YOU ARE WHAT THEY EAT:

THE INFLUENCE OF REFERENCE GROUPS ON

CONSUMERS’ CONNECTIONS TO BRANDS

Jennifer Edson Escalas

Eller College of Business and Public Administration
University of Arizona

James R. Bettman
Fuqua School of Business
Duke University

Forthcoming at the Journal of Consumer Psychology, v. 13, n. 4

 Requests for reprints should be sent to Jennifer Edson Escalas, Department of Marketing,

PO Box 210108, The Eller School of Business and Public Administration, University of Arizona,

Tucson, AZ 85721-0108. E-mail: [email protected].

1
 YOU ARE WHAT THEY EAT: THE INFLUENCE OF REFERENCE GROUPS

ON CONSUMERS’ CONNECTIONS TO BRANDS

ABSTRACT

The set of associations consumers have about a brand is an important component of

brand equity. This paper focuses on reference groups as a source of brand associations, which

can be linked to one’s mental representation of self to meet self-verification or self-enhancement

goals. We conceptualize this linkage at an aggregate level in terms of self-brand connections,

i.e., the extent to which individuals have incorporated a brand into their self-concept. Two

studies show that brands used by member groups and aspiration groups can become connected to

consumers’ mental representation of self as they use these brands to define and create their self-

concepts. Results from Experiment 1 show that the degree to which member group and

aspiration group usage influences individual self-brand connections is contingent upon the

degree to which the individual belongs to a member group or wishes to belong to an aspiration

group. Experiment 2 finds that for individuals with self-enhancement goals, aspiration group

brand use has a greater impact on self-brand connections; for individuals with self-verification

goals, on the other hand, member group use has a greater impact.

2
 INTRODUCTION

The set of associations consumers have about a brand is an important component of

brand equity (Keller 1993, Aaker 1991). Brand associations include both user imagery and

psychological benefits. Many consumer researchers have made the assertion that people engage

in consumption behavior to construct their self-concepts and to create their personal identity

(e.g., Richins 1994; Kleine, Kleine, and Kernan 1993; Ball and Tasaki 1992). We examine one

aspect of this construction process, namely the appropriation of brand associations that are

derived from brand usage by reference groups. We propose that consumers use brands to meet

self-needs, such as self-verification or self-enhancement. When brand associations are used to

construct the self or to communicate the self-concept to others, a connection is formed with the

brand. In particular, we consider reference groups as a source of brand associations that lead to

such connections. That is, associations about reference groups become associated with brands

those groups are perceived to use, and vice versa. The set of associations can then be linked to

consumers’ mental representations of self as they select brands with meanings congruent with an

aspect of their current self-concept or possible self, thus forging a connection between the

consumer and the brand. Two studies provide empirical support for the notion that brands used

by reference groups can lead to connections between consumers and brands as they use brands to

define and create their self-concepts. Findings from the second study also support the idea that

consumers' predominant self-goal (i.e., self-verification versus self-enhancement) determines

which reference groups' brand use (i.e., member groups versus aspiration groups) will have the

most influence on self-brand connections.

3
 THEORETICAL DEVELOPMENT

A positive brand image is created through building strong, favorable, and unique

associations to the brand in memory (Keller 1993), including user imagery and psychological

benefits (Aaker 1991). User imagery consists of associations about the typical brand user,

including demographic and psychographic associations. Reference group brand usage is an

important source of user imagery brand associations. Psychological benefits, including social

approval, personal expression, and outer-directed self-esteem, can also be associated with user

imagery (Keller 1993). Consumers value psychological brand benefits because these benefits can

help consumers construct their self-identity and/or present themselves to others.

These assertions about psychological benefits parallel consumer research on the

significance of important possessions and brands. Possessions and brands can be used to satisfy

such psychological needs as actively creating one’s self-concept, reinforcing and expressing self-

identity, and allowing one to differentiate oneself and assert one's individuality (Richins 1994,

Ball and Tasaki 1992, Belk 1988, Fournier 1998). Possessions and brands can also serve a social

purpose by reflecting social ties such as one's family, community, and cultural groups

(Wallendorf and Arnould 1988, Reingen et al. 1984). As symbols, brands can add to and/or

reinforce the way the consumer thinks about him/herself.

Thus, the set of brand associations can be used to create and define a consumer’s self-

concept. Reference group usage of a brand provides user image associations and psychological

benefit associations for brands (Muniz and O’Guinn 2001). Consumers construct themselves and

present themselves to others through their brand choices based on the congruency between brand

image and self-image. As a result of this process, the set of brand associations is linked to the

consumer’s mental representation of self. Thus, the meaning and value of a brand is not just its

4
ability to express the self, but its role in helping consumers create and build their self-identities

(McCracken 1989) by forming connections to brands.

Self-Brand Connections

We propose that the set of brand associations can be more meaningful the more closely it

is linked to the self. We conceptualize and operationalize this linkage at the aggregate level of

self-brand connections, the extent to which individuals have incorporated brands into their self-

concept (Escalas 1996, Escalas and Bettman 2000). To achieve their identity goals (Huffman,

Ratneshwar, and Mick 2000), people use products and brands to create and represent self-images

and to present these images to others or to themselves. As a result of this process, a link bridges

the brand and the self. We focus on self-brand connections, rather than specific brand

associations, because we believe that brand meaning is most often dependent upon the entire

constellation, or gestalt, or the set of brand associations. Thus, we believe that self-brand

connections capture an important part of consumers’ construction of self. Therefore, the primary

dependent variable in our studies is a measure of the degree to which consumers have

incorporated the brand into their self-concept, that is, formed a self-brand connection

(measurement is described below).

