead a consumer to form their attitude based on a self-conception of themselves as desiring to be or appear to be a freedom loving person, then it is not a problem to use a multi-attribute approach to simply ask people what their attitudes are based on specifying the set of social identity related beliefs. But, if one wants to know when a particular kind of attitudinal influence like social identification is likely to lead to an attitude and to understand how this influence occurs, then one is interested in understanding the process that underlies it. In line with the latter emphasis, this dissertation starts with an interest in social identity as one potential driving mechanism of an attitude formation process In this dissertation research there is no quarrel with the original multi-attribute model. It is designed as a measurement tool, it is definitionally correct and it is certainly acknowledged that the dissertation's empirical work cannot be distinguished from what can be post hoc explained by anv multi-attribute model. Nonetheless, research is more and more coming to understand that non-cognitive influences might be particularly important in certain instances, and it is doubtful whether that is nearly as well represented in various kinds of multi-attribute models (c.f Terry and Hogg, 1996). To the extent that social identity based influences do not rise to the level of beliefs or that they really should be characterized somewhat differently, then these kinds of qualitatively different influencesuch as social identity-might not translate into beliefs very well. An expanded view of social influence With respect to the behavioral intentions model, a very brief discussion is necessary in which the point will be made that the major point of difference with the classic behavioral intentions model is its view of social influence i.e., in treating all social influences as being normative in nature. Heavy use of

PAGE 93

83 the behavioral intentions model has led many consumer researchers to characterize social influence on attitudes as some outside force that stops a person from willingly doing what they would otherwise do. For example, imagine a mother whose son is doing poorly in school because he is spending too much time playing basketball. He does not like school. Besides, he says that he has to practice because he really wants to be like his favorite basketball players in the NBA. However, mom is now angry. She wants her son to do better and to be more favorable toward his studies. Consider two interventions: 1) The mother threatens to punish her child (withhold allowance) if he does not do well in school by maintaining balanced academic and athletic interests. 2) The mother shows her son an article where his favorite basketball players describe the value of maintaining balanced academic and athletic interests. Both interventions may lead to behavioral change, i.e., the son might study more. However, these two types of influence may be thought of quite differently. Normative influence (as might be described by the first intervention and heavily reflected in the behavioral intentions model) affects public expressions (Deutsch and Gerard, 1955). Social identification based influence (as described by the second intervention and reflected in this dissertation's framework) impacts on attitude generation. The only part of the behavioral intentions model that is challenged directly in the implications of this dissertation is the treatment of social influence where social influence is pulled outside of individual attitudes. The kind of social influence described in this dissertation contests the view that a person forms their individual attitude without ever considering social influence and that social influence can only modify the relationship between attitudes and intentions. Such a conceptualization might be considered a particularly narrow view of the importance of a social self This is an

PAGE 94

84 important point for the following reason: A model that defines all social influences as lying outside of a person's individual attitude will only account for social influence processes that are associated with the first example above. That is fine if the focus is on a normative kind of intervention. However, such a conceptualization will be less useful to a researcher whose primary goal is to understand the attitude formation process in order to affect/modify/change an attitude that is based on something other than the evaluative implications of motivational pressures to conform to some referent other. There may be many instances in which it is not enough to know that people tend to superficially modify their behavior in light of complying to some external contingency. Rather, in some instances, a more effective behavioral change strategy might be to lead someone to generate an attitude based on identification with a social identity. Therefore, if one has an a priori theoretical framework that specifies when this type of social influence is likely to come into play, then there may be instances where one may affect attitudes in a more enduring fashion by a clearer understanding of how identification may help to form a person's attitude in the first place. Contributions At any given point in time consumers have a subset of social identities that they may adopt in order to guide their thoughts and actions. These social identities can be very important bases for self-definition and in some instances, be the basis by which a consumer forms her attitude. In fact, academic research has consistently shown that categorizing consumers in terms of various social identities can lead to differences in how that consumer is likely to behave. Intriguingly, marketing practitioners certainly understand the value of targeting consumers, segmenting markets and positioning

PAGE 95

85 products based on the evaluative implications of these social identities. However, the field is in its infancy in understanding the precise nature of how consumers may generate their attitude on the basis of identification with a social identity. Therefore, the current framework presented here may be an important contribution to both attitude theory and marketing practice. Contribution to Attitude Theory. Consumer research has been heavily influenced by multi-attribute conceptualizations of attitude, and rightly so. This very parsimonious way to represent a consumer's evaluation of a product or brand is very useful in determining which product features should be emphasized in marketing strategy. However, some consumer decisions may not be based on a reasoned consideration of product features. Sometimes, a consumer may feel connected with a brand, product or behavior simply because it is "who they are." In fact, the notion that a consumer's sense of who they are should relate to their purchase decisions has been an important idea discussed by several scholars (e.g.. Levy, 1959); whether in terms of overall "congruency" between the self and a brand (Sirgy, 1982), in terms of precise roles that the consumer wishes to enact (Solomon, 1983) or in terms of particular personality associations embodied within the brand itself (Aaker, 1997). This extremely simple premise is not new. Though a few researchers have acknowledged that identification may lead a person to willingly adopt an attitudinal position to "maintain a desired relationship and the self-definition that is anchored in the relationship" (Kelman, 1958, page 62), and fimctional theorists have described social identity motives as reasons why a person might hold a particular attitude, almost no research has systematically examined a social identity based attitude formation process including identifying when and how it is

PAGE 96

86 likely to occur. This research attempts to make two contributions to attitude theory: 1) to provide a better understanding of the role of social identity as an alternative mechanism leading to attitude formation and change, and 2) to provide a more precise understanding of the moderating factors that are likely to determine when social identity will affect consumer attitudes. Contribution to Marketing Practice In terms of marketing strategy, brands and products are often created or positioned to embody a particular social identity oriented lifestyle. Often it appears that marketers are hoping people will look upon such lifestyle presentations favorably and connect them to the advertised brand. The model described in this research should be a useftil guide for developing various kinds of social identity based marketing strategy in several relevant substantive domains. The model uses an expansive, but precise definition of self (i.e., social identity). This allows marketers the flexibility to work with different kinds of social identities in formulating a particular strategy. The model should also contribute to marketing practice by identifying the precise roles likely to be played by salience, self-importance, object relevance and diagnosticity. The model introduced here and its subsequent validation in fixture research programs should be useful to marketers who wish to formulate or evaluate a social identity-based advertisement, positioning strategy, persuasive communication or market segmentation scheme because the model specifies each of the key components of a successftil execution. Finally, the model presented in this dissertation research should allow marketers to emphasize the most cost effective and easily alterable aspects of the proposed process— given a particular social identity that the marketer chooses to work with. In the future, it is hoped that the model can lead to improved insights-some

PAGE 97

87 discussed here— for marketers who wish to connect consumers to their brands or products through the use of social identification processes.

PAGE 98

APPENDIX A VISUAL IMAGERY STUDY UNIVERSITY OF FLORIDA Department of Marketing VISUAL IMAGERY STUDY: PRETEST OF INTERNET TYPEFACE FORIVIAT(S) INTRODUCTION: As you probably well know, the Internet has revolutionized information transmission. A much greater percentage of communication is taking place in "Cyber space." This research project deals with visual imagery and the effectiveness of Internet communication formats. One kind of imagery is connected to both the presentations of format and style That is, not just what we say, but how we say it. This is particularly important both in mass communication (e.g., Web page design) and specialized communications (e.g., interpersonal emails, etc.) One active area of research is the effectiveness of typeface used in electronic communications For example, look at the following types of presentation formats: HAVE A NICE DAY HAVE A NICE DAY HAVE A NICE DAY These presentation formats might convey different meanings. One may be perceived as more "warn & outgoing," another might be perceived as more "cold & impersonal." Your task in this study will be to review a set of typeface formats. Then you will be asked to evaluate each typeface format. Your responses are completely anonymous. To begin, please provide us with the following Administrative information: Last four digits of your student number: Please turn to the Next page > 88

PAGE 99

89 We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and we varied it slightly to slow you down just a little bit. Now, please evaluate the meaning of the word "friend" in terms of "how It comes across to you" in the particular type face presented. To provide us with the most accurate immediate response, read the phrase each time and try to rehearse it in your mind before providing your evaluations: Evaluate the Type face on the following dimensions provided in the box THINK ABOUT THE WORD FRIEND Warm a Impersonal Memorable a a Bland Fun a a Boring Common a Unique PONDER THE WORD FRIEND Warm Impersonal Memorable Bland Fun a a a Boring Common Unique REFLECT ON THE WORD Warm Impersonal Memorable Bland Fun a Boring Common Unique CONSIDER THE WORD FRIEND Warm Impersonal Memorable a Bland Fun a a Boring Common a Unique THINK ABOUT THE WORD FRIEND Warm a a Impersonal Memorable Bland Fun Boring Common a Unique PONDER THE WORD FRIEND Warm Impersonal Memorable Bland Fun Boring Common a Unique REFLECT ON THE WORD FRIEND Please turn to the Next page : Warm Memorable Fun Common Impersonal Bland Boring Unique

PAGE 100

90 We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and we varied it slightly to slow you down just a little bit. Now, please evaluate the meaning of the word "friend" in terms of "how it comes across to you" in the particular type face presented. To provide us with the most accurate immediate response, read the phrase each time and try to rehearse it in your mind before providing your evaluations: Evaluate the Type face on the following dimensions provided in the box THINK ABOUT THE IMPORTANCE OF BEING A FRIEND Warm Memorable Fun Common a a a a Impersonal Bland Boring Unique CONSIDER THE VALUE OF BEING A FRIEND Warm Memorable Fun Common a a Impersonal Bland Boring Unique PONDER THE SIGNIFICANCE OF BEING A Warm Memorable Fun Common Impersonal Bland Boring Unique THINK ABOUT THE VALUE OF BEING A FRIEND Warm Memorable Fun Common a a Impersonal Bland Boring Unique PONDER THE IMPORTANCE OF BEING A FRIEND Warm Memorable Fun Common Impersonal Bland Boring Unique THINK ABOUT THE IMPORTANCE OF BEING A FRIEND Warm Memorable Fun Common a a Impersonal Bland Boring Unique CONSIDER THE VALUE OF FRIEND Warm impersonal BEING A Memorable Bland Fun Boring Common Unique Please turn to the Next page >