Constructive Self-Concept/Self-Motivation

In our model, the set of brand associations can help consumers achieve goals that are

motivated by the self when these associations are linked or connected to the self. People are

motivated to create a favorable and consistent self-identity. Possible selves, e.g., individuals'

ideas of what they might become, what they would like to become, and what they are afraid of

becoming, also motivate behavior to achieve the realization of personal goals (Markus and

5
Nurius 1986). While there are many different self-construction and self-presentation motivations,

in this paper we focus on two general sets of self motives have been the focus of recent social

psychological research: self-enhancement and self-knowledge (need for consistency and self-

verification) (Banaji and Prentice 1994).

Self-Enhancement/Impression Management. People are heavily influenced by the need to

maintain and enhance self-esteem (Greenwald, Bellezza, and Banaji 1988). The impressions

psychologically healthy individuals hold of themselves are biased in a positive direction. People

judge positive traits to be more characteristic of themselves, and positive personality information

is more efficiently processed and recalled (Kuiper and Derry 1982). Attribution research shows

that people are more likely to attribute positive outcomes to aspects of self and negative

outcomes to circumstances unrelated to self (Miller and Ross 1975). This self-serving bias is

perpetuated because self-enhancement tendencies impose filters on incoming information.

Another important aspect of self-enhancement involves social interaction. People manage

their presentations of self in various situations in order to maximize positive feedback (Schlenker

1980). Usually, people strive to make a good impression, especially if they have high self-esteem

(Baumeister, Tice and Hutton 1989). People are motivated to create a good impression to gain

social approval and for the intrinsic satisfaction of projecting a positive self-image, even to

oneself (Schlenker 1980). Social psychological research has identified several techniques for

impression management: conforming to social norms, behavioral matching, self-promotion,

flattery, and projecting consistency between beliefs and behavior (Fiske and Taylor 1991).

Self-Verification. In addition to the need for self-enhancement and impression

management, people often have needs for self-knowledge, including self-verification. In general,

people seek out and interpret situations and adopt behavioral strategies that are consistent with

6
their existing self-conceptions. Similarly, they avoid situations and behaviors that yield

information contradictory to their existing self-conceptions. In our research, we focus on self-

verification, where people strive to preserve their self-conceptions by eliciting self-verifying

feedback. In this context, consistency provides individuals with a sense that the world is

predictable and controllable (Swann 1990). There are two primary strategies used to achieve

self-verification: 1) seeing more self-confirmatory evidence than actually exists and 2) striving to

influence the reactions of others by developing a self-confirmatory social environment, which

includes displaying identity cues such as driving a certain brand of automobile (Swann 1990,

Schlenker 1980).

Niedenthal et al. (1985; Setterlund and Niedenthal 1993) have found that people choose

situations (including products and brands) by 1) imagining the prototypical users for each item in

the choice set and 2) choosing the item that maximizes their similarity to a desired prototypical

user. Thus, people choose situations (including products) based on their need for self-

consistency: they select products by matching themselves to prototypical users, a heuristic

labeled “prototype matching.” For example, a person who perceives himself as an

environmentalist purchases the type of lawnmower he believes environmentalists use. Although

Neidenthal et al. explain their findings using self-consistency, self-enhancement could also be an

underlying rationale for their findings, because participants maximize their similarity to a desired

prototypical user, which may enhance their self-image.

Reconciling Self-Enhancement and Self-Verification Goals. It might appear that seeking

feedback that is favorable (self-enhancement) is incompatible with seeking accurate feedback

regardless of its favorability (self-verification). Social psychological research has shown that

various factors influence the relative degree to which each goal is active (e.g., cognitive

7
resources (Swann 1990); stable vs. malleable aspects of personality (Dunning 1995); intuitive-

experiential vs. analytical-rational modes of thought (Morling and Epstein 1997); or cognitive

vs. affective processes (Swann et al. 1987)). In terms of these motives being differentially

prevalent in different people, Sedikides and Strube (1995) find that people with high self-esteem,

high self-monitors, narcissists, and type B personalities are more likely than their counterparts to

be influenced by self-enhancement motives as opposed to self-verification motives. In our

experiments, we examine the relative degree of self-enhancement vs. self-verification by

measuring both. We predict that these two self-goals will differentially affect which reference

groups consumers look to for brand associations as they construct and present their self-

concepts.

Consumer Research on Reference Groups

Consumer research on reference groups has demonstrated congruency between group

membership and brand usage (e.g., Bearden and Etzel 1982; Burnkrant and Cousineau 1975;

Childers and Rao 1992; Bearden, Netemeyer, and Teel 1989; Mochis 1985). Reference groups

are defined as social groups that are important to a consumer and against which he/she compares

him/herself. Early research focused on classification systems for the reference groups to which

an individual turns as a standard for behaviors. For example, a member group is a reference

group to which an individual belongs, while an aspiration group is a reference group to which an

individual aspires to belong. More recent reference group research is based on conformity and

social comparison theory (see Folkes and Kiesler 1991 for a review). Consumers use others as a

source of information for arriving at and evaluating one’s beliefs about the world. This

assessment of opinions and abilities uses relevant others who share beliefs and are similar on

8
relevant dimensions. Three types of reference group influence have been identified in consumer

research: informational, utilitarian, and value-expressive (Park and Lessig 1977).