PAGE 101

91 We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and we varied it slightly to slow you down just a little bit. Now, please evaluate the meaning of the word "friend" in terms of "how It comes across to you" in the particular type face presented. To provide us with the most accurate immediate response, read the phrase each time and try to rehearse it in your mind before providing your evaluations: Evaluate the Type face on the following dimensions provided in the box THINK ABOUT THE WORD SON Warm a a a Impersonal Memorable a a a Bland Fun a a a Boring Common a a Unique PONDER THE WORD SON Warm a Impersonal Memorable Bland Fun a Boring Common a a Unique REFLECT ON THE WORD SOX Warm Impersonal Memorable Bland Fun a Boring Common a Unique CONSIDER THE WORD SON Warm a Impersonal Memorable a Bland Fun a Boring Common a Unique THINK ABOUT THE WORD SON Warm a Impersonal Memorable Bland Fun a Boring Common Unique PONDER THE WORD SON Warm a Impersonal Memorable a Bland Fun a Boring Common a a a a Unique REFLECT ON THE WORD SON Please turn to the Next page : Warm Memorable Fun Common a a Impersonal Bland Boring Unique

PAGE 102

92 We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and we varied it slightly to slow you down just a little bit. Now, please evaluate the meaning of the word "friend" in terms of "how it comes across to you" in the particular type face presented. To provide us with the most accurate immediate response, read the phrase each time and try to rehearse it in your mind before providing your evaluations: Evaluate the Type face on the following dimensions provided in the box THINK ABOUT THE IMPORTANCE OF BEING A SON Warm Memorable Fun Common a a a a Impersonal Bland Boring Unique CONSIDER THE VALUE OF BEING A SON Warm Memorable Fun Common a a a a Impersonal Bland Boring Unique PONDER THE SIGNIFICANCE OF BEING A SONWarm Memorable Fun Common a Impersonal Bland Boring Unique THINK ABOUT THE VALUE OF BEING A SON Warm Memorable Fun Common a a a a a Impersonal Bland Boring Unique PONDER THE IMPORTANCE OF BEING A SON Warm Memorable Fun Common a a a a a Impersonal Bland Boring Unique THINK ABOUT THE IMPORTANCE OF BEING A SON CONSIDER THE VALUE OF BEING A SON Please turn to the Next page Warm Memorable Fun Common Warm Memorable Fun Common Impersonal Bland Boring Unique Impersonal Bland Boring Unique

PAGE 103

93 We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and we varied it slightly to slow you down just a little bit. Now, please evaluate the meaning of the word "friend" in terms of "how it comes across to you" in the particular type face presented. To provide us with the most accurate immediate response, read the phrase each time and try to rehearse it in your mind before providing your evaluations: Evaluate the Type face on the following dimensions provided in the box THINK ABOUT THE WORD Daughter Warm a a Impersonal Memorable a Bland Fun a Boring Common a a Unique PONDER THE WORD Daughter Warm Impersonal Memorable Bland Fun a a Boring Common a a Unique REFLECT ON THE WORD 'DaugRter Warm a Impersonal Memorable Bland Fun Boring Common a Unique CONSIDER THE WORD Daughter Warm Impersonal Memorable a a a Bland Fun Boring Common Unique THINK ABOUT THE WORD Daughter Warm a a Impersonal Memorable a Bland Fun a Boring Common Unique PONDER THE WORD Daughter Warm a a a Impersonal Memorable a Bland Fun a Boring Common a a Unique REFLECT ON THE WORD DAUGHTER Please turn to the Next page : Warm Memorable Fun Common Impersonal Bland Boring Unique

PAGE 104

94 We would like you to do this exercise fairly rapidly, but we want you to carry out the task in a reasonably careful manner. In order to get you to thoughtfully evaluate each typeface, we used a lead in presented in a neutral font and we varied it slightly to slow you down just a little bit. Now, please evaluate the meaning of the word "friend" in terms of "how it comes across to you" in the particular type face presented. To provide us with the most accurate immediate response, read the phrase each time and try to rehearse it in your mind before providing your evaluations: Evaluate the Type face on the following dimensions provided in the box THINK ABOUT THE Daughter Warm Impersonal IMPORTANCE OF BEING A Memorable a Bland Fun a Boring Common a a a Unique CONSIDER THE VALUE OF Daughter Warm Impersonal BEING A Memorable Bland Fun Boring Common Unique PONDER THE SIGNIFICANCE Oaugkter Warm a impersonal OF BEING A Memorable Bland Fun Boring Common a Unique THINK ABOUT THE VALUE OF Daughter Warm a Impersonal BEING A Memorable a a Bland Fun Boring Common a Unique PONDER THE IMPORTANCE OF Daughter Warm a Impersonal BEING A Memorable Bland Fun Boring Common a Unique THINK ABOUT THE Daughter Warm Impersonal IMPORTANCE OF BEING A Memorable a Bland Fun Boring Common Unique CONSIDER THE VALUE OF DAUGHTER Warm Impersonal BEING A Memorable Bland Fun Boring Common Unique Please turn to the Next page > ~ ~

PAGE 105

95 Follow up Questions: Though the evaluations that you gave should mostly be affected by the typeface, it is possible that the use of actual words may have made a difference, i.e., if would have used other words, you might have responded differently. Therefore, in order to statistically control for the impact of a particular word, we need to obtain some evaluations of the word itself. Think about the word FRIEND standing by itself and not in any particular typeface: Now, answer the following questions: When I think about myself, I primarily think about myself as a: PRIMARILY A DAUGHTER 1 2 3 4 5 6 7 PRIMARILY A FRIEND EQUALLY BOTH How well does the term "friend" describe you? DOES NOT DESCRIBE ME 1 2 3 4 5 6 7 DESCRIBES ME PERFECTLY Indicate the extent to which you identify with your "friends". DO NOT IDENTIFY IN 1 2 3 4 5 6 7 STRONGLY IDENTIFY ANYWAY More generally speaking, how important do you consider your role in life as a friend to you peers? NOT IMPORTANT 1 2 3 4 5 6 7 VERY IMPORTANT Now, please answer the following general questions: Do you often notice different typefaces in your day to day electronic communications? YES NO Do you think typeface place a significant role in the attractiveness of electronic communications? YES NO THANK YOU FOR YOUR PARTICIPATION

PAGE 106

96 Follow up Questions: Though the evaluations that you gave should mostly be affected by the typeface, it is possible that the use of actual words may have made a difference, i.e., if would have used other words, you might have responded differently. Therefore, in order to statistically control for the impact of a particular word, we need to obtain some evaluations of the word itself. Think about the word SON standing by itself and not in any particular typeface: Now, answer the following questions: When I think about myself, I primarily think about myself as a: PRIMARILY A SON 1 2 3 4 5 6 7 PRIMARILY A FRIEND EQUALLY BOTH How well does the term "son" describe you? DOES NOT 1 2 3 4 5 6 7 DESCRIBES ME PERFECTLY DESCRIBE ME Indicate the extent to which you identify with your (parents). DO NOT IDENTIFY IN 1 2 3 4 5 6 7 STRONGLY IDENTIFY ANYWAY ^:,v More generally speaking, how important do you consider your role in life as a "son"? NOT IMPORTANT 1 2 3 4 5 6 7 VERY IMPORTANT Now, please answer the following general questions: Do you often notice different typefaces in your day to day electronic communications? YES NO '•' Do you think typeface place a significant role in the attractiveness of electronic communications? YES NO THANK YOU FOR YOUR PARTICIPATION

PAGE 107

97 Follow up Questions: Though the evaluations that you gave should mostly be affected by the typeface, it is possible that the use of actual words may have made a difference, i.e., if would have used other words, you might have responded differently. Therefore, in order to statistically control for the impact of a particular word, we need to obtain some evaluations of the word itself. Think about the word DAUGHTER standing by itself and not in any particular typeface: Now, answer the following questions: When I think about myself, I primarily think about myself as a: PRIMARILY A 1 2 3 4 5 6 DAUGHTER EQUALLY BOTH How well does the term "daughter" describe you? DOES NOT 1 2 3 4 5 6 7 DESCRIBES ME PERFECTLY DESCRIBE ME Indicate the extent to which you identify with your parents. DO NOT IDENTIFY 1 2 3 4 5 6 7 STRONGLY IDENTIFY IN ANYWAY More generally speaking, how important do you consider your role in life as a "daughter"? NOT IMPORTANT 1 2 3 4 5 6 7 VERY IMPORTANT Now, please answer the following general questions: Do you often notice different typefaces in your day to day electronic communications? YES NO Do you think typeface place a significant role in the attractiveness of electronic communications? YES NO -i 7 PRIMARILY A FRIEND THANK YOU FOR YOUR PARTICIPATION

PAGE 108

APPENDIX B WORD SEARCH STUDY UNIVERSITY OF P FLORIDA WORD SEARCH GAME(S) DON'T FLIP AHEAD TO THE NEXT PAGE UNTIL YOU HAVE READ ALL THE DIRECTIONS! We would appreciate your assistance on the following task. We are running a study on people's familiarity with word search games. On the following page is a word search grid that contains 18 words. We would like for you to take approximately TEN MINUTES to find as many words as possible. HOWEVER, WHEN YOU FIND A WORD, CIRCLE IT, LIKE THIS, AND THEN WRITE THE WORD BELOW ON ONE OF THE NUMBERED LINES (See below). WHAT'S IMPORTANT IS THAT YOU WRITE DOWN THE WORDS IN THE ORDER THAT YOU FIND (CIRCLE) THEM IN THE GRID. Continue this exercise for APPROXIMATELY TEN MINUTES. Don't cheat!!! Just try to find as many words as you can. After TEN minutes, stop the exercise • PLEASE DON'T CHEAT. • THIS IS NOT A TEST OF YOUR ABILITIES IN ANY WAY. • THE DATA WILL ONLY BE USEFUL IF YOU CAREFULLY FOLLOW DIRECTIONS. PLEASE TURN TO THE NEXT PAGE > 98

PAGE 109

99 Take a few seconds to peruse the grid. Then begin. When you find a word, circle it and then write it on one of the lines ON THE NEXT PAGE. CONTINUE THIS FOR APPROXIMATELY TEN MINUTES! THEN STOP! T 0 R N A D 0 A Z P D C H A 1 R B N E F A B S E R M R E S E A R C H J H E L H Y B Z X E T Y U H K E A S N M M U P A Q Q L D T 1 F R I E N D X K J K C P V E Z I E 0 K J T C V G I' N O C A H A V P L H Y X Z A E R D p R G H N S 0 N B u R G E R M 1 Y E S D A C G H X c 0 M P u T E R Z A c V N Y G P U J J J H E 0 E Y F 0 G X X R V 1 H M E M G U J 0 K J W N X U 0 P H 1 p P 0 Z R T H H N M 0 L M P S A N Y X C Y O H N I K L Z T X C V B N M P 0 A Q w E L R T Y U B A K E R Y 1 0 P L M K L K J 1 M N B V C S Z A R Z Q W E R 1 D Q A Z C W S F X E D c R F V T G B Y C H N U J E M 1 K R P 0 K J F L 0 R 1 D A Z Q Z A 0 W X S E 1 C D R V F R T B G Y N H P L F O M U J N E Y H B T A G D H W A A G D F L H J K Y 0 N H L J N Z C X P Q S A Z 1 H L J 0 A M H D J C G A Z D A z CONTINUE FOR ONLY TEN MINUTES!!!!! 17. ~2. ~ ~ "9^ 4. ~ la ~rr "6! ~l2. ~ STOP AFTER TEN MINUTES!!!!! THANK YOU FOR YOUR PARTICIPATION. 14. 15. 16. 17. W

PAGE 110

APPENDIX C CONSUMER SELF DESCRIPTION STUDY \ UNIVERSITY OF ^FLORIDA Department of Marketing SELF DESCRIPTION STUDY Self Description Study In your own words, please take a minute to describe yourself. Provide as much detail as you wish. Thank you for your participation. Student ID Number: 100

PAGE 111

APPENDIX D NEW PRODUCT ASSESSMENT STUDY University of Florida NEW PRODUCT ASSESSMENT STUDY 13 D D REA & R\RKER NIELSEN MEDIA [RESEARCH INSTRUCTIONS: Millions of dollars are spent on the development of new products. Therefore, it is critical that marketing analysts have a good idea of the likelihood of new product success. This way, they can weed out the "duds" from the "stars" and not waste resources. The University of Florida in conjunction with several marketing research firms is sponsoring this study on perceptions of a series of new product ideas. 1 Your job will be to examine each new product concept. 2. You will be provided with both 1) a picture of the product & 2) a description of the product. 3. You will then be asked to rate the product on a series of dimensions. You will also be asked to provide some comments / feedback regarding the products. ALL of your responses will be completely anonymous. Please be sure to read all information carefully. To not over burden you with too many evaluations, you will only evaluate four new product ideas in this study. For administrative purposes, please start by giving us the following background information: Last four digits of your student number: You will be shown information on the new product ideas. You will then be asked to answer a few questions regarding each product. Before you begin, remember that we appreciate your time in participating in this research. For this data to be useful, you must take your time and treat this seriously. Remember: YOUR OPINIONS ARE VERY IMPORTANT TO US AND ARE THE BASIS FOR THIS RESEARCH. Go to the next page. 101

PAGE 112

102 Hold instant access to all your important business information in the palm of your hand: ELECTRONIC COMPANION: Model BUS-3 Calendar, address book, to-do work list, expenses, business e-mail. Exchange synchronized data with your PC at the touch of a button, eliminating dual data-entry and ensuring the safety and accuracy of information. Plus, you can beam data to other Palm platform devices using the infrared port. Unique user-expandable design lets you add a modem, memory, software and more. Advanced LCD screen has razor-sharp clarity. The long-life lithium ion batteries recharge in just minutes using the HotSync cradle. Enjoy up to a month's operation on the road. Compatible with Windows 3.1, 95, NT 4.0 (optional adapter required for Mac). Modem operates on 2 AAA batteries (included). Measures 4.5" x 3. 1" x .4". 1) Assuming that the price was right, what is the likelihood that you would purchase this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 2) How likely is the Market success of this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 3) Off the top of your head and in your own words, what's the best reason you can give for someone buying this? 5) Think more about this. Are there anymore reasons? GO to the Next Page

PAGE 113

103 Reaching out to your Friend is now in the palm ofyour harid. ELECTRONIC COMPANION: Model SN-3 Calendar, address book, to-do list, expenses, e-mail. More importantly, exchange information and well wishes with your loved ones at the touch of a button, keeping you connected with the ones that you care about. This new product allows you to never miss important Friend events and/ or occasions that are dear to your heart. In fact, it will automatically remind you of them, and if you like, automatically send messages, flowers and e-cards. Custom messages can be programmed ahead of time and sent automatically. Enjoy up to a month's operation on the road. Compatible with Windows 3. 1, 95, NT 4.0 (optional adapter required for Mac). Modem operates on 2 AAA batteries (included). Measures 4.5" x 3. 1 x .4". 1) Assuming that the price was right, what is the likelihood that you would purchase this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 2) How likely is the Market success of this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 3) Off the top of your head and in your own words, what's the best reason you can give for someone buying this? 5) Think more about this. Are there anymore reasons?

PAGE 114

104 Reaching out to your friends is now in the palm of your hand. ELECTRONIC COMPANION: Model FR-3 Calendar, address book, to-do list, expenses, e-mail. More importantly, exchange information and well wishes with your close friends at the touch of a button, keeping you connected your buddies, pals and comrades. This new product allows you to never miss important events with friends and/ or occasions that are dear to your heart. In fact, it will automatically remind you of them, and if you like, automatically send messages, flowers and e-cards. Custom messages can be programmed ahead of time and sent automatically. Enjoy up to a month's operation on the road. Compatible with Windows 3. 1, 95, NT 4.0 (optional adapter required for Mac). Modem operates on 2 AAA batteries (included). Measures 4.5" x 3.1" x .4". 1) Assuming that the price was right, what is the likelihood that you would purchase this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 2) How likely is the Market success of this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 3) Off the top of your head and in your own words, what's the best reason you can give for someone buying this? 5) Think more about this. Are there anymore reasons? GO to the Next Page

PAGE 115

105 Reaching out to your Friend is now in the palm of your hand. ELECTRONIC COMPANION: Model (SN) DTR-3 Calendar, address book, to-do list, expenses, e-mail. More importantly, exchange information and wfell wisjles with your loved ones at the touch of a button, keeping you connected with your parents and loved ones. This new product allows you to never miss important family events and/ or occasions that are dear to your heart. In fact, it will automatically remind you of them, and if you like, automatically send messages, flowers and e-cards. Custom messages can be programmed ahead of time and sent automatically. Enjoy up to a month's operation on the road. Compatible with Windows 3.1, 95, NT 4.0 (optional adapter required for Mac). Modem operates on 2 AAA batteries (included). Measures 4.5" x 3.1" x .4". 1) Assuming that the price was right, what is the likelihood that you would purchase this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 2) How likely is the Market success of this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 3) Off the top of your head and in your own words, what's the best reason you can give for someone buying this? 5) Think more about this. Are there anymore reasons? GO to the Next Page

PAGE 116

106 Keeping Cool is now an easy thing to do! PERSONAL COOLING SYSTEM When it's blazingly hot, don't you dread leaving the air-conditioned comfort of your home? Maybe you've tried an ice-soaked towel around your neck — but that little bit of evaporative cooling is nothing compared to the long-lasting, total-body comfort of our all-new 2.0 version of the Personal Cooling System. It's the best way to stay comfortable — and avoid heat stress — when walking, mowing, gardening, watching sports or concerts.. .just about any activity, anywhere. This new and improved 2.0 model offers two levels of cooling: "regular" helps you feel up to 20F cooler than the hot, dry air around you; and "high" now offers 20 percent greater cooling — for even the hottest days. What's more, the tiny fan inside version 2.0 adds even more cooling by directing a gentle breeze over the back of your neck. The 2.0 model also is adjustable to fit a wider range of neck sizes comfortably and securely. It's trimmer and lighter, too: the sleek, die-cast aluminum body weighs just 10 oz. empty; 14 oz. filled. Runs on one AA battery (order separately). 1) Assuming that the price was right, what is the likelihood that you would purchase this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 2) How likely is the Market success of this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 3) Off the top of your head and in your own words, what's the best reason you can give for someone buying this? 5) Think more about this. Are there anymore reasons? GO to the Next Page

PAGE 117

107 Never get stuck on the road again. AIRMAN DELUXE AIR PUMP The next time you get a flat tire, don't panic. Instead, reach for our cordless, highpressure air pump. Operating at the touch of a button, it inflates a standard aato tire from completely flat to 35 P.S.J, in less than eight minutes. Unlike awkward hand pumps, it's powered by a rechargeable 9.6-volt NiCad battery that can inflate up to four tires on a single charge. Works up to six months after its last charge, but if you find yourself low on power, just use the included 1 2volt adapter to plug the unit into your car's cigarette lighter. Also ideal for homes filled with bikes, scooters and every kind of inflatable game ball. Equally adept at inflating everything from car and bike tires to basketballs and balloons. Three adapter nozzles are included for different types of valves. This super-fast air compressor allows you to inflate up to 150 P.S.I. An automatic pressure regulator ensures proper inflation. Practically effortless to use, the pump weighs less than three pounds, and measures just 2 3/4" x 3 1/4" x 8 1/4". 1) Assuming that the price was right, what is the likelihood that you would purchase this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 2) How likely is the Market success of this product? Unlikely 1 2 3 4 5 6 7 Highly Likely 3) Off the top of your head and in your o*n words, what's the best reason you can give for someone buying this? 5) Think more about this. Are there anymore reasons? THANK YOU FOR YOUR PARTICIPATION

PAGE 118

APPENDIX E HANDWRITING STUDY /^kl UNIVERSITY OF W FLORIDA DEPARTMENT OF MARKETING HANDWRITING STYLE & CONSUMER TENDENCIES Consider the following two handwriting samples: The sample on the left seems quite "cramped" while the sample on the right seems a bit more "spacious." Some research suggests that a cramped handwriting style could be positively related to thriftiness (i.e., high concern with conservative use of money) while a more spacious style could be related to "extravagance." Additional speculation exists for other relationships between handwriting style and consumer tendencies. For example, some research indicates that people's handwriting varies depending on such things as mood and emotionality. This research is designed to more carefully evaluate several of these types of predictions. More specifically, this research looks at two topics. 1. Whether HOW we write (handwriting style) is related to certain consumer tendencies. 2. Whether even small differences in mood can be seen in handwriting. Therefore, in collaboration with the department of psychology, the University of Florida, Department of Marketing is sponsoring this short study on handwriting. You will be asked to engage in a simple handwriting task The information that you provide us will be strictly anonymous. None of it can be linked back to you. The tasks that you engage in are by no means evaluative with respect to your intelligence or abilities. If you feel that you do not want to participate, you may withdraw at any time. Please Go to the Next Page> 108

PAGE 119

109 HAKD WRITING SAMPLE (BASELINE) : For this research to be meaningful, we need a baseline to compare your handwriting. So to begin, write the following 3 neutral sentences on the left in your own natural handwriting (cursive or not, depending on whatever style you use most frequently) : Write the sentence down Here: George Washington was the first president of the U.S. Potatoes are considered vegetables. The apple does not fall far from the tree Please turn to the next page >