We argue that consumers actively construct themselves, using brand associations that

arise through reference group usage and the resulting self-brand connections. By examining

consumers’ self-brand connections, we attempt to demonstrate that brand use by reference

groups is a source of brand associations, which become linked to consumers’ mental

representation of self as consumers actively construct themselves by selecting brands with

associations relevant to an aspect of their current self-concept or possible self. We propose that

when consumers appropriate the brand images of brands used by their reference group, they do

so to meet self-related needs. For example, self-enhancement needs would lead to forming

connections with brands associated with favorable aspiration groups and avoidance of brands

associated with unfavorable prototypical users. Consumers form self-brand connections that

become meaningful through this process.

STATEMENT OF HYPOTHESES

One source of brand associations is brand usage by reference groups with certain images.

These brands (and their associations) may then be connected to consumers as they construct their

self-identities. The extent to which such connections occur will depend upon the degree to which

the consumer is a member of a reference group. For example, if a consumer considers

him/herself to be an intellectual and his/her member group of intellectuals tends to drive Volvos,

he/she may choose to also drive a Volvo as a symbol of how intellectual he/she is and form a

connection between Volvo and the self. On the other hand, if being intellectual is not viewed as a

desirable aspect of the self, then the consumer will not form such a connection with Volvo. Thus,

9
consumers may form self-brand connections to the brands used by reference groups to which

they currently belong (i.e., member groups), depending upon the fit between the member group

and the self.

H1: Consumers’ perceptions that they belong to a member group will moderate the effect of
that group’s brand usage on self-brand connections. Consumers will have more positive
self-brand connections when they perceive that a member group uses a brand AND that
they have a positive fit with the member group.

Similarly, consumers may utilize the associations derived from groups to which they

would like to belong, that is, aspiration groups. When aspiration groups use a brand, consumers

may form associations about the brand that they attempt to transfer to themselves, despite the

fact that they are not yet members of the aspiration group. Nevertheless, the brand becomes

meaningful in the process of being used to construct one’s possible self. For example, if a

consumer wishes to be more hip, and he/she sees hip people wearing Versace clothing, he/she

may choose to wear Versace clothing in an attempt to appropriate the hip associations of that

brand. Thus, H1 refers to groups in which the individual belongs, whereas H2 refers to those

groups to which the individual aspires to belong.

H2: Consumers’ perceptions that they wish to belong to an aspiration group will moderate the
effect of that group’s brand usage on self-brand connections. Consumers will have more
positive self-brand connections when they perceive that an aspiration group uses a brand
AND that they have a positive fit with the aspiration group.

If an individual is guided by self-verification goals, that is, seeking accurate feedback

regardless of its favorability, then we would expect that he/she would strive to create and/or

project a realistic image of him/herself. An accurate self-image would reflect membership in

reference groups to which he/she belongs (i.e., member groups). On the other hand, if an

individual is guided by self-enhancement goals, that is, seeking feedback that is favorable and

creating favorable impressions on others, then we would expect a greater influence of aspiration

10
groups. Self-enhancement would be achieved by creating/projecting an image consistent with

groups to which the individual wishes he/she belonged.

H3: For the relationships proposed in H1 and H2, consumers who have self-verification goals
will be more influenced by their member groups, while consumers who have self-
enhancement goals will be more influenced by their aspiration groups.

Note that this third hypothesis goes beyond the predictions made by previous studies by

postulating an interaction between the goal and the source of influence on self-brand

connections.

EXPERIMENT 1

In this study, we examine how reference groups influence brand connections. Thus,

Experiment 1 tests H1 and H2. We test whether group membership or desired membership in an

aspiration group interacts with the group’s brand usage such that consumers are more likely to

have a self-brand connection with a brand used by their member or aspiration group.

Method

Participants. Forty-five undergraduate student participants at a public, Southwestern

university participated in this study to meet an introductory marketing course requirement.

Selection of Stimulus Brands. In a pretest designed to select brands that are meaningful to

the participant population, 20 students at the same university were asked to list up to five brands

they considered to be “really cool” and up to five brands they would “never use.” The 10 brands

listed most often by students (across both categories) were used to measure self-brand

connections in Experiment 1 (focal brands), with the next 10 brands used as filler brands for the

question about each group’s brand usage. These brands are listed in Table 1.

11
 Procedure. In a pencil and paper survey, participants first identified up to ten typical

social groups or student "types" found on their campus. They were provided with an example of

what was meant by “types” of people using seven types of elderly people. Although ten spaces

were provided, participants were told they need not complete ten types. Popular responses

include "Athletes" and "Greeks." Next, participants were instructed to write the types they had

identified on the top of each of 20 successive pages (each type was written in a blank space

provided on two pages, for up to ten types). Next, participants rated whether or not each social

group they identified was likely to use the ten focal brands and ten filler brands selected in the

pretest (see Table 1). Participants then indicated the degree to which they belonged to or aspired

to belong to each student group. In a final task, participants rated the degree to which they

themselves had formed self-brand connections with the ten focal brands. Participants were then

debriefed. The entire procedure took approximately one half-hour.