PAGE 120

110 As we discussed earlier, part of this research looks at whether or not WHAT people write about is related to their handwriting style. We want to assess if people's handwriting varies as to whether they are writing about things that may have different emotionality because of the topic. In the interest of full disclosure, different people participating will provide handwriting samples under different mindsets and about different topics. You have been assigned to a particular condition, which may induce a slight change in emotionality, but nothing major. PLEASE FOLLOW THE INSTRUCTIONS BELOW: We'd like you to write 4 independent statements (complete sentences in your usual handwriting style, i.e., cursive or not) each with about 10 to 15 words The statements should convey some positive, deeply moving, emotionally involving thoughts and or sentiments that describe the commitment to your relationship with one or both of your parents. Concentrate on how you have maintained your family ties while also STRENGTHENING a sense of independence as an individual young adult. 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 4 5. Please turn to the next page>

PAGE 121

Ill As we discussed earlier, part of this research looks at whether or not WHAT people write about is related to their handwriting style. We want to assess if people's handwriting varies as to whether they are writing about things that may have different emotionality because of the topic. In the interest of full disclosure, different people participating will provide handwriting samples under different mindsets and about different topics. You have been assigned to a particular condition, which may induce a slight change in emotionality, but nothing major. PLEASE FOLLOW THE INSTRUCTIONS BELOW: We'd like you to write 4 independent statements (complete sentences in your usual handwriting style, i.e., cursive or not) each with about 10 to 15 words The statements should convey some positive, deeply moving, emotionally involving thoughts and or sentiments that describe the commitment to your relationship with one or both of your parents Concentrate on how you have maintained your feunily ties while also STRENGHTHENING the sense of connectedness as a member of your family. 1. 2. 5. 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Please turn to the next page>

PAGE 122

112 As we discussed earlier, part of this research looks at whether or not WHAT people write about is related to their handwriting style. We want to assess if people's handwriting varies as to whether they are writing about things that may have different emotionality because of the topic. In the interest of full disclosure, different people participating will provide handwriting samples under different mindsets and about different topics. You have been assigned to a particular condition, which may induce a slight change in emotionality, but nothing major. PLEASE FOLLOW THE INSTRUCTIONS BELOW: We'd like you to write 4 independent statements (complete sentences in your usual handwriting style, i.e., cursive or not) each with about 10 to 15 words The statements should convey some positive, deeply moving, emotionally involving thoughts and or sentiments that describe the commitment to your relationship with your closest friends. Concentrate on how you have maintained your ties with your group of friends while also STRENGTHENING a sense of independence as an individual young adult. 1. 3. 5. 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 Please turn to the next page>

PAGE 123

113 As we discussed earlier, part of this research looks at whether or not WHAT people write about is related to their handwriting style. We want to assess if people's handwriting varies as to whether they are writing about things that may have different emotionality because of the topic. In the interest of full disclosure, different people participating will provide handwriting samples under different mindsets and about different topics. You have been assigned to a particular condition, which may induce a slight change in emotionality, but nothing major. PLEASE FOLLOW THE INSTRUCTIONS BELOW: We'd like you to write 4 independent statements (complete sentences in your usual handwriting style, i.e., cursive or not) each with about 10 to 15 words The statements should convey some positive, deeply moving, emotionally involving thoughts and or sentiments that describe the commitment to your relationship with your closest friends. Concentrate on how you have maintained your ties with your group of friends while also STRENGTHENING the sense of connectedness as a member of your group of friends 1. 2. Please turn to the next page>

PAGE 124

114 PART II. To the right of each of the five boxes that you provided handwriting samples, there is a scale that looks like this: EXAMPLE SCALE 1 2 3 4 5 The scales on page three have no endpoints. But, we would like for you to go back to each of the five statements that you wrote down and to rate the extent to which EACH STATEMENT reflects a 1=a sense of strong individuality and independence as a young adult or 5 = a sense of strong group interdependence on your family. Therefore, please go back and circle one number on the scale using the following endpoints: Individuality, Group, Interdependence on Independence as a young your family adult 12 3 4 5 Make sure that you RATE EACH statement that you provided!!! When you are done rating each statement of your handwriting sample, please go to the next page in the survey.

PAGE 125

115 PART II. To the right of each of the five boxes that you provided handwriting samples, there is a scale that looks like this: EXAMPLE SCALE 1 2 3 4 5 The scales on page three have no endpoints. But, we would like for you to go back to each of the five statements that you wrote down and to rate the extent to which EACH STATEMENT reflects a 1=a sense of strong individuality and independence as a young adult or 5 = a sense of strong group interdependence on your friends. Therefore, please go back and circle one number on the scale using the following endpoints: Individuality, Group, Interdependence on Independence as a young your friends adult 12 3 4 5 Make sure that you RATE EACH statement that you provided!!! When you are done rating each statement of your handwriting sample, please go to the next page in the survey.

PAGE 126

116 Part IV: Lastly are some personality variables that assess a consumer's "personal style." Answering these questions will allow us to discover if there are particular relationships between a person's handwriting style and their own self-perceptions. There are no "right" answers. Please answer each response as truthfully as you can. I love to spend lots of money (when I have it) O True O False I very rarely do any "impulse" buying O True O False I work hard to save money whenever I can o True o False Material things are not that important to me o True o False I enjoy watching television o True o False I would probably make a good actor o True o False In a group of people, I am rarely the center of attention o True o False I consider myself to have VERY high self esteem o True o False I am not oarticularlv aood m;=^ki r\n ot'her Df^nril p 1 i kp ttip True False I'm not always the person I appear to be True \j False I would not change my opinions (or the way I do things) in order to please someone else or win their favor — i L Uc Fa 1 s e T hirivp con ^ i H p TPfi Hpinrr pnt"P"r"t" ;=r i no?" J. i 1 t-i V ^ ^ — w 1 1 ^ v_t c x~ v3 kJ Jw" ^^ii^^ dil C J IC JLd -L 1 1 12 -LTrue False I have never been go.od at games like charades or improvisational acting o True o False It's very important to own a VERY nice car o True o False At a party I let others keep the jokes and stories going o True o False I feel a bit awkward in the company and do not show up quite so well as I should o True o False I can look anyone in the eye and tell a lie with a straight facie (if for the right end) o True o False I may deceive people by being friendly when I really dislike them o True o False Thank you for participating in this research!

PAGE 127

117 PART III. We would like for you to provide us with the following general information. This will help us control for individual differences. Please answer the following questions : The term "family" below refers to you and your relationship with your immediate family (i.e., your mother, father, and siblings.) Strougty Strongly Agree Agree Neutral Disagree Disagree 1. It makes me feel very good to be a member of my family. 5 4 3 2 2. I NEVER try to think like a typical member of my Family. 5 4 3 2 3. I strongly identify with my Family. 543 2 4. Jt is very easy for me to see the world in terms of my 5 4 3 2 Family, 5. Being a member of my Family is a very important part of 5 4 3 2 who I am, 6. When an issue is unclear to me, 1 otten adopt the perspective 5 4 3 2 that 1 think other members of my Family would have. 7. 1 NEVER think about the world in terms of how it relates to 5 4 3 2 my Family. 8. I am extremely proud of my Family. 543 2 9. All of what makes me who I am is my Family. 543 2 10. Beingamemberofmy Family makes me feel like I share a 5 4 3 2 common goal with others. 11. It is not important that other people know about my Family. 543 2 12. When I am around other Family members, I feel very good 543 2 about myself 13. My Family identity makes me feel very bad about myself 5 4 3 2 14. Being a member of my Family often affects how I tend to 543 2 view the world around me. 1 5. A big part of my emotional well being is tied to being a 5 4 3 2 member of my Family. When I think about myself, I primarily think about myself as a: PRIMARILY A FAMILY 12 3 4 5 6 7 PRIMARILY AN INDIVIDUAL MEMBER EQUALLY BOTH Please turn to the Next page :

PAGE 128

118 PART III. We would like for you to provide us wittl the following general inf omation This will help us control for individual differences Please answer the following questions : The term "friend" below refers to you and your relationship with your friends (i.e., the group of people that y your mother, father, and siblings.) Strongly Agree 16. It makes me feel very good to be a Friend. 5 17. I NEVER try to think like my group of close Friends. 5 18. I strongly identify with my Friends. 5 19. it is very easy for me to see the world in terms of my 5 Friends. 20. Being a member of my group of Friends is a very important 5 part of who I am. 2 1 When an issue is unclear to me, 1 often adopt the perspective 5 that 1 think other members of my group of Friends would have. 22. 1 NEVER think about the world in terms ofhow it relates to 5 my Friends. 23. I am extremely proud of my Friends. 5 24. All of what makes me who I am is my group of Friends. 5 25. Being a member of my group of Friends makes me feel like 5 1 share a common goal with others. 26. ft is not important that other people know about my Friends, 5 27. When I am around my group of Friends, I feel very good 5 about myself 28. Being a member of my group of Friends makes me feel very 5 bad about myself 29. Being a member of my group of Friends often afl'ects how I 5 tend to view the world around me. 30. A big part of my emotional well being is tied to being a 5 member of my group of Friends. Agree 4 Neutral 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 Disagree 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Strongly Disagree When I think about myself, I primarily think about myself as a: 12 3 4 5 6 7 PRIMARILY AN INDIVIDUAL PRIMARILY A MEMBER OF MY GROUP OF FRIENDS 2 3 4 5 6 EQUALLY BOTH Please turn to the Next page >

PAGE 129

APPENDIX F THE PANAS MOOD SCALE CONSUMER EMOTION & MOOD STUDY We are pre-testing the following scale. It consists of a number of words that describe different feelings and emotions. Read each item and then mark the appropriate answer in the space next to that word. Indicate to what extent you feel this way right now that is, at the present moment. Use the following scale to record your answers. very slightly or not at all a little moderately 3 quite a bit 4 extremely interested distressed excited upset strong guilty scared hostile enthusiastic proud irritable alert ashamed inspired nervous determined attentive jittery active afraid THANK YOU FOR YOUR PARTICIPATION 119