Independent Variables. Each participant listed up to 10 student types. These types are

idiosyncratic to each participant and are not of interest in the analysis. Rather, our analysis

focuses exclusively on the participants’ self-assessed fit with each type, regardless of what type

they listed. Member group fit for each type is a continuous variable, measured by the average of

three seven-point scale items: “I like the people in type X,” “I fit in/belong to the type X

category,” and “I have a negative opinion of the people in type X” (reverse scored). All three

items were anchored by not at all/very much so (α = .79). Similarly, aspiration group fit was

measured with the average of two seven-point scale items: “I respect the people in type X,” and

“I would like to be a part of the type X group” (anchored by not at all/very much so, r = .56, p <

.001). The likelihood that people of a particular student type use a brand (type use) was

12
measured on a five-point scale, anchored by never/always. These measures were taken for each

type listed by the participant.

Dependent Variables. Self-brand connections were measured using seven seven-point

scale items (see Table 2). The seven items were averaged to form one self-brand connection

score per participant per brand (α = .90).

Results

Participants listed an average of 4.74 types. The model used in the analyses to predict

self-brand connections is a within-subjects ANOVA model, with the categorical subject variable

first, followed by the type variable, brand variable, either member group fit (H1) or aspiration

group fit (H2), type use, and the member group fit (H1) or aspiration group fit (H2) by type use

interaction.1 Self-brand connections is a continuous variable, while member group fit, aspiration

group fit, and type use were each dichotomized using a median split.2

Hypothesis 1. This hypothesis asserts that the degree to which a consumer perceives

him/herself as fitting into a member group will moderate the effect of that group's brand usage

on the consumer's self-brand connection. H1 is supported by the data: the member group fit by

type use interaction is significant (F1,125 = 7.47, p < .01). Figure 1 shows the significant

interaction (Type use also has a significant main effect that is qualified by the significant

interaction [F1,125 = 33.20, p < .001].) For participants who identified with the group, whether the

group used the brand or not matters more than for those participants who did not identify with

the group.

---------------------------------
Insert Figure 1 about here
---------------------------------

13
 Hypothesis 2. This hypothesis asserts that the degree to which a consumer wishes to

belong to an aspiration group will moderate the effect of that group's brand usage on the

consumer's self-brand connection. Hypothesis 2 is supported by the data: the aspiration group fit

by type use interaction is significant (F1,125 = 7.74, p < .01). Figure 2 shows the significant

interaction (Type use also has a significant main effect that is qualified by the significant

interaction [F1,125 = 35.13, p < .001].) For participants who aspired to belong to the group,

whether the group used the brand or not matters more than for those participants who did not

aspire to belong.

---------------------------------
Insert Figure 2 about here
---------------------------------

Discussion of Results from Experiment 1

Experiment 1 provides empirical support for the notion that students are more likely to

develop a self-brand connection when there is a strong perceived usage association between the

member group and the brand and there is a strong connection between the member group and the

consumer’s self-concept or possible self (H1). The same relationships exist between aspiration

groups, brands, and self-brand connections (H2). When these scenarios exist, the consumer may

adopt the brand to meet a self-need, such as enhancement or consistency. Neidenthal et al.

(1985) demonstrate that people choose brands based on self-needs; however, Experiment 1 goes

beyond brand choice to show that member and aspirational group influence affects self-brand

connections.

Experiment 2 extends the results of Experiment 1 by considering the role of different

self-motives, such as self-enhancement and self-verification (H3). We argue that people

motivated to enhance their self-concept form connections to brands that are used by groups to

14
which they aspire to belong, whereas people motivated to verify their self-concept form

connections to brands used by groups to which they already belong.

EXPERIMENT 2

In this study, we examine the role of two self-motivations, self-enhancement and self-

verification (H3), in addition to revisiting H1 and H2. By differentiating self-enhancement goals

from self-verification goals, we are able to provide some support for the idea that brand

associations are transferred from a reference group to an individual as they construct their self-

concepts in a manner consistent with their predominant self-motivations. If individuals use the

user imagery and psychological benefits of brands to create their self-concept, then individuals

with different types of goals should be differentially influenced by the two types of groups that

provide these associations for brands. Specifically, self-enhancers will form connections to

brands used by aspiration groups to achieve their need to look better to themselves and others,

while self-verifiers will form connections to brands used by groups to which they actually belong

in order to achieve their goal of verifying or maintaining consistency with their existing self-

concept.

Method

Participants. As part of their requirements in an introductory marketing course and

marketing research course, 171 undergraduate student participants at a public, Southwestern

university participated in this study.

Stimulus Brands. A series of pretests were run to select 12 brands that are relevant to the

participant population and have specific brand images (that is, people perceive that very specific

user types are associated with them) as opposed to brands that are widely used by the population

15
in general. These brands are listed in Table 1. Although it is possible for people to develop self-

brand connections to almost any brand, in general, people should have stronger positive or

negative connections to brands with specific user images. These are the brands that communicate

the most about the user and thus are better suited for being incorporated into one’s self-concept.