PAGE 130

APPENDIX G WEB PAGE DESIGN STUDY student Mvunber Sponsored by CONSUMER WEB-PAGE ASSESSMENT STUDY ReA & R\RKER D D NIELSEN MEDIA RESEARCH INSTRUCTIONS: The World Wide Web has changed the face of Marketing. Companies can now sell stuff through E-conunerce by offering consumer products on Internet Web sites. However, success depends on being able to attract consumers to the electronic market place and to keep them interested long enough to make purchases. This is where Electronic Web Page Design becomes critically important. One of the things that Web page designers need to know is the best way to structure sites. Web pages need to make the most impact. There is not a lot of time to grab consumer's attention. Web sites need to be appealing and creative Before money is spent on designing a site, advertisers often pre-test possible site design and content attractiveness to get initial attitudes. We are testing about 150 example Web Page designs. One of these web pages has been randomly chosen for you to check out. Only looking at one will save you time and keep you from getting bored, which could affect the results. You will be asked to look at the one short, example web page. You will be asked for your opinion(s) regarding this example Web page. Please read the questions carefully. Think thoughtfully. Answer each question honestly, and accurately. Remember that there are no right answers. This is exploratory research and we are only interested in your opinions. Your opinions are completely anonymous and cannot be linked back to you. Turn the page now for further instructions. 120

PAGE 131

121 BACKGROUND INFORMATION: Internet E-commerce (Web site shopping) is growing quicicly. Companies who do not have a lot of financial resources can still take advantage of the Internet's ability to hook them up with consumers. Currently, Internet shopping is not as popular compared to shopping at regular stores. This is probably because it is so new. Very little is understood about how different types of consumers respond to shopping on the Internet and Web pages in general. Therefore, part of what is interesting to people is to understand how different groups of consumers respond to electronic media. Consumers are often "segmented" into different groups that are thought to have different likes and dislikes. This makes marketing easier. If a group of people share the same characteristics that are related to their shopping behavior, then a marketer can better serve needs by understanding the characteristics of those consumers. In fact, this ongoing research has already been conducted on three such segments. Each segment is described below. In order to grasp the full scope of this research, please carefully read about each consumer segment below. When you are finished turn to the next page and follow the instructions. This research has been conducted on an older segment conveniently entitled "Levittown, USA." This segment makes up about 3.1% of U.S. households. The Levittown, USA segment earns on average about $28,742. Members of the Levittown USA segment typically range from ages 55 and above. They are usually white couples who live in post war tract subdivisions. They are likely to read magazines like stereo review and watch shows like. Sale of the Century. Sample Zip codes include Cuyahoga Falls, Ohio 4421; Nashville, Tenn. 37214; Cheswick, Pa. 15024. This research has a been conducted on a segment entitled "Middle America." This segment makes up about 3.2% of the U.S. households. The Middle America segment earns on average about $24,43 1 Members of the Middle America segment are high school educated and are more likely to belong to Christmas clubs. They are likely to read magazines like the Saturday Evening Post and watch shows like Family Ties. Sample Zip codes include Marshall, Michigan 49068; Hagerstown, Md. 21740; This research has also been conducted on a younger segment conveniently entitled "Towns and Gowns." This segment makes up about 1.2% of U.S. households. The towns and gowns segment earns on average about $17,862. Members of the Towns and Gowns segment typically range from ages 1 8 to 34. They have been described as college educated, and highly likely to be the future leaders of this country. They are likely to read magazines like Natural History and watch television shows like Good Morning America. Sample ZIP codes include College Station, Texas 77840; Bloomington, Indiana 47401; Ithaca, New York 14850 and Ft. Lauderdale, Florida 33324. Please turn to the next page,

PAGE 132

122 BACKGROUND INFOR^UTION: Internet E-commerce (Web site shopping) is growing quickly. Companies who do not have a lot of financial resources can still take advantage of the Internet's ability to hook them up with consumers. Currently, Internet shopping is not as popular compared to shopping at regular stores. This is probably because it is so new. Very little is understood about how different types of consumers respond to shopping on the Internet and Web pages in general. Therefore, part of what is interesting to people is to understand how different groups of consumers respond to electronic media. Consumers are often "segmented" into different groups that are thought to have different likes and dislikes. This makes marketing easier. If a group of people share the same characteristics that are related to their shopping behavior, then a marketer can better serve needs by understanding the characteristics of those consumers. In fact, this ongoing research has already been conducted on three such segments. Each segment is described below. In order to grasp the full scope of this research, please carefully read about each consumer segment below. When you are finished turn to the next page and follow the instructions. This research has been conducted on an older segment conveniently entitled "Levittown, USA." This segment makes up about 3. 1% of U.S. households. The Levittown, USA segment earns on average about $28,742. Members of the Levittown USA segment typically range from ages 55 and above. They are usually white couples who live in post war tract subdivisions. They are likely to read magazines like stereo review and watch shows like. Sale of the Century. Sample Zip codes include Cuyahoga Falls, Ohio 4421; Nashville, Tenn. 37214; Cheswick, Pa. 15024. This research has a been conducted on a segment entitled "Middle America." This segment makes up about 3.2% of the U.S. households. The Middle America segment earns on average about $24,43 1 Members of the Middle America segment are high school educated and are more likely to belong to Christmas clubs. They are likely to read magazines like the Saturday Evening Post and watch shows like Family Ties. Sample Zip codes include Marshall, Michigan 49068; Hagerstown, Md. 21740; "This research has also been conducted on a segment conveniently entitled "Blue Blood Estates." This segment makes up about 1.2% of U.S. households. The blue blood estates segment earns on average about $17,862. Members of the blue blood estates segment typically range from ages 18 to 34. They have been described as high school educated, and highly likely to possess reasonable spending power. They are likely to read magazines like Natural History and watch television shows like Good Morning America. Sample ZIP codes include College Station, Texas 77840; Bloomington, Indiana 47401; Ithaca, New York 14850 and Ft. Lauderdale, Florida 33324." Please turn to the next page,

PAGE 133

123 Example 132: Smithsonian Magazine LAYOUT: Please look at the following example. Take a moment to look closely at this example page content. Try to imagine you sitting at a computer and surfing in to the main page shown below. Take several seconds. Think about how you would react to the site. After you are done looking at it, turn to the next page and answer the following questions. But carefully look at the design and try to read carefully all of the information in the layout. S mithsonian MAGAZINE TMtS ISSUE subscribe After surfing the Woild Wide Web, join us as we travel the w^ole wide world with Smithsonian and Air&Space/Stnithsonian Magailnes. Subscnb'? todayl IMACC CALLER^^ SUBSCRIB euesTBo FEEOBAC Copyri^ 1999 Smitltfonim Mtgudnt Allii^smivtd. Thif tit pradoctd Vy BhirLdtt Htw Mtdi^ Smithsonian ^ 3 SmiihMmian In^aitution MAGAZINE SMrrHS<^ii|iAN national ASSOfclATE MEMB6RSHII* ORDER ONUNE (fxiof m ftcun) Kt Mc.TibGrshlp Benefit* • Snuthsorusn Magazine (12 issues) • Book; record and video discounts • Dis c ounts at mus eum shop s in Washington D.C. and New York • Notice of lecture and seminar programs in Washington D.C. and other locales • Eligibility for travel pro grains • Free admission to the Cooper-Hewitt Nation Design Museum and the National Muse\jm of the American Indian in New York City Turn to the next page. Or call toU-free (877) 212-1940 RATES: Foreign (including Canada and Mexico): One Year (12 issues) for $26.00

PAGE 134

124 DESCRIPTION OF SERVICE. In addition to the web site layout on the adjacent page, a description of the service/product being offered on that Web page will be provided too. It will read as follows: Why would you want to be a Smithsonian National Associate Member? Just think about the pros and cons! Being a Smithsonian National Associate Member has many advantages. For example, each member receives twelve total issues of the Smithsonian Magazine at a special rate that has been discounted specifically for Association members. Twelve issues sent directly to your home or office. In addition to book, record and video discounts as well as other discounts at various museum shops, each member is also kept up to date on various notices of important events and intellectual activities going on in your area. Also, members of the Smithsonian National Association receive eligibility for specially designated travel programs, as well as free admission to a number of the country's best museums and other cultural events. All of these amazing features for just $26.00 for the entire year! Why don't you just weigh the costs and benefits! How can you lose? Please go to the next page...

PAGE 135

125 DESCRIPTION OF SERVICE: In addition to the web site layout on the adjacent page, a description of the service/product being offered on that Web page will be provided too. It will read as follows: Why would you want to be a Smithsonian National Associate Member? Just think about who might be likely to possess such a membership? Think about the type of people who would receive the twelve issues of the Smithsonian Magazine that is included in the membership. What type of person? Individuals who always present themselves with the highest level of intelligence and future success. For example, college educated individuals of this country might keep up to date on various notices of important events and activities going on in their area. These Smithsonian National Association members are the future leaders of this country. The magazines that they read tell others that they appreciate the more culturally involving things in life. Can't you imagine yourself as this type of person, i.e., a college educated, future leader of this country and Smithsonian Associate Member? All of these amazing features for just $26.00 for the entire year! How can you lose? Please go to the next page...

PAGE 136

126 REACTIONS TO THE WEB PAGE AND ITS CONTENT Now please answer the following questions. How would you rate the Web page and its content on the following dimensions: (Circle the number that best reflects your perceptions) Boring 1 2 3 4 5 6 7 Interesting Not Readable 1 2 3 4 5 6 7 Readable Not Visually Attractive 1 2 3 4 5 6 7 Visually Attractive Unpersuasive 1 2 3 4 5 6 7 Persuasive Simple 1 2 3 4 5 6 7 Complex Low Design Quality 1 2 3 4 5 6 7 High Design Quality What is your attitude toward the Smithsonian Association Membership? Unfavorable 1 2 3 4 5 6 7 Favorable In order to get more reliable data, modem market research often presents consumers with scenarios to help give consumers a context within which to make a decision. Please read the following scenario and put yourself in the situation. You reached the point on the website where you are asked to provide (click on using your mouse) further information (personal address, email, etc.) so that you may receive additional info on how to become a Smithsonian Association member. Imagine that as a class exercise, you and your high school teacher were surfing the site together. Given this scenario and the web site and web content that you saw, we would like to get more of your reactions to the web page. Suppose that a portion of the screen on the website asked you to circle the option that best reflects your intention to join the Smithsonian Association Membership. O 1 have a very strong intentions to join O 1 have a strong intention to join O 1 intend to join O Undecided O 1 do not intent to join O 1 have a strong intention to not join O 1 have a very strong intention to not join Thanks for participating in this brief survey. Please go to the next exercise in your task booklet.

PAGE 137

127 REACTIONS TO THE WEB PAGE AND ITS CONTENT Now please answer the following questions. How would you rate the Web page and its content on the following dimensions: (Circle the number that best reflects your perceptions) Boring 1 2 3 4 5 6 7 Interesting Not Readable 1 2 3 4 5 6 7 Readable Not Visually Attractive 1 2 3 4 5 6 7 Visually Attractive Unpersuasive 1 2 3 4 5 6 7 Persuasive Simple 1 2 3 4 5 6 7 Complex Low Design Quality 1 2 3 4 5 6 7 High Design Quality What is your attitude toward the Smithsonian Association Membership? Unfavorable 1 2 3 4 5 6 7 Favorable In order to get more reliable data, modem market research often presents consumers with scenarios to help give consumers a context within which to make a decision. Please read the following scenario and put yourself in the situation. You reached the point on the website where you are asked to provide (click on using your mouse) further information (personal address, email, etc.) so that you may receive additional info on how to become a Smithsonian Association member. Imagine that you were surfing the site in privacy of your own home with nobody else around. Given this scenario and the web site and web content that you saw, we would like to get more of your reactions to the web page Suppose that a portion of the screen on the website asked you to circle the option that best reflects your intention to join the Smithsonian Association Membership. O I have a very strong intentions to join O I have a strong intention to join O I intend to join O Undecided O I do not intent to join O I have a strong intention to not join O I have a very strong intention to not join Thanks for participating in this brief survey. Please go to the next exercise in your task booklet.

PAGE 138

APPENDIX H EXPERIMENT IB: OMINBUS ANOVA OF EFFECTS OF IDENTITY AND RELEVANCE ON PURCHASE LIKELIHOOD OF FOCAL PRODUCT Source DF SS Means Square F Value p value Model 3 57.46 19.15 8.90 < 0.0001 Error 58 124.82 2.15 Type III SS Social Identity (S vs D)a 1 .4503 .4503 .21 .6491 Object Relevance 1 57.40 57.40 26.67 <.0001 Social Identity x Object Relevane 1 .0052 .0052 0.00 .9607 1 "Subjects had the family social identity made salient. Male subjects had their son identity made salient and female subjects had their daughter identity made salient. 128

PAGE 139

APPENDIX I EXPERIMENT IB: OMINBUS ANOVA OF EFFECTS OF IDENTITY AND RELEVANCE ON PURCHASE LIKELIHOOD OF FILLER PRODUCTS Source DF SS Means Square F Value p value Model 3 3.92 1.31 1.03 .3871 Error 58 73.70 1.27 Type III SS Social Identity (S vs D)a 1 .0344 .0344 .03 .8698 Object Relevance 1 .0344 .0344 .03 .8698 Social Identity x Object Relevane 1 3,792 3.792 2.98 .0894 "Subjects had the family social identity made salient. Male subjects had their son identity made salient and female subjects had their daughter identity made salient. 129

PAGE 140

APPENDIX J EXPERIMENT 2: OMINBUS ANOVA OF EFFECTS OF SALIENCE RELEVANCE AND SELF-IMPORTANCE ON PURCHASE LIKELIHOOD OF FOCAL PRODUCT Source DF SS Means Square F Value p value Model 7 82.50 11.79 4.65 .0001 Error 113 286.29 2.53 Type 111 SS Self-importance 1 28.63 28.63 11.30 .0011 Salience 1.72 1.72 .68 .4116 Object Relevance 36.19 36.19 14.28 .0003 Self-Imp X Salience .00005 .00005 0.00 .9961 Salience x Object Relevance .3224 .3224 .13 .7219 Self-Imp X Object Relevance 10.96 10.96 4.33 .0398 Self-Imp X Salience x Object Relevance 1.08 1.08 .43 .5152 130

PAGE 141

APPENDIX K EXPERIMENT 2: OMINBUS ANOVA OF EFFECTS OF SALIENCE RELEVANCE AND SELF-IMPORTANCE ON PURCHASE LIKELIHOOD OF FILLER PRODUCTS Source DF SS Means Square F Value p value Model 7 11.95 1.71 1.27 .2719 Error 113 152.02 1.35 Type III SS Self-importance 8.90 8.90 6.62 .0114 Salience — .2856 .2856 .21 .6458 Object Relevance 2.10 2.10 1.56 .2137 Self-Imp X Salience .6474 .6474 .48 .4893 Salience x Object Relevance .0985 .0985 .07 .7872 Self-Imp X Object Relevance .00039 .00039 0.00 .9863 Self-Imp X Salience x Object Relevance .00037 .00037 0.00 .9868 131

PAGE 142

APPENDIX L EXPERIMENT 3: OMINBUS ANCOVA OF EFFECTS OF SALIENCE RELEVANCE AND SELF-IMPORTANCE ON EVALUATIVE DIMENSIONS Interestingness Dimension |Source DF SS Means Square F Value p value Model 7 29.17 4.17 7.25 0.0001 Error 90 51.73 0.5748 Type III SS Self-importance of Social Identity — 10.79 10.79 18.77 0.0001 Object Relevance 5.05 5.05 8.79 0.0039 Salience 4.95 4.95 8.62 0.0042 Self-importance of Social Identity X pbject Relevance 4.99 4.99 8.69 0.0041 Object Relevance X Salience 10.8 10.8 18.79 0.0001 Self-importance of Social Identity X Salience 5.27 5.27 9.17 0.0032 Self-importance of Social Identity X Object Relevance X Salience 10.83 10.83 18.84 0.0001 132

PAGE 143

Readability Dimension [Source DF ss Means Square F Value p value L [Model 7 4.51 0.6437 1.01 0.4268 Error 90 57.13 0.6347 Type III SS Self-importance of Social Identity 0.055 0.055 0.09 0.7689 Object Relevance — 2.36 2.36 3.72 0.057 Salience — 0.013 0.013 0.02 0.8863 Self-importance of Social Identity X Object Relevance 1.92 1.92 3.03 0.0852 Object Relevance X Salience 0.001 0.001 0 0.9565 Self-importance of Social Identity X Salience 0.08 0.08 0.13 0.7227 Self-importance of Social Identity X Object Relevance X Salience 0.3073 0.3073 0.48 0.4884

PAGE 144

Visual Appeal Dimension Source DF SS Means Square F Value p value Model 7 4.31 0.615 0.89 0.5182 Error 90 62.22 0.6913 Type III SS Self-importance of Social Identity 0.0022 0.0022 0 0.9543 Object Relevance 0.0687 0.0687 0.1 0.7533 OdllCIlCC Z.o4 i.oi U.Ujjj Self-importance of Social Identity X Object Relevance 0.0216 0.0216 0.03 0.8599 Object Relevance X Salience 0.0345 0.0345 0.05 0.8237 Self-importance of Social Identity X Salience 2.43 2.43 3.51 0.0641 Self-importance of Social Identity X Object Relevance X Salience 0.0622 0.0622 0.1 0.7577

PAGE 145

Persuasiveness Dimension Source DF SS Means Square F Value p value Model 7 68.89 9.841 13.40 0.0001 Error 90 66.1 0.7344 Type III SS |Self-Importance of Social Identity 1 26.58 26.58 36.2 0.0001 pbject Relevance 20.17 20.17 27.46 0.0001 fSalience 1 3.65 3.65 4.98 0.0282 Self-importance of Social Identity X pbject Relevance 32.51 32.51 44.27 0.0001 jObject Relevance X Salience 0.2134 0.2134 0.29 0.5912 Self-importance of Social Identity X |Salience 4.794 4.794 6.53 0.0123 ISelf-Importance of Social Identity X pbject Relevance X Salience 0.00009785 0.00009785 0 0.9908 1

PAGE 146

[Complexity Dimension f Source DF SS Means Square F Value p value Model 7 6.04 0.8633 1.63 0.1364 Error 90 47.59 0.5287 Type III SS Self-importance of Social Identity 0.6043 0.6043 1.14 0.2879 Object Relevance — 0.0239 0.0239 0.05 0.8318 Salience — 0.0321 0.0321 0.06 0.8057 Self-importance of Social Identity X Object Relevance 0.0768 0.0768 0.15 0.7039 Object Relevance X Salience 0.7434 0.7434 1.41 0.2388 Self-Importance of Social Identity X Salience 0.0506 0.0506 O.I 0.7576 Self-importance of Social Identity X Object Relevance X Salience 2.218 2.218 4.19 0.0435

PAGE 147

Design Quality Dimension Isource DF SS Means Square F Value p value Model 7 2.218 0.3169 0.56 0.7881 Error 90 51.129 0.5681 Type III SS Self-importance of Social Identity 0.401 0.401 0.71 0.403 Object Relevance — 0.2191 0.2191 0.39 0.5361 Salience 0.1017 0.1017 0.18 0.6732 Self-importance of Social Identity X Object Relevance 0.069 0.069 0.12 0.7282 Object Relevance X Salience 0.2015 0.2015 0.35 0.553 Self-importance of Social Identity X Salience 0.3837 0.3837 0.68 0.4133 Self-importance of Social Identity X Object Relevance X Salience 0.3335 0.3335 0.59 0.4456

PAGE 148

Attitudes Toward Product Source DF SS Means Square F Value p value Model 7 109.455 15.64 13.19 0.0001 Error 90 106.677 1. 185 Type III SS Self-importance of Social Identity 25.45 25.45 21.47 0.0001 Object Relevance — — 10.02 10.02 8.46 0.0046 |Salience 0.58 0.58 0.49 0.4854 ISelf-Importance of Social Identity X Object Relevance 36.55 36.55 30.84 0.0001 Object Relevance X Salience 0.9677 0.9677 0.82 0.3687 Self-importance of Social Identity X Salience 0.2814 0.2814 0.24 0.6273 Self-importance of Social Identity X Object Relevance X Salience 0.6858 0.6858 0.58 0.4489 1

PAGE 149

REFERENCES Aaker, J. (1997). Dimensions of brand personality. Journal of Marketing Research 34, 347-356. Ajzen, I., andFishbein, M. (1980). Understanding attitudes and predicting social behavior Englewood Cliffs, NJ: Prentice-Hall. Ajzen, I., and Madden, T. J. (1986). Prediction of goal directed behavior: Attitudes, intentions and perceived behavioral control. Journal of Experimental Social Psychology 22, 453-474. Baldwin, M. and Holmes, J. G. (1987). Salient private audiences and awareness of the self Journal of Personality and Social Psychology, 52, 1087-1098. Bern, D. J. (1967). Self perception: An alternative interpretation of cognitive dissonance phenomena. Psychological Review 74, 183-200. Bem, D. J. (1971). Self perception theory. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 6). New York: Academic Press. Bem, S. (1981) Gender schema theory: A cognitive account of sex typing. Psychological Review 88, 354-364. Berger, P. L. and Lukmann T. (1966). The social construction of reality Garden City, New York: Doubleday. Berger, P.L., and Luckmann, T. (1971). The social construction of reality Harmonds worth: Penguin. Burke, P. J. (1980). The self: measurement requirements from an interactionist perspective. Social Psychology Quarterly 43, 18-29. Chamg, H., Piliavin, J A., and Callero, P. L. (1988). Role identity and reasoned action in the prediction of repeated behavior. Social Psychology Quarterlv 5 1 303-3 1 7. Cohen, Joel B., and Reed, Americus. (2000) [When attitudes and group norms conflict]. Unpublished raw data. Deaux, K. (1997). Social Identification. Handbook of Social Psychological Principles (pp. 777-798). Deshpande, R., Hoyer, W. D., and Donthu, N. (1986), The intensity of ethnic affiliation: A study of the sociology of Hispanic consumption. Journal of Consumer Research 13 (September), 214-220. 139