Procedure. This study used a Visual Basic® program installed on the network in a

university decision behavior lab that allowed us to customize the experiment based on

participants' responses. After a short study introduction, participants entered up to six (at least

two) types of students found on campus, using instructions similar to those in Experiment 1. The

computer program enabled us to use these participant-specific types throughout the experiment

and randomize the order of the types and brands being evaluated. Next, participants completed a

series of scale questions that were answered on 1 to 100 sliding scales, using a mouse. First, they

indicated the degree to which they "fit" with each student type, both in terms of current group

membership and a wish to belong in the future (three scale items each, see below). Participants

then rated whether or not each social group they identified was likely to use the 12 brands in

Table 1. After a short, unrelated filler task, participants rated the degree to which they had

formed self-brand connections with these brands. At the end of the experiment, we measured the

degree to which participants had self-verification goals vs. self-enhancement goals, followed by

demographic information. The program ended with a debriefing statement. The entire procedure

took approximately one half-hour.

Independent Variables. Each participant listed up to six student types. As in Experiment

1, our analysis focuses exclusively on the participants’ self-assessed fit with the type, regardless

of what type they listed. In Experiment 2, all variables were measured on a 1-100 scale. Member

group fit was measured by the average of three scale items: I fit in with this group of people, I

16
belong to this group, and I consider myself to be this type of person. All three items were

anchored by strongly disagree/strongly agree (α = .96). Aspiration group fit was also measured

by the average of three scale items: I would like to be a part of this group, I look up to this type

of person, and I wish I had more friends in this group, again anchored by strongly

disagree/strongly agree (α = .90). The degree to which the student type was perceived as using

the brand (type use) was measured on one scale: to what extent do the X type of people use the

following brands (where the participant’s own type label replaces the “X”), anchored by

definitely does not use/definitely does use. Self-enhancement was measured with a single item, it

is important that people see me in the best possible light, while self-verification was also

measured with one item, it is important for me to have accurate information about my strengths

and weaknesses. Both items were anchored by strongly disagree/strongly agree (Swann 1990).

While the two items are significantly correlated (r = .26, p < .001), the relationship is not

particularly strong.

Dependent Variables. Self-brand connections were measured using the average of three

of the seven scale items used in Experiment 1, on a 1 to 100 scale: I feel a personal connection to

this brand, I can identify with this brand, and this brand reflects who I am (α = .93). These items

had the largest item to total correlation in Experiment 1.

Results

In this experiment, participants listed an average of 5.39 types. The model used in the

analyses to predict self-brand connections for H1 and H2 is a within-subjects ANOVA model,

with the categorical subject variable first, followed by the Brand variable, type variable, either

member group fit (H1) or aspiration group fit (H2), type use, and the member group fit (H1) or

aspiration group fit (H2) by type use interaction. The model used to test H3 is a mixed between

17
and within design, which is described below. Self-brand connections is a continuous variable,

while member group fit, aspiration group fit, type use, self-verification, and self-enhancement

were each dichotomized using a median split.

Hypothesis 1. This hypothesis asserts that the degree to which a consumer perceives

him/herself belonging to a member group will moderate the effect of that group's brand usage on

the consumer's self-brand connection. H1 is supported by the data: the member group fit by type

use interaction is significant (F1,172 = 25.67, p < .001), as shown in Figure 3 (Type use also has a

significant main effect that is qualified by the significant interaction [F1,172 = 39.80, p < .001].)

---------------------------------
Insert Figure 3 about here
---------------------------------

Hypothesis 2. This hypothesis proposes that the degree to which a consumer would like

to belong to an aspiration group will moderate the effect of that group's brand usage on the

consumer's self-brand connection. Hypothesis 2 is also supported by the data: the aspiration

group fit by type use interaction is significant in the model (F1,169 = 9.57, p < .01) (Type use also

has a significant main effect that is qualified by the significant interaction [F1,169 = 31.22, p <

.001].)

---------------------------------
Insert Figure 4 about here
---------------------------------

Hypothesis 3. This hypothesis asserts that consumers who have self-verification goals

will be more strongly influenced by their member groups, whereas consumers who have self-

enhancement goals will be more strongly influenced by aspiration groups. For this analysis, we

used participants who had either high self-enhancement goals, but low self-verification goals

(self-enhancers, SE) or high self-verification goals, but low self-enhancement goals (self-

18
verifiers, SV), using median splits to divide up the participants and eliminating those who were

high on both or low on both categories.3 To dichotomize member groups versus aspiration

groups, we first created an index of member group score minus aspiration group score that

ranged from +100 to -100. Next, we eliminated any type that scored between +50 to - 50 on the

index. In order to avoid unequal weighting of participants based on the number of times they are

represented in the dataset (which directly corresponds to the number of types they entered), we

only included the two most extreme types per participant (i.e., no participant in the analysis has

more than two types). This resulted in 46 participants, averaging 1.49 types per participant,

across 12 brands per type, for a total of 804 observations.4 The ANOVA model consists of one

between-subjects factor, self-motive (self-verifier vs. self-enhancer), followed by a subject

variable, brand variable, continuous type use variable, group type (member vs. aspiration), three

two-way interactions (self-motive by type use, self-motive by group type, and type use by group

type), and one three-way interaction (self-motive by type use by group type). 5

As predicted by H3, we find a significant three-way interaction for self-motive (self-

verifier vs. self-enhancer) by type use by group type (member or aspiration) on self-brand

connections (F1,50 = 3.68, p < .05, one-tailed). The nature of the three-way interaction, shown

graphically in Figure 5, supports our assertions. Specifically, for self-verifiers, when the member

group uses the brand, self-brand connections are significantly higher compared to when member

groups do not use the brand (F1,50 = 13.15, p < .001), while this is not true for aspiration groups

(F1,50 < 1.0). In the case of self-enhancement goals, self-brand connections are higher when

aspiration groups use the brand, compared to when aspiration groups don’t use the brand (F1,50 =

5.79, p < .05). However, the difference in self-brand connections when member groups use the

brand versus don't use the brand approaches significance (F1,50 = 2.66, p = .11). Importantly, the

19
effect of aspiration groups is significantly larger than the effect of member groups (F1,50 = 4.43, p

< .05) for self-enhancers (and the effect of member group usage on self-brand connections is

larger for self-verifiers compared to self-enhancers [F1,50 = 2.92, p < .05, one-tailed]), supporting

H3.

---------------------------------
Insert Figure 5 about here
---------------------------------

Discussion of Results from Experiment 2

Experiment 2 replicates H1 for member groups and H2 for aspiration groups.

Furthermore, the addition of self-motives (H3) provides insight into the self-construction

processes used by consumers. We find that people motivated to enhance their self-concept form

connections to brands that are used by groups they aspire to belong to, while people motivated to

verify their self-concept form connections to brands used by groups to which they already

belong. By differentiating self-enhancement goals from self-verification goals, we provide some

support for the idea that user imagery and psychological benefit associations are appropriated

from certain reference groups by individuals who construct their self-concepts in a manner

consistent with their predominant self-motivations. Individuals with different types of goals

appear to be differentially influenced by the member groups vs. aspiration groups that provide

images with which to create their self-concept. Specifically, self-enhancers can form connections

to brands used by aspiration groups to achieve their need to look better to themselves and others,

while self-verifiers can form connections to brands used by groups to which they actually belong

in order to achieve their goal of verifying or maintaining consistency with their existing self-

concept.

20
 CONCLUSION

These two studies examining undergraduate students’ attitudes and beliefs about

prototypical student types and the fit between brands and these student types indicate that

students are more likely to develop a self-brand connection when there is a strong usage

association between a reference group and the brand and there is a strong connection between

the reference group and the consumer’s self-concept. When this scenario exists, the consumer

may appropriate user imagery and psychological benefit associations of the brand to meet a self-

need, such as self-enhancement or the need for consistency. Additionally, in the case of self-

verifiers, member groups have a larger effect on self-brand connections, while for self-

enhancers, the effect of aspiration groups on self-brand connections is greater. We believe that

this finding is an important demonstration that consumers are motivated by their self-needs to

utilize brand associations derived from different types of groups in a contingent fashion to

construct and present their self-identities.

Marketing Implications

The set of associations consumers have about a brand is an important component of brand

equity, and we believe that forming a self-brand connection is a psychological manifestation of

such equity at the consumer level. For example, self-brand connections may lead to robust brand

attitudes, that is, attitudes that are not very susceptible to change. Consumers who have used brand

associations to construct their self-identities may be more forgiving of marketer blunders, be it a

poor advertising campaign or a temporary product quality problem. They may also be more brand

loyal and less likely to switch to competitors’ brands in response to price cuts, special displays,

bundling tactics, and coupons. Therefore, the notion that consumers form a link to a brand as they

use the brand’s associations for self-construction is important to marketing managers. When

21
consumers’ self-concepts are linked to a brand, then the company behind the brand may be able to

gain an enduring competitive advantage, because this type of connection is difficult for competitors

to imitate.

Limitations and Directions for Future Research

The main limitation of this research is the correlational nature of our data. Although our

results are consistent with the processes we propose (particularly in the case of H3), we do not

demonstrate causality in our two studies. One could argue that our participants were projecting

their own behavior onto others in their member or aspiration groups. Future research could

explore potential moderators to help establish causality. For example, in Experiment 2, we used

brands with specific images, since they are better suited for being incorporated into one’s self-

concept. However, our model would indicate that the effect of reference group brand usage on

self-brand connections brands should not occur for brands with broad, non-specific user

associations. Another similar moderator to explore is public versus private goods (Bearden and

Etzel 1982). We would expect a larger effect for public goods due to their increased visibility,

which makes them better suited for communicating one's self-concept to others.

Future research could also examine how consumers appropriate brand associations from

sources such as celebrities and sports figures to construct and communicate their own self-

identities. Specifically, in the case of celebrity influence on brand associations, consumers'

assessments of the degree of congruency between the celebrity and the brand could be influenced

by either celebrity use of the brand (for example, Tiger Woods and Ping golf clubs) or by the

perceived fit between the celebrity and the brand (for example, Tiger Woods and American

Express) (see McCracken 1989). Similarly, assessments of the fit between the consumer and the

celebrity could be either based upon the degree to which the consumer would like to become

22
more like a particular celebrity (more like aspiration groups) or upon the extent to which the

consumer feels they are similar to the celebrity on a variety of dimensions (more like member

groups). The process of association creation could be examined in more detail, therefore, by

manipulating the salience of these various aspects that could influence consumers' assessments

(e.g., manipulating whether use or perceived fit is made more salient). Although such factors

might also be differentially relevant in the case of reference groups, we believe that their impact

is likely to be more pronounced in the case of celebrity influence. There may also be additional

individual difference factors that come into play with celebrities, in addition to the self-goals

examined in this paper. For example, skepticism towards celebrity advertising may moderate the

degree to which consumers accept movement of associations from the celebrity to the brand to

themselves and form self-brand connections.

23
 REFERENCES

Aaker, David A. (1991). Managing Brand Equity: Capitalizing on the Value of a Brand Name.