PAGE 150

140 Deshpande, R., and Stayman, D. (1994). A tale of two cities: Distinctiveness theory and advertising effectiveness, Journal of Marketing Research 21 (February), 57-64. Deutsch, M., and Gerard, H. (1955). A study of normative and informational social influences upon individual judgment. Journal of Abnormal and Social Psychology 51,629-636. Dunnett, C. W. (1955). A multiple comparison procedure for comparing several treatments with a control. Journal of the American Statistical Association 50, 1096-1121. Eagly, A. H., and Chaiken, S. (1993). The psychology of attitudes New York: Harcourt Brace Jovanovich, Inc. Erikson, E. H. (1950). Childhood and Society New York: Norton. Erikson, E. H. (1956a). Ego development and historical change The Psychoanalytic Study of the Child 2, 359-396. Erikson, E. H. (1956b). The problem of ego identity. The Journal of the American Psychoanalytic Association 4, 56-121. Erikson, E. H. (1974). Dimensions of a new identitv New York: Norton Erikson,E. H. (1978). Adulthood New York: Norton. Fazio, R. H. (1986). How do attitudes guide behavior? In R. M. Sorrentino and E. T. Higgins (Eds.), Handbook of motivation and cognition: Foundation of social behavior (pp. 204-243). New York: Guilford Press. Fazio, R. H., Chen, J., McDonel, E. C, and Shemian, S. J. (1982). Attitude accessibility, attitude-behavior consistency, and the strength of the object-evaluation association. Journal of Experimental Social Psychology 18, 339-357. Fazio, R. H., and Zanna, M. P. (1978a). Attitude accessibility as a moderator of the attitude-perception and attitude-behavior relations: An investigation of the 1984 presidential election. Journal of Personality and Social Psychology 14,398-408. Fazio, R. H., and Zanna, M. P. (1978b). On the predictive validity of attitudes: The roles of direct experience and confidence. Journal of Personality 46, 228-243. Feldman, J., and Lynch, J. G. (1988). Self-generated validity and other effects of measurement on belief, attitude, intention and behavior. Journal of Applied Psychology 73(3), 421-435.

PAGE 151

141 Fishbein, M., andAjzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research Reading, MA: AddisonWesley. Fishbein, M., and Middlestat, S. (1995). Noncognitive effects on attitude formation and change: Fact or artifact. Journal of Consumer Psychology Vol. 4, no. 2. Fiske, S. T., and Pavelchak, M. A. (1986). Category based affective responses: Developments in schema-triggered affect. In R. M. Sorrentino and E. T. Higgins (Eds.), Handbook of motivation and cognition: Foundations of social behavior (pp. 167-203). New York: Guilford Press. Foote, N. (1951). Identification as the basis for a theory of motivation. American Sociological Review 16, 14-21. Forehand, M., and Deshpande R. (in press), When what we see makes us who we are: Priming ethnic self-awareness and advertising response. Journal of Marketing Research French, J. R. P., and Raven, B. (1959). The bases of social power. In D. Cartvraght (Ed.), Studies in social power (pp. 150-167). Ann Arbor: University of Michigan. Freud, S. (1985). Civilization, society, and religion: Group psychology, civilization and its discontents and other works Middlesex, England: Penguin Books. (Original work published 1955). Giles, H., and Johnson, P. (1987). Ethnolinguistic identity theory: A social psychological approach to language maintenance. International Journal of the Sociology of Language 68, 256-269. Goffman, E. (1959). The presentation of self in everyday life New York: Anchor. Goffman, E. (1961a). Asylums New York: Doubleday Anchor. Gorsuch, R. L., and Ortberg, J. (1983). Moral obligation and attitudes: Their relation to behavioral intentions. Journal of Personality and Social Psychology 44, 10251028. Granberg, D., and Holmberg, S. (1990). The intention-behavior relationship among U.S. and Swedish voters. Social Psychology Quarterly 53, 44-54. Gregan-Paxton, Jennifer, and John, Deborah Roedder (1997) Consumer learning by analogy: A model of internal knowledge transfer. Journal of Consumer Research 24(3), 266-284. Higgins, E. T., Loeb, I., and Moretti, M. (1995). Self-discrepancies and developmental shifts in vulnerability: Life transitions in the regulatory significance of others. In

PAGE 152

142 D. Cicchetti and S. L. Toth (Eds.)., Rochester symposium on developmental psychopathology. Volume 6: Emotion, cognition and representation (pp. 191230). Rochester, NY: University of Rochester Press. Hinkle, S., and Brown, R. J. (1990). Intergroup comparisons and social identity: Some links and lacunae. In D. Abrams and M. A. Hogg (Eds.), Social identity theory: Constructive and critical advances (pp. 148-167). London and New York: Harvester Wheatsheaf and SpringerVerlag. Hirschman, E. C. (1981), American Jewish ethnicity: Its relationship to some selected aspects of consumer behavior, Journal of Marketing 45(3), 102-1 10. Hogg, M. A. (1992). The social psychology of group cohesiveness: From attraction to social identity New York: New York University. Hogg, M. A. (1996), Intragroup processes, group structure and social identity. In W. P. Robinson (Ed.), Social groups and identity: Developing the legacy of Henri Tajfel (pp. 173-190). Oxford: Butterworth-Heinemann. Hogg, M.A., and Abrams, D. (1988), Social identifications: A social psychology if intergroup relations and group processes. London and New York: Routledge. Jaffe, L. (1991). The Impact of positioning and sex-role identity on women's responses to advertising. Journal of Advertising Research 3 1 (3), 57-64. Kahn, B. E. and Meyer, R. J. (1991). Consumer multiattribute judgments under attributeweight uncertainty," Journal of Consumer Research 17 (March), 508-522. Kallgren, C.A., Reno, R.R., and Cialdini, R.B. (2000), A focus theory of normative conduct: When norms do and do not affect behavior. Personality and Social Psychology Bulletin 26(8), 1002-1012. Katz, D. (1960). The functional approach to the study of attitudes. Public Opinion Quarterly, 24, 163-204. Keller, Punam Anand, and McGill, Ann L. (1994). Differences in the relative influence of product attributes under alternative processing conditions: Attribute importance versus attribute ease of imagability. Journal of Consumer Psychology 3(1), 2949. Kelman, H. C. (1958). Compliance, identification, and internalization: Three processes of attitude change. Journal of Conflict Resolution 2. 51-60. Kelman, H. C. (1961). Process of opinion change, Public Opinion Quarteriy 25, 57-78.

PAGE 153

143 Kelman, H. C. (1974). Further thoughts on the processes of comphance, identification, and intemaHzation. In J. T. Tedeschi (Ed.), Perspectives on social power (pp. 125-171). Chicago: Aldine. Keppel, G. (1991). Design and analysis. A researcher's handbook. Englewood Cliffs; NJ: Prentice Hall-Inc. Kravitz, D. A. (1995). Attitudes toward affirmative action plans directed at blacks: Effects of plan and individual differences. Journal of Applied Social Psychology 25,2192-2220. Levy, S. J. (1959), Symbols for Sales. Harvard Business Review 37, 1 17-124. Lichtenstein, H. (1977). The dilemma of human identity New York: Jason Aronson. Lingle, J. H., and Ostrom, T. M. (1979). Retrieval selectivity in memory-based impression judgments. Journal of Personality and Social Psychology 37, 2. 180194. Liska, A. E. (1984). A critical examination of the causal structure of the Fishbein/Ajzen attitude-behavior model. Social Psychology Quarterly 47, 61-74. Lutz, R. J. (1975) Changing brand attitudes through modification of cognitive structure, Journal of Consumer Research 1 (March), 49-59. Lynch, J. G. (1985). Uniqueness issues in the decomposition modeling of multiattribute overall evaluations: An information integration perspective. Journal of Marketing Research 22(1), 1-19. Markus, H., and Kunda, Z. (1986). Stability and malleability of the self concept. Journal of Personality and Social Psychlogy 15, 858-866. McCall, G. J., and Simmons, J. L. (1978). Identities and interactions (rev. ed.). New York: Free Press. McGuire, W. J., McGuire, C. V., Child, P., and Fujioka, T. (1978) Salience of ethnicity in the spontaneous self concept as a function of one's ethnic distinctiveness in the social environment. Journal of Personality and Social Psychology 36, 51 1-520. McGuire, W. J., McGuire, C. V., and Winton, W. (1 979). Effects of household sex composition on the salience of one's gender in the spontaneous self concept. Journal of Experimental Social Psychology 15, 77-90. Mead, G. H. (1934). Mind, self and society Chicago: University of Chicago Press.