New York, NY: The Free Press.

Ball, A. Dwayne, and Lori H. Tasaki (1992). "The Role and Measurement of Attachment in

Consumer Behavior." Journal of Consumer Psychology, 2, 155-172.

Banaji, Mahzarin R. and Deborah A. Prentice (1994). “The Self in Social Contexts.” Annual

Review of Psychology, 45, 297-332.

Baumeister, R. F., D. M. Tice, and D. G. Hutton (1989). “Self-Presentational Motivations and

Personality Differences in Self-Esteem.” Journal of Personality, 57, 547-579.

Bearden, William O., and Michael J. Etzel (1982). “Reference Group Influence on Product and

Brand Purchase Decisions.” Journal of Consumer Research, 9, 183-194.

Bearden, William O., Richard G. Netemeyer, and Jesse E. Teel (1989). “Measurement of

Consumer Susceptibility to Interpersonal Influence.” Journal of Consumer Research, 15,

473-481.

Belk, Russell W. (1988). "Possessions and the Extended Self." Journal of Consumer Research,

15, 139-168.

Burnkrant, Robert E., and Alain Cousineau (1975). “Informational and Normative Social

Influence in Buyer Behavior.” Journal of Consumer Research, 2, 206-215.

Childers, Terry L., and Akshay R. Rao (1992). “The Influence of Familial and Peer-based

Reference Groups on Consumer Decisions.” Journal of Consumer Research, 19, 198-

211.

Dunning, David (1995). “Trait Importance and Modifiability as Factors Influencing Self-

Assessment and Self-Enhancement Motives.” Personality and Social Psychology

24
 Bulletin, 21, 1297-1306.

Escalas, Jennifer Edson (1996), “Narrative Processing: Building Connections between Brands

and the self.” unpublished dissertation, Duke University, June 1996.

Escalas, Jennifer Edson, and James R. Bettman (2000). "Using Narratives to Discern Self-

Identity Related Consumer Goals and Motivations." In Ratti Ratneshwar, David Mick,

and Cynthia Huffman (Eds.), The Why of Consumption: Perspectives on Consumer

Motives, Goals, and Desires, (pp. 237-258). New York, NY: Routledge Press,.

Fiske, Susan T., and Shelly E. Taylor (1991). Social Cognition. New York, NY: McGraw-Hill

Inc.

Folkes, Valerie. S., and Tina Kiesler (1991). “Social Cognition: Consumers’ Inferences about the

Self and Others.” In Thomas S. Robertson and Harold H. Kassarjian (Eds.), Handbook of

Consumer Behavior, (pp. 281-315). Englewood Cliffs, NJ: Prentice Hall.

Fournier, Susan. (1998). “Consumers and Their Brands: Developing Relationship Theory in

Consumer Research.” Journal of Consumer Research, 24, 343-373.

Greenwald, A. G., F. S. Belleza, and Mahzarin R. Banaji (1988). “Is Self-Esteem a Central

Ingredient of the Self-Concept?” Personality and Social Psychology Bulletin, 14, 34-45.

Huffman, Cynthia, S. Ratneshwar, and David G. Mick (2000). "Consumer Goal Structures and

Goal-Determination Processes: An Integrative Framework." In R. Ratneshwar, D. G.

Mick, and C. Huffman (Eds.) The Why of Consumption: Perspectives on Consumer

Motives, Goals, and Desires, pp. 9-35. New York, NY: Routledge Press,.

Keller, Kevin L. (1993). “Conceptualizing, Measuring, and Managing Customer-Based Brand

Equity.” Journal of Marketing, 57, 1-22.

25
Kleine, Robert E., III, Susan S. Kleine, and Jerome B. Kernan (1993). “Mundane Consumption

and the Self: A Social-Identity Perspective.” Journal of Consumer Psychology, 2, 209-

236.

Markus, Hazel, and Paula Nurius (1986). “Possible Selves.” American Psychologist, 41, 954-

969.

McCracken, Grant (1989). “Who Is the Celebrity Endorser? Cultural Foundations of the

Endorsement Process.” Journal of Consumer Research, 16, 310-321.

Miller, D. T., and M. Ross (1975). “Self-Serving Biases in the Attribution of Causality: Fact or

Fiction?” Psychological Bulletin, 82, 213-225.

Morling, Beth, and Seymour Epstein (1997). “Compromises Produced by the Dialectic Between

Self-Verification and Self-Enhancement.” Journal of Personality and Social Psychology,

73, 1268-1283.

Moschis, George P. (1985). “The Role of Family Communication in Consumer Socialization of

Children and Adolescents.” Journal of Consumer Research, 11, 898-913.

Muniz, Jr. Alber M., and Thomas C. O’Guinn (2001). “Brand Community.” Journal of

Consumer Research, 27, 412-423.

Niedenthal, Paula, M., N. Cantor, and J.F. Kihlstrom (1985). “Self to Prototype Matching: A

Strategy for Social Decision-Making.” Journal of Personality and Social Psychology, 48,

575-584

Park, C. Whan, and V. Parker Lessig (1977). “Students and Housewives: Differences in

Susceptibility to Reference Group Influence.” Journal of Consumer Research, 4, 102-

110.

26
Reingen, Peter H., B. L. Foster, J. J. Brown, and S. B. Seidman (1984). “Brand Congruence in

Interpersonal Relations: A Social Network Analysis.” Journal of Consumer Research,

11, 771-783.

Richins, Marsha L. (1994). “Valuing Things: The Public and Private Meanings of Possessions,”

Journal of Consumer Research, 21, 504-521.

Schlenker, Barry R. (1980). Impression Management: The Self-Concept, Social Identity, and

Interpersonal Relations. Monterey, CA: Brooks/Cole.

Sedikides, Constantine, and Michael J. Strube (1995). “The Multiply Motivated Self.”

Personality and Social Psychology Bulletin, 21, 1330-1335.

Setterlund, Mark B. and Paula M. Niedenthal (1993). “Who Am I? Why Am I Here?” Self-

Esteem, Self-Clarity, and Prototype Matching.” Journal of Personality and Social

Psychology, 65, 769-780.

Swann, Jr., William. B. (1990). “To Be Adored or To Be Known? The Interplay of Self-

Enhancement and Self-Verification.” In E. T. Higgins and R. M. Sorrentino (Eds.)

Handbook of Motivation and Cognition: Foundations of Social Behavior (pp. 408-448).

New York, NY: Guildford.

Swann, Jr., William B., John J. Griffin, Steven C. Predmore, and Bebe Gaines (1987). “The

Cognitive-Affective Crossfire: When Self-Consistency Confronts Self-Enhancement.”

Journal of Personality and Social Psychology, 52, 881-889.

Wallendorf, Melanie, and Eric J. Arnould (1988). "'My Favorite Things': A Cross-Cultural

Inquiry into Object Attachment, Possessiveness, and Social Linkage." Journal of

Consumer Research, 14, 531-547.

27
 TABLE 1

BRANDS USED IN EXPERIMENTS

Experiment 1 Experiment 2

Focal brands: Abercrombie & Fitch

Adidas
Adidas Birkenstock
Coca-Cola BMW
The GAP Doc Martin
Hershey FUBU
J. Crew The GAP
Kellogg’s Tommy Hilfiger
Levi’s Levi’s
The Limited Polo
Nike Mountain Dew
Reebok Nike

Filler brands:

Crest
Diet Coke
Kodak
Macanudo Cigars
Mountain Dew
Polo
Snapple
Sony
Timberland
Victoria’s Secret

28
 TABLE 2

SELF-BRAND CONNECTION SCALE ITEMS

1. Brand X reflects who I am (not at all/extremely well)

2. I can identify with Brand X (not at all/extremely well)

3. I feel a personal connection to Brand X (not at all/very much so)

4. I (can) use Brand X to communicate who I am to other people (not at

all/extremely well)

5. I think Brand X (could) help(s) me become the type of person I want to be (not at
all/extremely well)

6. I consider Brand X to be “me” (it reflects who I consider myself to be or the way
that I want to present myself to others) (not “me”/“me”)

7. Brand X suits me well (not at all/extremely well)

Scale source: Escalas 1996.

29
 FIGURE 1

SELF-BRAND CONNECTIONS IN EXPERIMENT 1

MEMBER GROUPS (H1) (7 POINT SCALE)

4.00

Does Not Belong

3.50 to Group

Belongs to Group
3.00

2.50
Type Does Not Use Type Uses

30
 FIGURE 2

SELF-BRAND CONNECTIONS IN EXPERIMENT 1

ASPIRATION GROUPS (H2) (7 POINT SCALE)

4.00

Does Not Wish to

3.50 Belong to Group

Wishes to Belong
3.00 to Group

2.50
Type Does Not Use Type Uses

31
 FIGURE 3

SELF-BRAND CONNECTIONS IN EXPERIMENT 2

MEMBER GROUPS (H1) (1 – 100 SCALE)

55

Does Not Belong

45 to Group

35 Belongs to Group

25
Type Does Not Use Type Uses Brand
Brand

32
 FIGURE 4

SELF-BRAND CONNECTIONS IN EXPERIMENT 2

ASPIRATION GROUPS (H2) (1 – 100 SCALE)

55.00

Does Not Wish to

45.00
Belong to Group

35.00 Wishes to Belong

to Group

25.00
Type Does Not Use Type Uses

33
 FIGURE 5

SELF-BRAND CONNECTIONS IN EXPERIMENT 2

ASPIRATION GROUPS VS. MEMBER GROUPS (H3) (1 – 100 SCALE)

SELF-VERIFICATION GOALS

55

45 Type Does Not

Use
Type Uses
35

25
Aspriration Group Member Group

SELF-ENHANCEMENT GOALS

55

45 Type Does Not

Use
Type Uses
35

25
Aspiration Group Member Group

34
1
We use the member group fit (H1) or aspiration group fit (H2) by type use by subject

interaction as the error term in the models to test H1 and H2 throughout the paper.
2
Analysis results using the continuous measures of member group fit, aspiration group fit, and

type use are virtually identical to the categorical analysis results presented in the paper.
3
Of 170 participants, 47 were low SV, low SE; 38 were high SV, low SE; 36 were low SV, high

SE; and 49 were high on both SE and SV. When a model is run with all four groups, the three-

way interaction is significant and the qualitative results are consistent with H3.
4
After eliminating types that were neither high member/low aspiration groups nor high

aspiration/low member groups, some participants in the dataset were left with only one type.
5
For the three way interaction and all contrasts, the mean square of self-motive by type use by

group type by subject is used as the error term. Additionally, type use is dichotomized using a

median split for the contrasts.

35