PAGE 154

144 Meyers-Levy, J. (1988) The influence of sex roles on judgments," Journal of Consumer Research 14 (March) 522-530. Miniard, P. W., Cohen, J. B. (1981). An examination of the Fishbein-Ajzen behavioral intentions model's concepts and measures. Journal of Experimental Social Psychology 17, 309-339. Newman, Leonard S, Duff, Kimberley, Schnopp-Wyatt, Nicole, Brock, Bradley, and Hoffman, Yonit (1997). Reactions to the O. J. Simpson verdict: "Mindless tribalism" or motivated inference processes? Journal of Social Issues 53(3), 547562. O'Guinn, T. C, and Meyer, T. P. (1984). Segmenting the Hispanic market: The use of Spanish-language radio. Journal of Advertising Research 23 (6), 9-16. Peart, N. (1981). The weapon. On Signals [CD]. New York: Polygram Music. Pham, Michel Tuan. (1998). Representativeness, relevance, and the use of feelings in decisionmaking. Journal of Consumer Research 25(2), 144-159. Regan, D. T., and Fazio, R. H. (1977). On the consistency between attitudes and behavior: Look to the method of attitude formation. Journal of Experimental Social Psychology 13, 28-45. Rokeach, M. (1968). Beliefs, attitudes and values: A theory of organization and change San Francisco: Jossey-Bass. Rokeach, M. (1973). The nature of human values New York: Free Press. Rokeach, M. (1980). Some unresolved issues in theories of beliefs, attitudes, and values. In H. E. Howe, Jr., and M. M. Page (Eds.), Nebraska Symposium on Motivation, 1979 (Vol. 27, pp. 261-304). Lincoln: University of Nebraska Press. Rosenberg, M. J., and Hovland, C. I. (1960). Cognitive, affective and behavioral components of attitudes. In C. I. Hovland and M. J. Rosenberg (Eds.), Attitude organization and change: An analysis of consistency among attitude components (pp. 1-14). New Haven, CT: Yale University Press. Saegert, J., Hoover, R.J., Hilger, M.T. (1985), Characteristics of Mexican American consumers. Journal of Consumer Research 12(1), 104-109. Schlenker, B. R. (1985). Identity and self identification. In B. R. Schlenker (Ed.), The self and social life (pp. 65-99). New York :McGraw Hill.

PAGE 155

145 Schwartz, S. H., and Tesser, R. C. (1972). A test of a model for reducing measured attitude behavior discrepancies. Journal of Personality and Social Psychology 24, 255-236. Schwarz, N., and Clore, G. L. (1988). How do I feel about it? The informative fiinction of affective states. In K. Fiedler and J. Forgas (Eds.), Affect, cognition and social behavior: New evidence and integrative attempts (pp. 44-62), Toronto: Hogrefe. Schwarz, N., and Bohner, G. (2000). The construction of attitudes. Draft for a chapter for A. Tesser and N. Schwarz (Eds.), Blackwell handbook of social psychology: Intrapersonal processes Oxford, UK: Blackwell. Schwarz, N. (1990). Feelings as information: Informational and motivational ftinctions of affective states. In E.T. Higgins and R.M. Sorrentino (Eds.), Handbook of motivation and cognition: Foundation of social behavior (Vol. 2, pp. 527-561). New York: Guilford Press. Shavitt, S. (1989). Operationalizing functional theories of attitudes. In A. R. Pratkanis, S. J. Brecklyer, and A. G. Greenwald (Eds.), Attitude structure and function (pp. 31 1-338). Hillsdale, NJ: Erlbaum. Shavitt, S. (1 990). "The role of attitude objects in attitude ftinctions," Journal of Experimental Social Psychology 26, 124-148. Shih, Margaret; Pittinsky, Todd L; Ambady, Nalini (1999). Stereotype susceptibility: Identity salience and shifts in quantitative performance. Psychological Science 10(1), Jan, 80-83. Sigall, Harold; Mills, Judson (1998). Measures of independent variables and mediators are useful in social psychology experiments: But are they necessary? Personality and Social Psychology Review 2(3) 218-226. Sirgy, J. M. (1982). Self-concept in consumer behavior: A critical review. Journal of Consumer Research 9, 287-300. Smith, M. B., Bruner, J. S., and White, R. W. (1956). Opinions and personality New York: Wiley. Solomon, Michael R. (1983). The role of products as social stimuli: A symbolic interactionism perspective. Journal of Consumer Research 10,319-329. Songer-Nocks, E. (1976). Situation factors affecting the weighting of predictor components in the Fishbein model. Journal of Experimental Social Psych ology 12, 56-69.

PAGE 156

146 Stayman, D. and Deshpande, R. (1989). Situational ethnicity and consumer behavior. Journal of Consumer Research 16 (December), 361-371. Stem, Paul C, Dietz, Thomas, Kalof, Linda, and Guagnano, Gregory A. (1995). Values, beliefs, and proenvironmental action: Attitude formation toward emergent attitude objects. Journal of Applied Social Psychology 26(18), 161 1-1636. Stone, Gregory P. Appearance and the self: A slightly revised version. (1990) In D. Brissett, D., Edgley, C, et al., (Eds.), Life as theater: A dramaturgical sourcebook Communication and social order (2"** ed., pp. 141-162). New York: Aldine De Gruyter. Stryker, S. (1980). Symbolic interactionism Menlo Park, CA: Benjamin Cummings. Stryker, S., and Serpe, R. T. (1982). Commitment, identity salience, and role behavior: Theory and research example. In W. Ickes and E. S. Knowles (Eds.), Personality, roles and social behavior (pp. 199-218). New York: SpringerVerlag. Stryker, S., and Statham, A. (1985). Symbolic interaction and role theory. In G. Lindzey and E. Aronson (Eds.), The handbook of social psychology (3''' ed.. Vol. 1, pp. 31 1-378). New York: Random House. Tajfel, H. (1959), Quantitative judgment in social perception. British Journal of Psychology 50, 16-29. Tajfel, H. (1969), "Social and cultural factors in perception," In G. Lindzey and E Aronson (Eds.), Handbook of social psychology (Vol. 3, pp. 315-394). Reading, MA: AddisonWesley. Tajfel, H. (1978). Differentiation between social groups London: Academic Press. Tajfel, H. (1981). Human groups and social categories: Studies in social psychology Cambridge, UK: Cambridge University Press. Tajfel, H. (1982). Social psychology of if intergroup relations. Annual Review of Psychology 33. 1-39. Tajfel, H., and Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin and S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33-47). Monterey, CA: Brooks/Cole. Terry, D. J., and Hogg, M. A. (1996). Group norms and the attitude behavior relationship: A role for group identification. Personality and Social Psychology Bulletin 22, 776-793.

PAGE 157

147 Thoits, P. A., and Virshup, L. V. (1995). Me's and we's: Forms and functions of social identities. Paper presented at the symposium on self and social identity, Rutgers University, New Brunswick, NJ. Triandis, H. C. (1977). Interpersonal behavior Monterey, CA: Brooks/ Cole. Triandis, H. C. (1980). Values, attitudes and interpersonal behavior. In H. E. Howe, Jr., and M. M. Page (Eds.), Nebraska symposium on motivation (Vol. 27, pp. 195259). Lincoln: University of Nebraska Press. Turner, J. C. (1982), Towards a cognitive redefinition of the social group. In H. Tajfel (ed.), Social identity and intergroup relations (pp. 15-40). Cambridge, England: Cambridge University Press. Turner, J. C. (1985), Social categorization and the self-concept: A social cognitive theory of group behaviour. In E.J. Lawler (Ed.), Advances in group processes: Theory and research (Vol. 2, pp-22-122). Greenwich, CT: JAI. Turner, J. C, and Giles, H. (Eds.). (1981), Intergroup behaviour Oxford, UK: Blackwell. Turner, J. C, and Oakes, P. J. (1986). The significance of the social identity concept for social psychlogy with references to individualism, interactionism and social influence. British Journal of Social Psychology 25, 237-252. Turner, J. C, Hogg, M. A., Oakes, P. J., Reicher, S. D., and Wetherell, M. S. (1987). Rediscovering the social group. A self categorization theory Oxford, UK: Basil Blackwell. Turner, J. C, Oakes, P. J., Haslam, S. A.., and McGarty, C. (1994). Self and collective: Cognition and social context. Personality and Social Psychology Bulletin 20 454-463. Turner, R. H. (1956). Role-taking, role standpoint, and reference-group behavior. The American Journal of Sociology 61, 316-328. Van Knippenberg, A., and Ellemers, N. (1990). Social identity and intergroup differenfiation processes. In W. Stroebe and M. Hewstone (Eds.), European review of social psychology (Vol. l,pp. 137-168). Chichester, UK: Wiley. Wallendorf, M. and Reilly, M.D. (1983). Ethnic migration, assimilation, and consumption. Journal of Consumer Research 10 (December), 292-302. Watson, David, Clark, Lee A, and Tellegen, Auke. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6). 1063-1070.

PAGE 158

148 Weigert, A. (1983). Identity: Its emergence within sociological psychology. Symbolic Interaction, 6(2), 183-206. Wilson, T. D., and Hodges, S. D. (1992). Attitudes as temporary constructions. In Leonard L. Martin and Abraham Tesser (Eds.), The construction of social judgments (pp. 37-65). Hillsdale, NJ: Erlbaum. Wilson, T.D., Lindsey, S. and Schooler, T. Y. (2000). A model of dual attitudes. Psychological Review 107 (1), 101-126. Wooten, D. B. (1995), One of a kind in a full house: Some consequences of ethnic and gender distinctiveness. Journal of Consumer Psychology 4(3), 205-224. Ybarra, Oscar, and Trafimow, David. (1998) How priming the private self or collective self affects the relative weights of attitudes and subjective norms. Personality and Social Psychology Bulletin 24(4), 362-370. Zuckerman, M., and Reis, H. (1978). Comparison of three models for predicting altruistic behavior. Journal of Personality and Social Psychology 36, 498-510.

PAGE 159

BIOGRAPHICAL SKETCH Americus Reed II was bom on August 13, 1970, in Queens, New York. He grew up in Atlanta, Georgia, devoting most of his free time to pondering the paradoxes of life, drawing cartoons, writing science fiction stories and playing drums, classical guitar and bass in a progressive art-rock trio called "Spear." He attended Georgia State University where he received an undergraduate degree in business management and two Master of Science degrees in organizational behavior and market research methods, respectively. He became interested in consumer behavior after taking an undergraduate course and got really excited about research after taking a master's level elective in social psychology. He went into industry in between working on his two master's degrees and in 1990 started his own telecommimications company called Reed Communications, Inc. In 1995 he was awarded a five year minority tuition stipend to attend the Ph.D. program in marketing at the University of Florida. He was the 1999 American Marketing Association Doctoral Consortium Fellow from the University of Florida. In August 2000, he accepted a position as Assistant Professor of Marketing at The Wharton School of the University of Pennsylvania and is still trying to master the paradiddle. 149

PAGE 160

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. foe] B. Cohen, Chairman /Distinguished Service Professor of Marketing I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Dolores AJbarfacin Assistant Professor of Psycho/ogy I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.

PAGE 161

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy, y ^ Professor of Psychology This dissertation was submitted to the Graduate Faculty of the Department of Marketing in the College of Business Administration and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 2000 Dean, Graduate School

PAGE 162

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy, y ^ Professor of Psychology This dissertation was submitted to the Graduate Faculty of the Department of Marketing in the College of Business Administration and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 2000 Dean, Graduate School


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EZ8SKHHZ0_646RMG INGEST_TIME 2014-12-18T23:04:58Z PACKAGE AA00026523_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES