This Brand's for Me: Brand Personality and User Imagery Based Self-Congruity

Material Information

This Brand's for Me: Brand Personality and User Imagery Based Self-Congruity
PARKER, BRIAN THOMAS ( Author, Primary )
Copyright Date:


Subjects / Keywords:
Brand image ( jstor )
Brands ( jstor )
Consumer behavior ( jstor )
Consumer research ( jstor )
Mathematical congruence ( jstor )
Modeling ( jstor )
Psychological attitudes ( jstor )
Self ( jstor )
Self image ( jstor )
Self perception ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Brian Thomas Parker. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
Resource Identifier:
496175031 ( OCLC )


This item is only available as the following downloads:

Full Text




Copyright 2005 by Brian Thomas Parker


This document is dedicated to my family.


iv ACKNOWLEDGMENTS This dissertation was made possible by Marilyn Roberts, Gail Baker, David Miller, and particularly John Sutherland, each providing unique intellectual and personal support often beyond the scope of the academic endeavor, contributing to my growth as an individual and academic professional.


v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES.............................................................................................................x ABSTRACT.......................................................................................................................xi CHAPTER 1 INTRODUCTION........................................................................................................1 Research Problem.........................................................................................................4 Dissertation Agenda......................................................................................................7 2 LITERATURE REVIEW...........................................................................................10 Self-Congruity Theory................................................................................................11 Self-Theory and Self-Image................................................................................12 Antecedents of Self-Brand Congruity.................................................................15 Image Congruence Hypothesis............................................................................16 Theoretical Paradigms: Image M easurement and User Imagery........................18 Congruity Indicator Scores..................................................................................21 “Global” Self-Congruity Measure.......................................................................22 Theoretical Summary..........................................................................................23 Brand Image................................................................................................................23 Brand Personality................................................................................................26 Brand personality source variables..............................................................27 Brand personality scales and framework.....................................................28 Symbolic function of brand personality.......................................................31 Brand Attitude............................................................................................................32 Summary of Literature Review..................................................................................35 Research Objectives and Questions............................................................................36 Research Objective One......................................................................................38 Research Objective Two......................................................................................38


vi 3 METHODOLOGY.....................................................................................................40 Brand Selection...........................................................................................................40 Questionnaire Development.......................................................................................43 Brand Attitude Op erationalization......................................................................43 User Image Self-Congruity (UIC) Operationalization........................................44 Brand Personality Self-Congruity (BPC) Operationalization.............................45 Pilot Tests...................................................................................................................47 Primary Survey...........................................................................................................48 Description of Data Analysis......................................................................................48 4 RESULTS...................................................................................................................50 Brand Familiarity and Attitude...................................................................................50 Brand Personality........................................................................................................51 Research Questions.....................................................................................................53 Actual-BPC vs. Actual-UIC................................................................................54 Abercrombie & Fitch...................................................................................55 Birkenstock...................................................................................................56 Banana Republic..........................................................................................57 Nike..............................................................................................................58 Ideal-BPC vs. Ideal-UIC.....................................................................................59 Abercrombie & Fitch...................................................................................59 Birkenstock...................................................................................................60 Banana Republic..........................................................................................61 Nike..............................................................................................................62 Simultaneous Comparison of Congruity Types..................................................63 Abercrombie & Fitch...................................................................................64 Birkenstock...................................................................................................66 Banana Republic..........................................................................................67 Nike..............................................................................................................69 Summary of Results....................................................................................................71 5 DISCUSSION AND CONCLUSION........................................................................74 Theoretical Implications.............................................................................................75 Research Question One.......................................................................................75 Research Question Two.......................................................................................76 Research Question Three.....................................................................................77 Theoretical Summary..........................................................................................79 Professional Implications............................................................................................80 Limitations and Future Research................................................................................81 Conclusion..................................................................................................................84 APPENDIX A QUESTIONNAIRE....................................................................................................85


vii B DESCRIPTIVE STATISTICS: INDEPENDENT VARIABLES..............................92 LIST OF REFERENCES...................................................................................................93 BIOGRAPHICAL SKETCH.............................................................................................98


viii LIST OF TABLES Table page 2-1. Independent variables: self-brand congruity indicators............................................37 3-1. Top twenty-five p ublicly consumed brands.............................................................41 3-2. Brand Personality Scale (B PS) dimensions and indicators......................................46 4-1. Brand attitude i ndices: desirability, impressi on, and purchase intentions................51 4-2. Brand personality......................................................................................................52 4-3. Respondent actual and ideal BPS self measures......................................................53 4-4. Abercrombie & Fitch correlatio n matrix: actual BPC and actual UIC.....................55 4-5. Abercrombie & Fitch stepwise re gression: actual BPC and actual UIC..................55 4-6. Birkenstock correlation ma trix: actual BPC and actual UIC....................................56 4-7. Birkenstock stepwise regr ession: actual BPC and actual UIC.................................57 4-8. Banana Republic correlation ma trix: actual BPC and actual UIC............................57 4-9. Banana Republic stepwise regr ession: actual BPC and Actual UIC........................58 4-10. Nike correlation matrix : actual BPC and actual UIC..............................................58 4-11. Nike stepwise regression: actual BPC and actual UIC............................................59 4-12. Abercrombie & Fitch correlati on matrix: ideal BPC and ideal UIC.......................60 4-13. Abercrombie & Fitch stepwise re gression: ideal BPC and ideal UIC.....................60 4-14. Birkenstock correlation matrix: ideal BPC and ideal UIC......................................61 4-15. Birkenstock stepwise regres sion: ideal BPC and ideal UIC....................................61 4-16. Banana Republic correlation ma trix: ideal BPC and Ideal UIC..............................62 4-17. Banana Republic st epwise regression: ideal BPC and ideal UIC............................62


ix 4-18. Nike correlation matrix : ideal BPC and ideal UIC..................................................63 4-19. Nike stepwise regressi on: ideal BPC and ideal UIC...............................................63 4-20. Abercrombie & Fitch correlati on matrix: actual/ideal BPC and UIC.....................64 4-21. Abercrombie & Fitch stepwise regression: actual/ideal BPC and UIC...................65 4-22. Birkenstock correlation ma trix: actual/ideal BPC and UIC....................................66 4-23. Birkenstock stepwise regres sion: actual/ideal BPC and UIC..................................67 4-24. Banana Republic correlation ma trix: actual/ideal BPC and UIC............................67 4-25. Banana Republic st epwise regression: actu al/ideal BPC and UIC..........................68 4-26. Nike correlation matrix : actual/ideal BPC and UIC................................................69 4-27. Nike stepwise regression: actual/ideal BPC and UIC.............................................70 4-28. Summary of variance explained in brand attitude by c ongruity indicators.............71 4-29. Summary table: actua l congruity indicators............................................................72 4-30. Summary table: ideal congruity indicators..............................................................73


x LIST OF FIGURES Figure page 2-1. Illustration of the image congruence hypothesis........................................................17 2-2. Self-brand congruity scales example..........................................................................18 2-3. Absolute simple-difference model..............................................................................21 2-4. Brand personality framework.....................................................................................30 2-5. Cognitive and affective brand image factors..............................................................34 3-1. Selected brands and image descriptors.......................................................................42 3-2. Difference squared model...........................................................................................47


xi Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THIS BRAND’S FOR ME: BRAND PERSONALITY AND USER IMAGERY BASED SELF-CONGRUITY By Brian Thomas Parker December 2005 Chair: John Sutherland Major Department: Journalism and Communications A vital benchmark in the development of brand communication strategies is an understanding of what drives favorable brand at titudes, a determinant of future consumer behavior. At times, consumers have favorable attitudes towards brands that project images suitable for self-expressive motivations. In this domain, most research falls within the self-congruity model of consumer be havior. Self-brand congruity, the match or mismatch between one’s self-image and a br and’s image, partially drives consumer behavior by influencing the favor ability of brand attitudes. To study the self-brand congruity ph enomenon, researchers measure the similarity, that is congruence, between consumer’s self-per ceptions and their perceptions of a brand’s image. The dominant theoretica l paradigm conceptualizes brand image as consumers’ view of a brand’s stereotypical us er, referred to as brand user imagery. In a slightly different paradigm, researchers use the construct brand pe rsonality rather than user imagery to study self-brand congruit y. However, no studies have empirically


xii compared these two brand image constructs within the framework of self-congruity theory, providing the motiva tion for the present study. This dissertation was an exploratory study designed to co mpare self-brand congruity operationalized with both the br and personality and brand user imagery constructs as a basis for modeling the structure of self-expressive brand attitudes. To facilitate the study, the author proposed two c oncepts that represent different types of self-brand congruity: user image self-congruity (UIC ) and brand personality selfcongruity (BPC). A survey of 272 college students collected UIC, BPC, and attitude ratings for a set of socially consumed brands, predetermined in preliminary research to be relevant to the target respondents: Abercrombie & Fitc h, Birkenstock, Banana Republic, and Nike. Correlation and stepwise regression procedures illustrated (1) that UIC and BPC are not linear dependent and add unique explanatory pow er in brand attitude prediction models, (2) UIC was a stronger predictor of brand at titude, and (3) the re lationship between UIC and BPC differed across brand types. Theoreti cal and professional implications, as well as limitations and areas for future research, are discussed.


1 CHAPTER 1 INTRODUCTION A primary role of advertising and rela ted promotional communications is to convey a brand’s core message in support of an overall brand positioning strategy, typically intended to garner favorable c onsumer reactions and influence purchase decisions. Towards this end, researchers stri ve to understand what factors predispose target consumers to like or dislike products and brands and then adjust strategies accordingly. Ultimately, the reasons underlying consumers’ affection towards brands and products exert a strong influence on their purch ase decisions, while brands that promise to satisfy needs and wants better than competitive choices garner favorable responses. Most research in this domain falls w ithin attitude theory, which provides the theoretical basis for explaining and researching consumers’ li kes and dislikes in relation to different needs that drive consumption be havior. The present dissertation fits within the general theoretical area of attitudes, specifically desi gned towards an understanding of the structure of brand att itudes in regards to the relatio nship between consumers’ selfexpressive needs and the image projected by brands. Derived from social psychology, the concept “attitude” is central to the field of communication research, in particular advert ising and related fields, such as brand management (Batra, Myers & Aaker, 1996) . Communication scholar s study attitudes because they are insightful in predicting cons umer behavior, such as purchase and loyalty (Broniarczyk & Alba, 1994; Ke ller & Aaker, 1992; Mitchell & Olson, 1981). Formally defined, an attitude is “a l earned predisposition to respond in a consistently favorable or


2 unfavorable manner with respect to a gi ven object” (Fishbein & Ajzen, 1975 p. 6). Fishbein and Ajzen’s (1975) attitude model, widely applied in advertising and brand research, distinguishes among three psychol ogical constructs that drive behavior: cognition, affect, and conation. Affect, the evaluative component of attitudinal predispositions, is the positive or negative evaluations of an object along bi polar dimensions (e.g., good vs. bad, like vs. dislike, favorable vs. unfavorable). Beliefs and opinions, the cognitive components of attitudes, are the “informational base” that ultimately determines affect and intentions, and are considered the “fundamental” buildi ng blocks of attitudes (Fishbein & Ajzen, 1975; Kardes, 1999). Most important, all att itudes consist of beliefs, though not all beliefs are attitudes; hence, it is in the cogni tive belief structure (i.e., the informational base) that one finds reason for attitudi nal evaluations and judgments (Katz, 1960). Conations, personal action tendenc ies, are intentions to perf orm the behavior in question. Behaviors are “observable acts” typically studied in their own right; it is predispositions to act that matter most in attit ude research (Fishbein & Ajzen, 1975). Because humans have limited mental capacity, usually focusing on and remembering salient information, ultimately, a person’s attitude towards an object is a function of “salient beliefs” and the evalua tion of those beliefs regarding the object (Fishbein & Ajzen, 1975; Keller, 1993; W ilkie, 1994). Attitude towards a brand (Ab), referred to as brand attitude, is a key i ndicator of consumer market behavior and primarily a function of salient brand imag e beliefs both functional and non-functional (i.e., symbolic). In regards to the multi-a ttribute model, brand image represents the cognitive belief structure that influences the positive or nega tive evaluation of the brand,


3 and hence determines consumers’ predisposi tions towards action. Therefore, the brand image construct provides insight into the structure of consumer attitudes towards the brand. Essentially, brand image is a subjec tive perception of functional and nonfunctional information regarding the produc t or service (Patterson, 1999; Sutherland, Marshall & Parker, 2004). When consumers characterize a brand image, they think about two types of information rela ted to the brand: attributes and benefits (Aaker, 1996; Keller, 1993). Attributes are de scriptive features that repres ent what a consumer thinks that a product has or does and ascribes to the brand. Brand attribut es are either product related, that is, features needed for performing the pr oduct function, or non-product related, for example, brand pe rsonality and user imagery. Brand benefits are the “personal value” (i.e., what the consumer thinks the product attribute will do for them) that consum ers ascribe to the brand. Brand benefits are consequences of attributes, cl assified as functional, experi ential (sensory), and symbolic (self-expressive) (Park, Jaworski & Macinnis , 1986). Both functional (i.e., desire for problem removal) and experienti al (i.e., what is feels like to use a product) benefits are associated with product related attributes (Keller, 1993). In contrast, symbolic brand benefits relate to non-product attributes and are associated with individuals’ social and self-expressive needs (Solomon, 1983). Brand at titudes based on symbolic brand benefits (e.g., brand personality) can se rve a “value-expressive” f unction by allowing consumers to express their self-image (A aker, 1996; Katz, 1960). In this domain, much research falls within the self-congr uity theory framework (Sirgy, 1986).


4 Self-congruity theory adds to the discu ssed attitudinal model, by proposing that brand attitudes are partially a function of the self-congruity phenomenon. Self-congruity is the mental process of comparing self to ot her objects and stimuli, characterized as the match or mismatch between consumer self-ima ge and a product image, brand image, or company image (Sirgy, 1986). Self-brand congru ity, the comparison of self to brand, affects brand attitudes particularly when the so cial signaling value of a brand is high (i.e., used in a public situation) and when symbo lic, self-expressive motiv ations are involved. Positive brand attitudes should result as th e similarity between brand image and the consumer’s self-image increases. Self-congruity theory is important fo r communication researchers because it facilitates an understanding of how to build favorable bra nd attitudes. However, selfbrand congruity researchers, until recently, have paid little attention to multiple facets of the brand image construct, limiting the explanatory power of the congruity model in regards to the structure of consumer brand attitudes. The primary empirical goal of the present dissertation was to examine the stru cture of brand attitude s in regards to the conceptualization of the brand image constr uct within the congru ity theory framework. Doing so will expand the range of phenome non explained by the self-brand congruity model and lead to a better unders tanding of the role of differe nt brand image constructs in building favorable brand attitudes. Research Problem The notion that brands communicate information beyond functional product utility that is psychologically important to the consumer motivates an important area of consumer research. Levy (1959) often receives credit for initiating this debate, arguing that consumers are not functionally orient ed. As a result, consumer behavior is


5 significantly affected by the “symbols” us ed to identify goods, primarily the image projected by different products and bra nds. Scholars from across disciplines (e.g., psychology, anthropology, and consumer behavior) argue this is the case because individuals often use symbo lic brand meaning for pers onal-expression and social communication in order to sustain a soci al self (Belk, 1988; Katz, 1960; McCraken, 1986; Zinkham & Hong, 1991). To research th e effect of self-e xpressive needs on consumer brand attitudes, research ers employ self-congruity theory. Self-congruity theory (Sirgy, 1986), grounded in the behavioral sciences, provides a framework for studying the structure of c onsumer brand attitudes via the self-brand congruity phenomenon. Self-brand congruity is an internal (i .e., mental) comparison that consumers make in regards to the similarity or dissimilarity of a brand’s image and their own self-image (i.e., the way one sees him/hers elf to be as an indi vidual) (Dolich, 1969; Gould, 1991; Graeff, 1996; Sirgy, 1982, 1986; Sutherland et al., 2004). The central premise of congruity theory suggests that as the similarity (i.e., image congruence) between one’s self-image and a brand’s image increases, so should the favorability of brand evaluations and hence the likelihood of positive action (e.g., purchase) in regards to that brand (Gould, 1991; Graeff, 1996; Sirgy, 1982, 1986). In other words, self-brand congruity partially drives consumer behavior by effecting attitudes toward the brand. The dominant paradigm in self-brand congr uity research is the comparison of a consumer’s self-image with the image of the typical brand user . According to selfcongruity theorists, “brand image” is best c onceptualized as consum ers’ perceptions of the brand’s stereotypical or ge neralized user (i.e., user im agery) and “self-image” often defined as a multidimensional construct, consisting of perceptions of an “actual self” and


6 an “ideal self” image (Sirgy, 1986). Research ers traditionally operationalize self-brand congruity by measuring responde nts’ perceptions of a brand’ s image and their self-image perceptions, and then calculat e similarity or distance scores that represent the image congruence between brand user image and self-image (Birdwell, 1968; Dolich, 1969; Graeff, 1996; Grubb & Grathwhohl, 1967; Sirgy, 1986). To facilitate discussion and the present research agenda, the author refers to this paradigm of self-brand congruity operationalization as “user image congruity” (UIC). In a slightly different paradigm, resear chers interested in the construct brand personality (i.e., the brand personified as a person), considered by some a primary component of a brand’s overall marketplace image, have applied the congruity theory framework towards its investig ation. Rather than employing us er imagery to assess brand image, researchers employ the brand personal ity construct in its place to study self-brand congruity (Aaker, 1996; Hogg, Alastair, & Ke lling, 2000). In other words, researchers operationalize self-brand congruity by comp aring brand personality and self-image measures. The author refers to this paradi gm as “brand personality congruity” (BPC). In doing so, researchers make the assumpti on that the brand personality and user imagery constructs are theoreti cally transposable, or at best , overlooked their conceptual differences and the empirical implications of doing so. Literature cl early indicates these two brand image facets are conceptually distinct; brand personality is a more encompassing and direct representation of a brand’s composite image, derived from multiple direct and indirect sources vari ables, such as product endorsers, company personal, product user, and marketing co mmunications (Aaker, 1996; Patterson, 1999;


7 Plummer, 2000). In comparison, consumers’ us er imagery percepti ons are only one of many variables that may or may not compose a brand’s overall image. Nonetheless, literature does not report a ny studies that have compared the brand personality and user imagery constructs in th e contexts of the self -brand congruity model. Since self-brand congruity is an experience th at results from compar ing one’s self view with the image a brand projec ts, then different aspects ot her than user imagery that comprise a brand’s marketplace image (e .g., brand personality) may result in unique image congruence experiences and should be em pirically cataloged in regards to their relationship to brand evaluations . Therefore, it is the overall goal of this dissertation to provide new insight into the self-brand congr uity model, by incorpor ating a more holistic conceptualization of brand image, which includes both brand personality and user imagery. Dissertation Agenda The primary research objective of this dissertation was to attain a better understanding of the structure of brand att itudes via the self-brand congruity phenomenon by focusing attention on different facets of the br and image construct. In its essence, selfbrand congruity is a comparison of simila rity between self and brand. Substantial research evidences different dimensions of th e self-image construct and is standard in self-congruity research studies. As argued, br and image does not consist of a single facet (i.e., user imagery), and therefore self-bra nd congruity should differ depending on which aspect of a brand’s image individuals compar e to their own self-image. Specifically, the designed study’s objective was to compare th e constructs brand pe rsonality and user imagery in regards to the self-congruity mode l when used independently to operationalize self-brand congruity as a basis for mode ling self-expressive brand attitudes.


8 When used independently to operationa lize self-brand congruity, two different types of self-brand congruity are evident for comparison: “user image congruity” (UIC) and “brand personality congruity” (BPC). As discussed, UIC represents the traditional paradigm, the congruity experienced between brand “user image” perceptions and selfimage, and BPC is the congruity experienced between “brand pers onality” perceptions and self-image. Moreover, combining BPC a nd UIC independently with the two selfimage dimensions (i.e., ideal self-image and ac tual self-image), four different congruity variables (i.e., indicators) are available fo r comparison: actual-BPC, actual-UIC, idealBPC, and ideal-UIC. To achieve the research objectives, a multi-phased study was designed to (1) generate brands for analysis that are public ly consumed and commonly used by the target respondents, (2) develop a quest ionnaire via pre-testing desi gned to assess user image congruity (UIC), brand personal ity congruity (BPC), and brand attitudes, and (3) survey a large sample of respondents to ensure robust st atistical power. In rega rds to analysis, the four congruity indicators menti oned above served as independ ent variables, and attitude toward the brand (Ab) was the primary dependent variab le. In addition, because brands are the primary units of analys is, it was necessary to conduc t separate analysis for each brand analyzed. Since there are no published reports of a comparison as conducted in this study, there were no grounds to develop or test spec ific hypotheses. The author’s intent was to determine if differences exist between UI C and BPC and empirically verify their conceptual differences and/or similarities in self-brand congruity research. This will allow future researchers to have a basis to test hypotheses. Chapter 2 further synthesizes


9 relevant literature in regards to theoretical models and the measurement tools used in the study design, and presents formal research que stions used to guide data collection and analysis.


10 CHAPTER 2 LITERATURE REVIEW Understanding why consumers like or dislik e a brand is central to managing an organization’s brand market position. In esse nce, the goal of bra nding is to establish favorable consumer attitudes towards one’s br and, that is, entice consumers to like one’s brand in an effort to energize product purchas e. In this endeavor, researchers study the structure of consumer attitudes, that is, why consumers like or dislike a particular brand or product, primarily to better understand how to build favorable brand attitudes. In general, consumers tend to favor brands and products that satisfy their needs and wants better than competitive choices. Consumers have a variety of different needs and wants that drive their consumption behavior patterns. Isolating and understanding the needs and wants that relate to likes or dislikes of brands and products is of utmo st importance to marketers. At times, consumers base their choices on brand evaluations; if consum ers like a brand, they are more likely to purchase it. Other times, c hoice is attribute based; for example, one may buy the least expensive brand, or the brand believed to function best on a particular attribute (e.g., fast pain relie f or refreshing taste) . One important consumer need is the consumption motivator self-expression. Consum ers at times purchase brands (e.g., Gucci and Rolex) that communicate a particular image or social role. To better understand consumers’ self-expressive needs and the infl uence on the structure of brand attitudes, much research falls in the dom ain of self-congruity theory.


11 The following literature review encompasses f our sections. First is an overview of self-congruity theory in regards to its development, the operationalization of key constructs, and empirical studies demonstra ting traditional resear ch paradigms. Second the author reviews the brand image construct, focusing on the brand personality and user imagery constructs, central to the present st udy. Since concepts are the building blocks of theory, used to guide resear ch, it is important that self -brand congruity researchers distinguish between different brand image con cepts. Third is an overview of the attitude model applied in this study and a discussi on of the self-brand c ongruity phenomenon as an addition to the model and the role of the brand image construct as the central cognitive determinant of brand evaluations. After synt hesizing the relevant literature, the final section of this chapter presents formal rese arch questions that a ddress the study’s primary empirical goals. Self-Congruity Theory Self-congruity theory suggest s that since people often use brand meaning for selfexpression, then consumers will have favorab le predispositions towards, that is like products and brands that serve to enhance perceptions of their own self-image (Sirgy, 1986). More specifically, consumers often prefer brands, products, and stores that project images similar (i.e., congruent) with how they perceive themselves to be as individuals, that is, their self-image (Belk, 1988; Graeff, 1996; Sirgy 1982; Solomon, 1983; Sutherland et al., 2004). The mental process of comparing self to other objects, called self-congruity, characterizes the match or mi smatch between consumer self-image and a product image, brand image, or company image (Sirgy, 1986). Derived from consumer psychology, the phenomenon self-congruity is a psychological (i.e., an internal) comparison that consumers make between a product,


12 store, and/or brand’s perceived image and th eir own self-image pe rceptions (Sirgy, 1986). High self-congruity occurs when perceptions of a produc t or brand’s image match perceptions of one’s self-image, and low self -congruity, incongruity, is experienced when there is a mismatch between product and/or product/brand image and self-image. Note that self-congruity is a comparison that can take place between di fferent possible stimuli and objects such as stores and brands. The present study focuses on image congruence experience between self-image and brand image termed “self-brand congruity.” In essence, the self-brand congruity e xperience is a function of two perceptual constructs: self-image and brand-image. In order to study this phenomenon, researchers typically employ psychometric techniques to measure consumers’ self and brand image perceptions and then use standard calculati ons to determine the difference or distance between the image measures (Gould, 1991; Graeff, 1996; Sirgy, 1982, 1986; Sutherland et al., 2004). Calculated congr uity indicators are typically the focal point of analysis. Hence, the conceptualization of both the self-image and br and-image constructs is of central importance when studying self-brand co ngruity. The self-image concept, derived from self-theory, provides the “major theoreti cal impetus” of self-congruity theory and explains the motivational drives of self-b rand congruity on consumer behavior (Sirgy, 1986, p. x ). Self-Theory and Self-Image Formally classified, self-c ongruity theory is a social cognition theory that involves the self (Sirgy, 1986). It is a social cognition theory because a conscious sense of “self” plays a fundamental role in life experience, interw oven in the completeness of life (Rychlak, 1981). Formally defined, the “self” is an organized, consistent, conceptual gestalt composed of percepti ons of the characteristics of the “I” or “me” and the


13 perceptions of the relatio nship of the “I” or “me” to others and to various aspects of life, together with the values attached to these perceptions (Rogers, 1959, p. 200). Self-image is a “subself” that embodies an individual’s awareness of self that can change in different social roles such as bei ng a father, co-worker, teacher, brother, friend (Sirgy, 1986). The construct self -image is central to the di scipline of social psychology, and a key variable in understanding not only c onsumer behavior but also human behavior in general. Self-image, in the most basic sens e, refers to the way in which one perceives her/himself to be as an individual, cal led a self-perception (Grubb & Grathwhohl, 1967; Graff, 1996). Self-image is a multidimensional perception of one’s self that changes from situation to situation, and is comprised of at leas t two major facets: th e “real/actual-self” and, the “ideal-self” (Aaker, 1999; G ould, 1991; Graeff, 1996 ; Sirgy, 1982, 1986; Sutherland et al., 2004; Zinkham & Hong, 1991). The real/actual self (i.e., “me as I am”), is one’s perceptions of the self as now e xperienced, and the ideal self (i.e., “the good me”), is one’s perceptions of the self as an imagined ideal, the image of the self as one desires to be (Grubb & Grathwhohl , 1967; Rogers, 1959; Sirgy, 1982, 1985). The way people perceive themselves to be as individuals signif icantly influences their behavior. According to psychologist Carl Rogers (1959), a central figure in self and personality theory, every i ndividual is motivated by a fundamental “actualizing tendency” that serves the pur pose to develop all abilities in ways that maintain and enhance one’s self-image (i.e., self-enhancem ent). Self-preservation is the postulated central motive of most if not all human functioning; people ar e motivated to maintain an


14 integrated self-image (Sirgy, 1986; Ziegler & Hjelle, 1992). The following quotations echo this perspective on human behavior. There is great deal of research which show s that the self-conce pt is, perhaps, the basis for all motivated behavior. It is th e self-concept that gi ves rise to possible selves, and it is possible selves that create the motivation for behavior. (Franken, 1994, p. 439) The self represents a totality which b ecomes a principal value around which life revolves, something to be safeguarded a nd, if possible, to be made still more valuable. An individual’s evaluation of himself will greatly influence his behavior, and thus, the more valued the self, the more organized and consistent becomes his behavior. (Gr ubb & Grathwhohl, 1967, p. 24) Individuals are not born w ith self-perceptions; one’s self-image develops over time through the processes of so cial experience and interpers onal interactions (Ziegler & Hjelle, 1992; Rychlak, 1981). From the reactions of others, in part icular “significant references” (e.g., parents, peers, teachers, etc.), an individual devel ops his/her selfimage perceptions (Grubb & Grathwhohl, 1967; R ogers, 1959). Because self-enhancement and preservation are dependent on reactions of signi ficant others, the “individual will strive to direct his behavior to obtain a positive reac tion” from references (Sirgy, 1982, p. 286). In order to garner proper interpretation of his/her social performance, individuals attempt to control both environmental settings and “personal attire” (Grubb & Grathwhohl, 1967). In regards to consumer market behavior, po ssessions and “items” (e.g., products and brands) are used as sym bolic communication devices in the social interaction process to project a proper image and garner fa vorable reactions (Belk, 1988; Grubb & Grathwhohl, 1967; Katz, 1960). In other words, self-enhancement occurs through associations with goods that have de sirable social meaning that also bring favorable reactions from significant references.


15 Antecedents of Self-Brand Congruity The image that a person has of her/himsel f often influences the brands consumers purchase (Aaker, 1996; Bel k, 1988; Plummer, 2000; Sir gy, 1982; Zinkham & Hong, 1991).The simple presence of a particul ar brand (e.g., Gucci, FuBu, Birkenstock, Porsche, or Toyota) can serve to define a pers on with respect to others, particularly when social identity is involved (Belk, 1988). C onsumer behaviorists s uggest that by choosing brands with particular image associations (e.g., sophisticated or r ugged), individuals can communicate to others or themselves the type of person they are or want to be, in turn enhancing one’s own self-image and therefore maintain psychological equilibrium and health (Aaker, 1996; Graeff, 1996; Gr ubb & Grathwhohl, 1967; Keller, 1993; Underwood, Bond & Baer, 2001). Congruity theory applies self -image theory to explain the influence of self-brand congruity on consumer brand attitudes thr ough mediating influences of two motivations: self-esteem and self-consi stency (Sirgy, 1986). From a self-esteem perspective, consumers are motivated to purchase a positiv ely valued brand to maintain a positive self-image or to enhance self-image per ceptions by approaching an ideal image brand. Self-consistency, on the other hand, suggests that a consumer will be motivated to purchase a product with an image that is cons istent with actual self-image beliefs. According to congruity theorists, the result ant motivation state toward a given product or brand is the net effect of the motivationa l state arising from self-esteem and selfconsistency (Sirgy, 1986). Likewise, the “value expressi ve function” of human att itudes argues for the effect of product symbolism on the activation of self -image in consumption related situations (Katz, 1960). Value expressive attitudes func tion by giving “positive expression” to one’s


16 core values and to the type of person individuals’ thinks they are or want to be (Katz, 1960). The psychological reward an d satisfaction an individual gains from the expression of attitudes, which reflect one’s most important beliefs and self-image, results from establishing self-identity and c onfirming the perception of the type of person one believes to be. As stated by Katz (1960, p. 173), “value-e xpressive attitudes not only give clarity to the self-image but also mold the self-image closer to the heart’s desire.” Therefore, because self-image is of value to an indivi dual, behavior manifests towards its protection and enhancement, partially via consumed brands and products. Image Congruence Hypothesis Grubb and Grathwhohl (1967) we re first to formally hypothesize a relationship between self-brand congruity and brand att itudes. The authors argued that because products and brands are an important part of how individuals define themselves, consumers should have favorable attitudes towa rd brands perceived to be congruent with their self-image, and relatively less favorable attitudes towards bra nds perceived to be incongruent with their self-image. Figur e 2-1 illustrates the image congruence hypothesis: as the distance betw een brand and self-image d ecreases (i.e., high self-brand image congruity), positive brand ev aluation increases and vice-versa. Numerous studies on a variety of product categories and services have produced results supportive of the image congruence hypothesis. Moreover, decades of studies provide evidence that self-congru ity affects not only brand at titudes, but also variables such as brand satisfaction, preference, purch ase intentions, and c hoice (Birdwell, 1968; Dolich, 1969; Erickson, 1996; Grubb & Grat hwhohl, 1967; Graeff, 1996; Levy, 1959; Malhotra, 1981; Ross, 1971; Sirgy, 1982). In addition, self-congruity researchers have refined the measurement of key constructs (e.g., self-image), developed techniques for


17 calculating self-brand congruity i ndicators, and identified critic al variables that moderate the image congruence hypothesis. Figure 2-1. Illustration of th e image congruence hypothesis In regards to theoretical development, it is important to note a number of variables moderate the image congruence hypothesis. Self-brand congruity is not an important driver of consumer brand att itude in all product categories, nor does it drive consumption equally for all individuals (Aaker, 1999; Dolich, 1969; Gould, 1991; Graff, 1986; Hogg et al., 2000; Sirgy, 1982; Sirgy et al., 1997; Zinkham & Hong, 1991). In particular, the variables conspicuousness of brand use (i.e., pr ivate vs. publicly consumed), brands that that rely heavily on image, brand preference, and consumers’ “sel f-monitoring behavior” each are evidenced to moderate the imag e congruence hypothesis (Aaker, 1999; Gould, 1991; Zinkham & Hong, 1991). BRAND IMAGE NEGATIVE BRAND EVALUATION POSITIVE BRAND EVALUATION BRAND IMAGE BRAND IMAGE BRAND IMAGE SELF IMAGE SELF IMAGE SELF IMAGE SELF IMAGE


18 Theoretical Paradigms: Image Measurement and User Imagery Semantic differential scales, develo ped by Osgood, Suci and Tannenbaum (1957), originally intended for resear ching high-level cognitions a nd to study the meanings of concepts, are the dominant measures utilized to acquire consumer self-image and brand image perceptions. Use of semantic type scal es pervades all aspects of social science research, and literature shows that most congruity studies empl oy either standard semantic scales (i.e., based on opposite adj ective pairs where re spondents can indicate their degree of belief or feeli ng about objects of research intere st) or modified versions to accomplish particular research goals (e.g., Dolich, 1969; Grubb & Hupp, 1968). Figure 2-2. Self-brand congruity scales example Traditionally, self-congruity is assessed by measuring respondents’ perceptions of a product and/or brand image and their own se lf-image perceptions (i.e., actual and ideal) using the same set of image scales for both br and and self-perceptions. In one of the first and most widely replicated image congruen ce studies, Dolich (1969) applied a set of semantic differential scales to measure both self-image and brand image perceptions. In this early self-brand congruity study, the auth or simply measured “brand image” without any reference to the conceptual ization of the construct. Sema ntic scale ratings presented to respondents to describe real-self image (i .e., the person I am) and the image of brand Scale Person I am 1 impulsive ___:___:_ X _:___:___:___:___ deliberate 2 simple ___:___:___:___:_ X _:___:___ complex (to scale 22) Scale The brand of beer I most prefer is ____________ 1 impulsive _ X _:___:___:___:___:___:___ deliberate 2 simple ___:___:___:___:_ X _:___:___ complex (to scale 22)


19 of beer most preferred were similar to the example illustrated in figure 2-2 (Dolich, 1969, p. 81). Pioneer studies also set the groundwork in regards to conceptualizing the brand image construct with brand-user imagery pe rceptions. User imagery is a stereotyped perception of the “generalized” user of a particular brand, fo rmally defined as the set of human characteristics associated with the ty pical brand user (Sirgy, 1982). For example, Grubb & Hupp (1968), self-congruity pioneers evidenced self-brand congruity based on user imagery, hypothesizing that consumers of a specific brand would have self-images similar to self-images that were “attributed” to “other” consumers of the same brand that they owned, and different than owners of competing brands. Respondents rated semantic adjectives to the degree of similarity between perceptions of “stereotypes” of other brand owners and own self-image. The same scales measured brand-image, self-image, and “other” brand user image. Results suppor ted the hypothesis and demonstrated that consumers of a specific brand of automobile ha ve self-image perceptions similar to others that consume the same brand, while having diffe rent self-image per ceptions from owners of a competing brand. Birdwell (1968) was one of the first re searchers to measure the relationship between self-image congruity and purchase behavior. Although the study was simple, the results showed that respondent self-image was significantly more congruent with the brand of car owned, likewise employing sema ntic scales to measure user imagery perceptions. Landon (1974) was one of the first researchers to argue for the addition of the “ideal” self-image dimension. Landon showed that some subjects match product-user image with actual self-image, while others matched product-user im age with ideal self-


20 image. Subjects used Likert scales to rate brands according to how similar they believed the brand was to their own selves: “very st rongly like me” and “very strongly unlike me”. In addition, the study demonstrated a signifi cant correlation betw een self-image and purchase intentions and differen ces between respondent gender. In recent decades, researchers contin ue to operationalize brand image via respondent user image perceptions and em ploy semantic type scales to acquire respondents’ self and brand image perceptions. For example, Graeff (1996) examined whether promotional messages (i.e., advertisem ents) activate consumers’ self-image, and if activating their self-image increased the effect of self-image congruity on purchase intentions. The author developed a set of image dimensions (e.g., thrifty-indulgent, excitable-calm, youthful-mature, etc.) base d on a compilation of semantic dimensions from previous congruity studies. Respondent s rated the degree to which each image dimension associated with the “typical user” of different brands and described their own actual and ideal self-percep tions using the same image scales, on which the author employed a distance model to calcul ate congruity indicator scores. Hogg et al., (2000) examined the public a nd private contexts of self-image and differential impact of self-congruity on bra nd evaluation and choice. The authors used twelve “unipolar” semantic differential scal es for respondents to describe the user imagery of different brands. Respondents rated how well each semantic adjective described the “stereotypical consumer” of brands and thei r “situational” self-image perceptions. Results indicated that respondents prefer differe nt product types in different situations, and that brand selection correlates to brands that have si milar (i.e., congruent) user image perceptions.


21 Congruity Indicator Scores Congruity indicator scores depict the level of image congruence between selfimage perceptions and brand image perceptions . Researchers calculate congruity scores by comparing the data from two image measur es typically employing either a d-score or discrepancy ratio that indicates the degree of self-brand congruity (Aaker, 1999; Dolich, 1969; Graeff, 1996; Grubb & Grathwhohl, 1967; Martin, 1982; Sirgy, 1982, 1997; Sutherland et al., 2004). With the d-score ap proach, congruity vari es by distance: the larger the d-score (i.e., distance) between brand and the respondent’s image, the less congruity, and the lower the d-score between the brand and the respondent’s image, the more congruity. Figure 2-3. Absolute simple-difference model To employ the discrepancy ratio, one cal culates a discrepancy score for each image dimension measured, and then sums th e discrepancy scores across all dimensions (Sirgy, 1997). A variety of math ematical indexes are availabl e to estimate discrepancies (e.g., see Sirgy et al., 1997 for a review of fo rmulas applied to congr uity calculation). As an example, figure 2-3 shows the most often us ed formula, the absolute simple-difference model. The discussed method is the traditio nal and most widely used formula to determine and study self-brand congruity. n = ( Pij Sij) i=1 where Sij = self-image (i) of individual (j) Pij = brand image perception (i) of individual (j)


22 “Global” Self-Congruity Measure Sirgy et al. (1997) reviewed the “tra ditional methods” discussed above for measuring self-congruity and examined the pred ictive validity of a “global” congruity measure in a series of six comprehensive st udies. The authors ar gued that traditional techniques have three methodological flaws. On e, discrepancy and distance scores tend to be unreliable and have systematic correlations with their components, restricting variance in statistical analyses. Two, th e use of pre-determined imag e characteristics may or may not be relevant to brands evaluated. Three, the use of the compensatory decision rule implies that respondents experience self-congrui ty with more than one image dimension and then integrate the information across dimensions. The argument against the compensatory decision rule is that consumers process value expressive brand associations “holistically.” Hence, the authors claim that the traditional methods do not “capture” the actual congruity experience. To address these problems, the researchers introduced the “global” measurement approach, grounded in the notion that self-image congruity is a “holistic, gestalt-like perception” rather than a “piecemeal pro cess” (Sirgy et al., 1997 p. 230). The method bypasses the use of discrepancy scores and does not use predetermined images, allowing respondents to generate free association im age descriptors salient at the time of measurement. Respondents are instructed to think about a particular brand and the person who typically uses the brand a nd then describe the brand, once they do this, respondents indicate their agreement or di sagreement to how consistent that person is with how they seem themselves (i.e., actual self-image) and how they would like to see themselves (i.e., ideal self-image).


23 Participants then mark their responses on a Likert scale in regards to self-image statements and repeated for all self-image dime nsions of interest. The predictive validity of this measure held up across consumer beha vior indicators such as brand preference and brand attitude, using different products and consumer segments. Additionally, this technique proved predictive in all regressi ons, producing substantially higher beta weights in all cases compared to th e traditional approach (Sirgy, 1997). Theoretical Summary Self-congruity theory explai ns why consumers like and di slike particular brands and products as a function of self-expressive needs, a common denominator in human functioning that also influences consumption patterns. The dominant paradigm of selfbrand congruity research is to measure self -brand congruity with either traditional semantic differential type image scales or the recently develope d “global” congruity measure. The central contribution of self-congr uity theory is an understanding of brand attitudes in regards to the structure of the relationship between brand image and the consumer. Since, self-brand congruity takes plac e at the symbolic level of the brand, than a more encompassing conceptualization of the brand image construct (the primary symbolic component of a brand) will st rengthen the model both practically and theoretically. Brand Image Brand image has been a central c oncept for communication scholars and marketers at least since the 1950’s (Patte rson, 1999). In today’s highly competitive information landscape, marketers attempt to establish images for their products, symbolized by the brand in order to garn er a position unique from the competitive choices. An established brand image position is critical in order to protect against


24 competitive offering and enhance long-term market performance (Shocker & Srinivasan, 1979). A brand image is established in the consumer’s mind via multiple sources of information regarding the brand and/or pr oduct. In particular, various forms of communication both controlled and uncont rolled by marketers (e .g., word of mouth, advertising, packaging) and even via the obs ervation of the type of people who use a particular brand, such as Izod shirts for “ yuppies” or Oakley sunglasses for “surfers” or “snowboarders.” In regards to formal definitions, brand literature provides two related conceptualizations of the brand image constr uct, both discussed pr esently to provide a comprehensive overview. First, researchers s uggest that via the us e and consumption of brands, consumers form “subjec tive perceptions” of how diffe rent brands perform across functional and non-functional cr iteria important for evaluati on and choice decisions (Patterson, 1999). Consumers organize subjective perceptions into a “succinct picture of the brand” that (1) are important determin ants of consumption behavior and (2) are “wholly resident” in the consumer’s thought process called a brand image (Patterson, 1999, p. 412). Essentially, brand image is a su bjective perception of functional and nonfunctional information regarding the product or service (Patterson, 1999; Sutherland et al., 2004). A slightly different concep tualization, suggest that brand image is an overall perception, a mental repres entation, about a brand as indicated by informational associations stored in consumer memory (Aaker, 1996; Keller, 1993; Schultz & Barnes, 1999). Informational brand associations, link to a brand symbol in memory are what consumers think about when characterizing a brand’s image (Batra et al., 1996; Keller,


25 1993). Brand associations that consumers beli eve will satisfy their needs, wants, and desires better than competitive choices; ga rner favorable respons es (Keller, 1993). In other words, brand associations are what consumers think about when characterizing a brand’s image, which influences how one f eels towards a brand. Brand image perceptions develop from two general types of informati onal associations: attributes and benefits. Attributes are descriptive features that represent what a consumer thinks that a product has or does and ascribes to a brand. Attributes are both “product related” (i.e., features needed for performing the product function) and “non-product related” (e.g., price, packaging, brand pe rsonality) (Aaker, 1996; Sc hultz & Barnes, 1999). In comparison, benefits are the “personal valu e” (i.e., what the consumer perceives the product attribute will do for them) that c onsumers ascribe to a brand (Aaker, 1996; Keller, 1993). The benefits a consumer ascribes to a brand image, relate to three underlying consumption motivations (i.e., needs): functiona l needs (i.e., desire for problem removal), experiential/sensory needs (i.e., what it feels like to use a product), and symbolic needs (i.e., social approval and personal expre ssion) (Park, Jaworski & MacInnis, 1986). Functional and experiential brand benefits ar e associated with product related attributes (physical product characteristics and features ) (Keller, 1993). For example, characterized by consumers, the brand Crest is most often salient on sensory brand image associations such as “fresh” or “clean.” In contrast, sym bolic benefits relate to non-product attributes, such as user imagery (Solomon, 1983). A case in point, consumers of ten characterize the brand Nike with symbolic brand image associ ations such as “youthful” and “sporty.” A


26 major symbolic brand associati on, that has garnered the attent ion of researchers, is the concept brand personality (Aaker, 1997). Brand Personality Throughout history, advertiser s have attempted to humanize and personify brands and products using animism tactics, the pr ocesses by which inanimate objects are imbued with characteristics that make them somehow person like (Aaker, 1997). Grubb & Grathwhohl (1967) suggested that brands, products, and stores have “psychic values” to certain market segments that relate to so me consumer market behaviors. David Ogilvy (1983, p. 14) proposed that image means person ality. He stated, “products, like people, have personalities, and they can make or break them in the market place.” The personified brand has captivated the attention of advertising researchers, particularly in the decades of the 1980’s and 1990’s. In the 1980’s, advertising a nd brand image researchers transposed the unit of observation from consumer personality to brands by ascribing to them a personality (Ogilvy, 1983; Plummer, 1985). Over time, re search has supported the contention that brands develop unique personali ties that can serve as units of observation and analysis (Aaker, 1996; Aaker, 1997, 1999; Ogilvy, 1983; Plummer, 2000; Sutherland et al., 2004). In essence, brand personality is the act of applying human ch aracteristics (i.e., personality traits) to a brand, inducing cons umers to think of a brand as if it had pers on like qualities (Aaker, 1997; Keller, 1998). Associating human personality characteris tics with a brand is possible because people anthropomorphize, that is, transfer human charac teristics (i.e., personhood) to inanimate objects on a regular basis (Bower, 1999; Boyer, 1996). Typical examples are animating a “pet rock” or when one referen ces an object, such as a motor boat by saying,


27 “she is a beauty.” Not only do individuals view the anthropopathic, inanimate object as she/he would another person, but also treats the object as another person (Boyer, 1996). Likewise, consumers have a natural tendency to imbue brands with human characteristics (e.g. rugged, sophisticated, inte llectual), thinking of them as if they are more than inanimate objects. For example, respondents char acterized the brand personality of Oil of Olay as “upscale” and “aspira tional,” while Absolut vodka’s brand personality has been characterized as “cool,” “h ip,” and “contemporary” (Aaker, 1997; Plummer, 2000). Formally defined, brand personality is “the se t of human characteristics associated with a brand” (Aaker, 1997, p. 347), considered a ma jor symbolic brand image association. Literature evidences that companies that employ brand personality as a part of an overall positioning strategy, when properly a nd consistently comm unicated, can affect consumer perceptions in far more e nduring ways than other marketing and communication strategies (Burke , 1994). Brand personality pr ovides an enduring point of differentiation, particularly in categories wh ere products have reached functional parity and/or when symbolic consumption motivations market behavior. This differentiate (1) facilitates consumer choice by simplifying the decision process, (2) increases awareness and attachment (i.e., builds loyalty), and (3 ) enhances the favorability a brand’s image (Phau & Lau, 2001; Sutherland et al., 2004). Brand personality source variables Plummer (2000, p. 80) described “two faces” of brand personality that are critical to “better grasp the totality and power of this useful strategic concept.” The first face is “input,” that is, what marketers want cons umers to think and feel about a brand’s personality image. In other words, the way in which a company presents its brand to consumers through all marketing and communica tion efforts. The second face of brand


28 personality is “out-take,” which is what c onsumers actually do think and feel about a brand’s personality image. Since consumers’ view of a brand are influenced by many factors such as market noise, personal e xperience, both direct and indirect source variables, then the out take likely will differ from the or iginal input (Plummer, 2000). A brand’s human personality traits result fr om any direct or indirect contact an individual has with the brand (Aaker, 1997; Plummer, 1985). Di rect source brand personality traits originate from any indivi dual associated with the brand (e.g., endorsers, spokespersons, company CEO, family members, and typical user), and transfer to the overall brand personality perception (Aak er, 1997). In comparison, indirect brand personality traits originate from such inform ational sources as product attributes, product category, brand name and symbol, advertis ing approach, price, and demographic characteristics (e.g., gender and social class). An example of a direct brand personal ity source, spokespersons and endorsers enable the consumer to feel as if the “spi rit” of the endorser is available via brand consumption (McCraken, 1986). In a similar ma nner, brands are personified by enduring associations with family members such as th e brand of beer dad used to drink, the brand of soap grandma always had in the bathroom, or “hand-me-down” brands can all be strongly associated with anothe r individual from the past and transfer to a brand’s image (Solomon, 1983). Overall, human characteristic s associated with a brand are drawn from many possible sources, resulting in a global perc eption of a brand as if it has an enduring human like personality. Brand personality scales and framework Given that brand personalities are exte nsions of human personalities, pioneer researchers adapted widely accepted human pers onality theories and related measures in


29 order to characterize brands, with little avail. A central prob lem with early brand personality research was the use of scales and m easures with little regard to their validity and reliability. Kassarjian (1971) often cited in marketing literature, su ggested in order to achieve unequivocal results ac ross studies, researchers must develop their own scales and measures to reliably and validly measure the particular concepts and phenomenon under investigation. To this end, Aa ker (1997) suggested brand personality operates differently than human personality, advancing a framewor k of the brand persona lity construct and a set of indicators for measur ing construct dimensions. Aaker (1997) advanced a model of the br and personality cons truct and developed a set of “trait” scales for measuring distinct dimensions of brand personality. The Brand Personality Scales (BPS) are grounded in a framework that psychologist recently have advocated as a comprehensive classification of human personality based on five core human personality traits [i.e., Extrov ersion, Agreeableness, Conscientiousness , Emotional Stability, and Culture (or Intell ect)] (McCrae & Cost a, 1985, 1987). In the process of developing the Bra nd Personality Scale (BPS), th e author util ized over 1000 respondents, 60 brands, and included an orig inal 114 personality traits (Aaker, 1997). Results demonstrated a robust, five-factor so lution emerged for the classification of a brand’s personality: Sincerity, Excitement, Competence, Sophisti cation, and Ruggedness. In addition, results evidenced that the BPS fr amework and scales are reliable, valid, and generalizable across brands and consumer segments. Each BPS personality dimension is subdivided into “facets” that provide descriptors of each dimension and indicat ors for operationalization (Aaker, 1997). In total, the standard BPS measures five pe rsonality dimensions and fifteen facets,


30 illustrated in Figure 2-4. To date, the brand pe rsonality dimensions assessed via the BPS are the only published and most widely employed brand personality measures, shown to be reliable and generalizable across different brands and product categories (Aaker, 1997, 1999). Figure 2-4. Brand personality framework The “Sincerity” dimension captures the idea of a warm and accepting brand personality and represented by four facets: down-to-earth, honest, wholesome, and cheerful. “Excitement,” similar to the human personality dimension extroversion, characterizes a sociable, energetic, and activ e brand personality, also represented by four facets: daring, spirited, imaginative, a nd up to date. The “Competence” dimension characterizes a brand personality that is re sponsible, dependable, and secure, represented by three facets: reliable, intelligent, and successful. According to Aaker (1997) the brand pers onality dimensions “Sophistication” and “Ruggedness” differ from any known major di mension of human personality and the authors suggest that these two dimensions may relate to or tap into brand personality dimensions that individuals’ desire but do not necessarily have. The “Sophistication” dimension is represented by two facets, uppe r class and charming, and two facets also represent “Ruggedness”: outdoorsy and tough. BRAND PERSONALITY Sincerity Down-to-Earth Honest Wholesome Cheerful Sophistication Upper Class Charming Competence Reliable Intelligent Successful Excitement Daring Spirited Imaginative Up-to-date Ruggedness Outdoorsy Tough


31 Symbolic function of brand personality At the core of a brand’s image percepti on are its human charact eristics, key points of brand differentiation and paramount imag e associations (Ogilvy, 1963; Parente, 1996; Phau & Lau, 2001). Derived from multiple informational sources, the construct brand personality is a salient facet of a brand’s overall image th at serves a symbolic, selfexpressive function for the consumer (K atz, 1960; Keller; 1993; Plummer, 2000). Suggested by brand strategist Plummer (2000, p. 81), brand personality plays a critical role in the “for me” choice, or “I see myse lf in that brand” choice. Moreover, brand personality, may be an important driver of brand evaluation when the “social signaling” value of a brand is greater, when the consumer is in a socially important situation, and/or in particular product categories (Batra, Myers & Aaker, 1996). Since brand personality serves a symbo lic and self-expressive purpose for the consumer, then consumers should be attrac ted to brands that project congruent personalities. This is because symbolic human characteristics compose a brand personality, which allows the processes of image congruence to function. Image congruence relies on symbolic interactionism: the psychological process that allows an individual to symbolically associate one’s se lf-image with the symbolic image of a brand (Dolich, 1969; Sirgy, 1982). As a result, bra nd personality researchers have employed the congruity theory framework. For example, Aaker (1999) examined consumers’ self-concepts and selfmonitoring behavior by employing the brand pe rsonality traits to study brand and selfimage suggest in text, “the crux of self c ongruity is that consumers prefer brands associated with a set of personality traits congruent with their own.” Employing the selfcongruity framework, the author investigat ed the hypothesis that respondents would


32 prefer brands with personality characterist ics congruent, versus incongruent with their self-concept perceptions. In another example, Wee (2004, p. 320) in an examination of the brand personality construc t states in text, “there should be congruence between the brand personality and the consumers self-concept because consumers aim to express themselves in brand choices and prefer pr oducts that match thei r self-concept since purchase offers a vehicle for self-expression.” Common to the aforementioned studies, congr uity theory provides the theoretical foundations of the research. However, self-congr uity theory research does not report any studies that have empirically incorporated this construct into the model. This is not to discount the logic in employing self-congruity principles to study brand personality. As argued, it is a symbolic brand image associati on that provides self-e xpressive value to the consumer. However, the history of congruity theory development in regards to self-brand congruity research focused on a single lim ited conceptualization of the brand image construct. Hence, the incorporation of the brand personality construct into the selfcongruity model by making comparisons to prev ious findings is firs t necessary. By doing so will, foster a better understanding of the structure of brand attitudes when selfexpressive needs motivate consumption behavior. Brand Attitude Derived from social psychology, the concept “attitude” is central to the fields of advertising, brand, and related research fields (Batra et al., 1996). Attitudes are often researched in these disciplines because they are (1) insightful in predicting consumer behavior (Mitchell & Olson, 1981) and (2) social psychology offe rs long used theories to guide such research. As an example, co mmunication researchers study attitudes in a variety of research domains, such as copy te sting, concept testing, brand evaluations, and


33 brand tracking studies used to indicate poten tial market behavior (Hastak & Olson, 1989; Kardes, 1999; Keller & Aaker, 1992). Formally defined, an attitude is “a learned predisposition to respond in a consistently favorable or unfavorable manner with respect to a give n object” (Fishbein & Ajzen, 1975 p. 6). Individuals can form attitudes toward any object, symbol, or aspect of his/her world (Katz, 1960): in this study a brand. Attitu des develop as a result of information that includes not onl y facts and figures, but also a ny stimulus or cue that is “psychologically meaningful” or influential (Kardes, 1999). Most important, attitudes are enduring psychological states that tend to “energize” behavior (Spears & Singh, 2004). Fishbein and Ajzen’s (1975) multiattribute m odel of attitudes is widely applied in communication research. This model is base d on the relationship between affect (i.e., evaluation), cognition (i.e., belief s), conation (i.e., behavior in tention), and behavior (i.e., observable overt acts). Intentions and behavi ors are typically studied in their own right. Rather, it is evaluative predis positions to act, primarily a function of one’s cognitive belief structure that matters most in a ttitude research (Fishbein & Ajzen, 1975). Therefore, the present discussion focuses on the affective and cognitive attitudinal constructs. Affect, the most essential part of the attitude concept, is the positive or negative evaluations (e.g., like vs. dislike, favorable vs. unfavorable) of an object along bipolar dimensions. One should construct the concep t attitude with a “bipolar affective dimension” measure because affect is an “evaluative judgment” that consists of two components: direction (positive or negative) and extremity (e.g., slightly positive, somewhat positive, fairly positive, very pos itive, extremely positive) (Kardes, 1999;


34 Osgood, Suci, & Tannenbaum, 1957). In comparison, beliefs, the cognitive components of attitudes, are the “informational base” that ultimately determines affect and behavior, and considered the “fundamental” building bl ocks of attitudes (F ishbein & Ajzen, 1975; Kardes, 1999). Most important, all attitudes consist of beli efs, though not all beliefs are attitudes (Katz, 1960). Therefore, it is in the cognitive belief structure (i.e., the informational base) that one finds reason for attitudinal evaluations. The present study focuses on consumers’ ev aluative disposition towards brands, referred to as brand attitude (Ab). Brand attitude, formally de fined in the present study, is a learned predisposition to respond in a cons istently favorable or unfavorable manner towards a brand name, symbol, logo, or other brand identifier (Keller, 1993). Because humans have limited mental capacity a nd usually focus on salient information, ultimately, a person’s attitude towards an objec t is a function of “salient beliefs” about that object (Fishbein & Ajzen, 1975; Keller, 1993). Figure 2-5. Cognitive and affective brand image factors Both favorable and unfavorable brand eval uations result from consumer beliefs about associated “attributes” and “benefits” that are salient in a brand’s perceived image (Aaker, 1996; Keller, 1993; Wilkie, 1994). Figur e 2-5 illustrates this perspective on brand Ab Brand Evaluation Brand Image N E G ATI V E P OS ITI V E ATTRIB U TE S BE N EFIT S


35 attitudes; brand image is the cognitive bra nd attitudinal factor that affects brand evaluations to form one’s pr edisposition towards a brand (Ab). Summary of Literature Review Brand researchers interested in studying th e relationship between consumers’ selfexpressive needs and the st ructure of brand attitudes have employed self-congruity theory. Self-congruity theor y, grounded in principals derived from social psychology, explains that the comparison consumers ma ke between a brand’s projected image and their own self-image partially drive consumer behavior by infl uencing brand attitudes. At times, consumers like brands that are perceived to be similar to the self-image they desire to project in social situati ons, either an ideal self or and actual self. The dominant paradigm in self-brand congruity research, spanning four decades of studies (1960-2000), is constructing brand image as the typical br and user and comparing this to consumers’ self-image motivations as a basi s for predicting brand attitudes. User imagery, that is, the view of a brand’ s typical user, is only one aspect of a brand’s composite image, and may not always capture the actual congruity experience. Brand personality, a major facet of a brand image, is similar to user imagery in that both constructs represent human characteristics a ssociated with brand. Wh en characterizing a brand’s user imagery, one is only describi ng a hypothetical person, not necessarily the brand image. However, brand personality is a direct perception of a brand’s image, at least the symbolic components, and encompasses multiple source inputs such as product endorsers, celebrity spokesperson, animated characters, etc. Because it is a symbolic, self-expressi ve image association, brand researchers have intuitively replaced user imagery per ceptions with brand personality as the key component of brand’s image in the self-brand congruity model. Even so, literature does


36 not indicate that researchers have compared brand personality and user imagery within the same congruity model. The argument made here is that the constructs user imagery and brand personality are conceptually di fferent, necessitating empirical comparison, fostering a better understanding of the role br and image plays in forming brand attitudes when self-expressive needs motivate consumer behavior. Research Objectives and Questions The ultimate goal of this study was to compar e conceptually distinct facets of the construct brand image within the self-congruity model framework, as a basis for understanding the cognitive struct ure of self-expressive brand attitudes. Specifically, to compare the constructs brand personality a nd user imagery to determine if they are empirically distinct when used independently in the operationalization of the self-brand congruity phenomenon. In other words, to de termine if the relationship between selfbrand congruity and brand attitudes differs depending on which brand image construct consumers compare to their self-image perceptions. In doing so, this research will foster a better understanding of the brand image construc t and therefore a bette r understanding of the cognitive structure of brand attitude s in the self-brand congruity model. Given that these two constructs represen ts conceptually distinct brand image facets, when made salient to respondents a nd compared to their self-image perceptions different congruity experiences should result for the same br and. In order to facilitate research and discussion, the author proposed two concepts that represent two types of self-brand congruity. First is the perceptual comparison ex perienced between consumer self-image and brand user imagery percepti ons, termed “user imagery congruity” (UIC). Second, the congruence experienced between self-image perceptions and a brand’s personality perceptions, termed “b rand personality congruity” (BPC).


37 Moreover, a primary challenge in self-bra nd congruity research is to determine which consumer self-image perceptio n (e.g., actual vs. ideal) is most congruent/incongruent with a brand imag e and predictive of brand evaluation. Researchers accomplish this task by calculati ng and comparing congruity indicator scores that represent different levels of congru ence between a brand’s image and consumer’s self-image and then statistically compar ing the self-brand congruity indicators to attitudinal measures. Table 2-1. Independent variable s: self-brand congruity indicators Congruity Indicator Self-Image Dimension Brand-Image Dimension Actual-UIC Actual self-image User imagery Ideal-UIC Ideal self-image User imagery Actual-BPC Actual self-image Brand personality Ideal-BPC Ideal self-image Brand personality In regards to the present study, combining UIC and BPC with dimensions of selfimage results in four distinct self-image congruity indicators, re presented in Table 2.1 that served as independent variables. “Act ual” and “Ideal” UIC indicators represent the congruence between the two dimensions of self -image and the image of the typical brand user as viewed by the respondents. Likewi se, “Actual” and “Ideal” BPC indicators represent the congruence betw een self-image dimensions and consumers’ view of a brand’s personality. The nature of this study was explorator y and descriptive. Because no previous studies have compared BPC and UIC indicato rs, there was no basis to derive hypotheses. Rather, towards discovery, the study addresse d research questions intended to achieve two primary objectives derived to accomplis h the overall empirical goal of the study. Doing so will provide a basis for future re searchers to both deduce testable hypotheses that may support or falsify the present fi ndings and also compare the influence of


38 moderating variables, already shown to influe nce the relationship between UIC and brand attitudes, to the relationship be tween BPC and brand attitudes. Research Objective One The first research objective was to comp are UIC and BPC as a basis for predicting brand attitudes for a set of publicly consumed brands relevant to ta rget respondents and to determine if the results are similar or diffe rent across brands with unique brand images. Since user imagery is only one direct source variable that may or may not influence brand personality perceptions, then brand pers onality will at times differ (Aaker, 1996; Plummer, 2000). The ultimate goal was not to compare the actual constructs user imagery and brand personality, but rather the self-bra nd congruity that results when these two constructs are independently us ed in the calculation (i.e., ope rationalization) of self-brand congruity. In addition, it is important to inco rporate different self-image dimensions, that is, actual and ideal self-ima ge, and make comparisons betw een UIC and BPC. Therefore, the following two research questions addressed this overall objective: Research question 1. What is the empirical relati onship between “actual” brand personality-self congruity (Actual-BPC) and “actual” brand user image-self congruity (Actual-UIC) when employe d in the same model as distinct independent variables used to predict brand attitudes for publicly consumed brands with unique images? Research question 2. What is the empirical rela tionship between “ideal” brand personality-self congruity (Ideal-BPC) and “ideal” brand user image-self congruity (Ideal-UIC) when employed in the same model as di stinct independent variables used predict brand attitudes fo r publicly consumed brands with unique images? Research Objective Two As discussed, a central goal of self-congruity research is to determine which selfimage dimensions is most congruent with brand image and strongly associated with


39 positive brand attitudes. The second research objective was to determine which of the four congruity indicator scores adds the most explanatory power in a series of attitudinal prediction equations for different brands. In other words, when applied in the same model, the objective was to determine which of the four indicators from table 2.1 account for the most variance when used to model fa vorable brand attitudes. The purpose was to isolate the self-brand congruity relationship(s) that garner the most favorable reactions. Note this possibly could involve a single indicator as the best predictor of brand attitude, or any combination of the four indicator type s. Hence, the following research question: Research question 3. Which of the self-brand imag e congruity indicators (i.e., Actual-BPC, Actual-UIC, Ideal-BPC, and Ideal-UIC), or combination of, accounts for the most variance when all four variables are employed in the same model as a basis for predicting brand attitudes for a set of publicly cons umed brands with unique images?


40 CHAPTER 3 METHODOLOGY This study utilized a survey to assess the relationship between user-image selfcongruity (UIC), brand personality self-congr uity (BPC) on brand attitude, impression, desirability, and purchase inten tion each related to the attitu de construct. Prior to the primary survey, preliminary research genera ted an extensive list of publicly consumed brands meaningful to respondents, of wh ich the researcher selected four (i.e., Abercrombie & Fitch, Birkenstock, Banana Re public, and Nike) to serve as units of analysis. Questionnaire devel opment involved two pre-tests on sub samples of the target population for the primary study. After questi onnaire refinement, two hundred seventy two (n=272) undergraduate college students at a large southern university, participated in the primary survey. Brand Selection Preliminary research identified a set of socially consumed brands (1) judged relevant and meaningful to a sub sample of target respondents, (2) that had unique described images, and (3) were readily ava ilable and consumable by either female or male respondents. On two occasions, brand names were generated from undergraduate college students (n=87) drawn from the sa me population used to sample respondents in the primary survey. Given verbal instructions , participants listed and characterized the image of three brands that th ey currently own and use prim arily in a public or social situation. As instructed, partic ipants used two or more adje ctives to describe each brand


41 based on their free associations. Frequency of free association was determined to be a good preliminary assessment of each brand’s images (Malhotra, 1981). Table 3-1. Top twenty-five publicly consumed brands Brand Name Number of Mentions Percent of Respondents Most frequent descriptors Coca-Cola 58 66% Cool/Fun Nike 49 56% Sporty/Competitive Pepsi 45 52% Young/Active Taco Bell 45 52% Friendly/Different Budweiser 44 51% Traditional/Relaxing Banana Republic 31 43% Preppy/Classy Abercrombie & Fitch 28 32% Hip/Trendy (young) Dr. Pepper 28 32% Unique/Different Mountain Dew 26 30% Upbeat/Energetic Old Navy 26 30% Casual/Comfortable Gap 25 29% Classic/Familiar Tommy Hilfiger 25 29% Preppy/Sporty Adidas 22 25% Athletic/Hip Starbucks 20 23% Warm/Sociable Levi’s 20 23% Average/Classic American Eagle 20 23% Young/Easy going Gatorade 20 23% Trustworthy/Confident Miller Lite 16 18% Cheap/Average Volkswagen 15 17% Fun/Safe Ralph Lauren 14 16% Sophisticated/Preppy Doritos 12 14% Quality/Good Birkenstock 12 14% Earthy/Laid back Evian 11 13% Clean/Fresh Victoria’s Secret 11 13% Sexy/Mature Seven-Up 10 12% Old/Crisp This procedure generated a list of over two hundred brands, many with only a single mention. Table 3.1 displays the top twenty-five brands ba sed on frequency of mention by respondents and the two most common adjective descriptors used by respondents to characterize the im age of the brand. From this list, the researcher selected four brands to serve as the units of analysis in the primary survey. Four brands are sufficient to achieve the research objectives and not overly cumbersome to respondents,


42 allowing enough time to complete the questionna ire. Likewise, most congruity studies examine between two and four brands in a study. The criteria utilized to select the brands from those displayed table 3.1 were, freque ncy of mention, relative homogeneity of the product categories, and distinct iveness of image based on par ticipants’ free association descriptions. Figure 3-1. Selected brands and image descriptors Selected from this list were two clot hing apparel brands (Banana Republic and Abercrombie & Fitch) and two footwear bra nds (Nike and Birkenstock), illustrated in figure 3.1. The first brand, Nike mentioned by about 56 percent of respondents (n=49), and most frequently characterized as “co mpetitive” or “athletic.” Second, the brand Banana Republic was mentioned by about 43 percent of respondents (n=31) most frequently characterized as “preppy” or “cl assy.” Third, the brand Abercrombie & Fitch was mentioned by about 32 percent of respondent (n=28) most frequently characterized as “hip,” “trendy,” and “young.” The fourth brand, Birkenstock was mentioned by about 14 percent of respondents (n=12) and most fre quently characterized as “earthy” or “laid back.” Even though not mentioned as frequently as the other brands selected, the brand C OMPETITIVE A THLETIC NIKE L AID B ACK E ARTHY BIRKENSTOCK C LASSY P REPPY BANANA REPUBLIC H IP T RENDY ABERCROMBIE & FITCH


43 Birkenstock’s free association image is most distinct from the other three brands, providing a point of descriptive comparis on within the clothing apparel category. Questionnaire Development The research instrument consisted of four sections. Section one of the questionnaire collected basic demographic data including gender, year in college, ethnicity, and age. Also in this section, the instrument collected familiarity ratings for each brand to allow for the exclusion of s ubjects not familiar with any of the brands analyzed. Section two presented brand att itude measures including brand impression, brand desirability, and purchase intenti on scales. Section three employed the user imagery based “global” self-congruity scales (Sirgy, 1997) to assess “actual” and “ideal” UIC. The fourth section employed tradit ional self-brand congruity techniques to operationalize BPC. This was accomplished ut ilizing the BPS scales (Aaker, 1997), to measure both brand and self-image (i.e., act ual and ideal). These image measurements served in the calculation of BPC indicator scores for each brand. Finally, based on the pilot test, it was decided not to counterbalan ce the order of presentation of user imagery and brand personality scales to avoid the brand personality descriptors biasing the user imagery descriptions. An overview of the ope rationalization of each variable follows. Brand Attitude Operationalization The second section of the questionnaire co ntained the brand attitude measures. The researcher employed multiple brand attitude measures; if results are similar across measures, one can have more certainty in the findings (Fishbein & Ajzen, 1975). Researchers typically operationalize the c onstruct attitude with “bipolar affective dimension” measures because attitudinal affect is an “evaluative judgment” that consists of two components: direction (i.e., positive or negative) and extremity (e.g., slightly


44 positive to extremely positive) (Kardes, 1999; Osgood, Suci & Tannenbaum, 1957). Towards this end, a three-item, seven-point attitude scale (fa vorable/unfavorable, good/bad, and likeable/unlikable), commonly used to measure brand attitude (Ab), was employed (Mackenzie, Lutz & Belch, 1989). Bra nd attitude was operationalized as an index of these three semantic di fferential attitudinal scales. Three related brand attitudinal construc ts, brand desirabil ity, brand impression, and purchase intentions, were included to enhance measurement rigor (Kardes, 1999). Desirability and impression constructs were operationalized using five -point Likert-scale items (1 = strongly agree to 5 = strongly di sagree) in response to evaluative questions. For brand desirability, respondents indicated th eir reactions to the statement, “Overall, the brand [X] is very desirable.” For impr ession, respondents indica ted their reaction to the statement, “My overall impression of bra nd [X] is favorable.” Purchase intentions also was operationalized with a five-point Likert scale (1 = I definitely would buy it to 5 = I would definitely not buy it) in response to the question, “Think about the brand [X]. What is the likelihood that you would purchase th e brand if it were available locally and priced comparably to the leading brands?” User Image Self-Congruity (UIC) Operationalization The third section of the que stionnaire contained the “g lobal” congruity scales (Sirgy, 1997) employed to catalog brand user im agery descriptors and measure self-brand congruity (i.e., UIC) based on these descriptor s. For each of the four brands analyzed, respondents were instructed to first describe the brand’s “typical user” using any personal adjective descriptors that came to mind and th en characterize congrue nce with their selfimage; repeating the follow ing procedure for each brand.


45 Take a moment to think about [Brand x]. Think about the kind of person who typically uses [Brand x]. Imag ine this person in your mind and then describe this person using one or more personal adjectiv es such as stylish, classy, masculine, sexy, old, athletic, or whatever personal adjective you can use to describe the typical user of [Brand x]. Once you have done this, indicate your agreement or disagreement to the following statement: [B rand x] is consistent with how I see myself” (actual self congruity ); [Brand x] is consistent with how I like to see myself” (ideal self congruity). After writing one or two descriptive adjec tives, participants then marked their responses on five point Likert scales (e.g., 1 = str ongly agree to 5 = str ongly disagree) in regards to the self-image statements above, repeated for both actual and ideal self-image. Note that this self-brand congruity measure does not directly asse ss ones self-image, but rather attempts to capture the congruity expe rience directly. Hence, this technique does not require calculating distance or difference scores. Rather, UIC is operationalized as the average on each scale (i.e., idea l and actual) for each brand; the lower the average, the more congruence experienced between brand and self. Brand Personality Self-Congrui ty (BPC) Operationalization Presently, researchers have not develope d a global measure similar to the user image congruity scales that incorporates the brand persona lity construct. Rather, BPC was operationalized by employi ng Aaker’s (1997) fifteen item brand personality scales (BPS) and traditional self-brand congruity techniques by measuring brand image and respondent self-image perceptions with the sa me set of image scales. Self and brand BPS image scores provided the base for calculati ng BPC self-brand congruity indicator scores for each brand. The BPS measures five facets of brand personality, employing indicators of each dimension displayed in table 3.2. There are two primary reasons to empl oy BPS scales in self-brand congruity research. First, the BPS scales were derived in research employing techniques outline by


46 Malhotra (1981), who developed a process of scale development for measuring self, person, and product concepts, which is commonl y applied in congrui ty research. Second, the BPS scales have been used in self-bra nd congruity research in order to measure consumers' self-image, in which respondents rated themselves on the BPS traits, used to determine brand image congruence (Aaker, 1999). Table 3-2. Brand Personality Scal e (BPS) dimensions and indicators Brand Personality Facet Indicators Down-to-earth Honest Wholesome Sincerity Cheerful Daring Spirited Imaginative Excitement Up-to-date Reliable Intelligent Competence Successful Upper Class Sophistication Charming Outdoorsy Ruggedness Tough Respondents described each brands’ personal ity on five-point Li kert scales (1 = not at all descriptive to 5 = extremely descriptive) by i ndicating the number from the scale that best described th e brand as a person. The higher the number, the more that personality trait described the brand as a person. After characterizing each brand’s personality, the respondents used the same scales to charact erize both their own ‘actual’ and “ideal” self-image. Appe ndix A contains the research instrument, including the instructions given to responde nts for both brand personality and self-image measures. BPC self-congruity was operationalized as the difference, or distance, between respondent’s self-image scores and each brand personality score. BPC congruity


47 indicators were calculated with the distan ce-squared model, depicted in figure 3-2, for each BPS dimension: Figure 3-2. Difference squared model With this approach, congruity varies by di stance: the larger the d-score (distance) between brand and the respondent’s personal ity, the less congruity, and the lower the dscore between the brand and the responde nt’s personality, the more congruity. Pilot Tests Two separate pilot tests assessed how well the questionnaire gathered the intended information and provided insights used to refine the instrument. The first pilot test was an “active” pre-test, the researcher administered the questionnaire to a small group of target respondents (n=30) from the population used to draw the primary survey sample. Insights from an active pilot test result from discussion be tween researcher and respondents after questionnaire administration, rather than collected data (Davis, 1997). After completing the questionn aire, participants provided feedback in a question and answer group session in regard s to the instrument instruct ions, flow, and other issues derived in discussion. In comparison, the second pilot test was a “passive” pre-test, used after initial refinement from the first pilot test. A second small group (n=20) of target audience members took the questionnaire, no di scussion occurred; rather the researcher simply assessed the response quality and the final construction of the instrument. n = ( Pij Sij)2 i=1 n = ( Pij Iij)2 i=1 & where Pij = brand image perception (i) of individual (j) Sij = self-image (i) of individual (j) Ii j = self-image (i) of individual (j)


48 Primary Survey Respondents were college students, recr uited from undergraduate classes at a large, southeastern university. A studen t sample was deemed appropriate by the researcher, because many self-brand congruity studies use student samples (Malhotra, 1981) and the self-image scales derived fr om psychology are based on student sample research (Osgood et al., 1957; Wylie, 1975) . A convenience sample of two-hundred seventy two (n=272) undergraduate college st udents participated in the primary survey, recruited from four differen t undergraduate large courses, representing a wide spectrum of the institution’s student population. Administration of the questionnaire occu rred in classroom group settings during the regular scheduled time. Extra credit provi ded an incentive for participating in the survey. From the 272 complete surveys, a majority of the respondents were females (n=170) 62 percent, while 38 percent were male (n=102). Respondents represented the full range of college classification with 24 percent freshman (n=66), 17 percent sophomores (n=46), 38 percent juniors (n=101) , and 22 percent seniors (n=58), while 87 percent of respondents were between the ages 17 and 21. After coding and entering the data, analyses occurred vi a SPSS statistical package. Description of Data Analysis The researcher first employed Cronbach’s Alpha reliability test on the scales employed in the survey instrument. Descriptive statistics are reported in regards to UIC, BPC, and the attitudinal measures for each analyzed brand. In addition, brand personalities and respondent self-image are reported. Analysis addr essed each research question by conducting separate anal yses for each evaluated brand.


49 Statistical techniques employed included correlation and multiple regression analyses. Correlation analysis provides insight into the relationships between variables, while stepwise regression analysis investig ated the linear relations hip between self-brand congruity indicators when modeling brand atti tudes. The independent variables were the four self-brand congruity indi cators: actual-BPC, ideal-BPC, actual-UIC, and ideal-UIC, while the dependent measure is attitude toward the brand (Ab). In addition, analysis includes correlations between each self-brand indicator and the additional brand attitude constructs (i.e., desirability, im pression, and purchase intentions). The regression analyses followed the a pproach to multivariate model building outlined by Hair et al., (1998). To examine di fference across brands, a similar regression model for each brand is included in the final analysis. Moreover, the regression analyses compared R2 and beta weight values in the st epwise technique to determine which congruity indicator explained the most varian ce, hence added the most explanatory power to the model and indicating a strong predictor of brand attitudes. The results of this analysis indicate (1) the self-i mage dimension that best pred icts the tested relationship and (2) which congruity phenomena (i.e., BPC or UIC) or combination of, in relation to self-image, provides the str ongest attitude predictor for each brand. The following chapter 4 outlines and synthesizes the results from the analysis.


50 CHAPTER 4 RESULTS This chapter presents the descriptive stat istics for brand familiarity and attitude ratings, followed by a description of each bran d’s personality and self-measures of the respondent sample indicated by the BPS scales. Following the descriptive statistics, the researcher addressed each research objectiv e presented in chapter two via correlation and stepwise multiple regression analyses. Brand Familiarity and Attitude Familiarity ratings for each brand were measured to exclude respondents from analysis that were not familiar with a part icular brand. As planned, from preliminary research, each brand had high levels of familiarity. However, the brand Birkenstock did display low familiarity (i.e., two or less on a seven-point scale) ratings for 37 of the 272 total respondents. Hence, the analysis fo r Birkenstock was limited to 235 respondents, while 272 surveys were deemed useful for the brands Abercrombie & Fitch, Banana Republic, and Nike. Brand attitude (Ab) was measure via three seven-point semantic scales (good vs. bad, favorable vs. unfavorable, likable vs. unlikable) and oper ationalized as an index of those scales. To ensure the reliability of the attitude measures, thr ee additional attitudinal measures were included: brand desirabilit y, impression, and intention, measured with five-point Likert scales. Illustrated in Ta ble 4-1, the brand Nike ranked highest on each attitudinal measure, Banana Republic second, Abercrombie & Fitch third, and Birkenstock was rated the lowest on each scale.


51 Correlation analysis showed a strong relationship between the brand desirability, impression, and intention measures with the br and attitude index for each of the four brands analyzed. This indicates that the semantic-differential based brand attitude indexed scales provide a reliable attitude measure; thus to avoid redundancy, the remaining analysis need only to utilize the indexed brand atti tude scores as the dependent variable. Table 4-1. Brand attitude indices: desirability, impressi on, and purchase intentions Brand n Brand Attitude Index Rank Brand Desirability Rank Brand Impression Rank Purchase Intention Rank Abercrombie & Fitch 272 4.98 3 3.46 3 3.41 3 3.57 3 Birkenstock 253 4.54 4 3.03 4 3.20 4 2.90 4 Banana Republic 272 5.42 2 3.72 2 3.82 2 3.75 2 Nike 272 6.10 1 4.15 1 4.21 1 4.06 1 Additionally, the researcher examined th e internal consiste ncy of the brand attitude indices via Cronbach’s Alpha reliab ility coefficients for each brand. The results showed high reliability for the brand attit ude indices for Abercrombie & Fitch (.9514), Birkenstock (.9699), Banana Republ ic (.9579), and Nike (.9433). Brand Personality Each brand and definition of self (i.e., actual and ideal ) was analyzed based on the mean for each BPS dimension and assigned a personality based on the two highest rated BPS dimensions. Respondents rated personal ity dimension indicators (e.g., see Chapter 3) to how well they described each brand as a person. Each brand personality dimension represents an index of its i ndividual indicators for that BPS facet. Table 4-2 shows the results of the mean index scores of the pers onality dimensions associated with each brand and the brand personality indicated by the two BPS dimension rated highest by respondents.


52 The review of the BPS litera ture (i.e., Aaker, 1997), for profiles of the personality types provides more detail of the brand personality type. For example, the brand Abercrombie & Fitch, rated highest on the br and personality dimension “sophistication,” is characterized as charming and upper cl ass, while rated high on the excitement dimension, capturing a sociable, en ergetic, and activ e personality. The brand Birkenstock, rated highest on “sin cerity,” captures the idea of a warm and accepting brand personality. Birkenstock also rated high on the “competence” dimension, capturing a responsible, dependable, and secure personality. In comparison, the brand Banana Republic, rated highes t by respondents on “s ophistication” and “competence” dimensions, captures an upper cla ss and charming personality that is also dependable and secure. Table 4-2. Brand personality Abercrombie & Fitch Birkenstock Banana Republic N ike Mean Std. dev. Mean Std. dev. Mean Std. dev. Mean Std. dev. Sincerity 2.35 .772 3.46 .833 3.02 .850 3.12 .893 Excitement 3.22 .884 2.50 .864 2.92 .824 3.87 .785 Competence 2.63 .920 3.28 .970 3.86 .902 3.44 .978 Sophistication 3.30 .998 2.65 .974 3.97 .864 2.63 .926 Ruggedness 2.27 1.08 3.02 1.21 1.76 .892 4.32 .857 Overall Sophistication/ Excitement Sincerity/ Competence Sophistication/ Competence Ruggedness/ Excitement The last brand, Nike, rated highest by respondents on the BPS dimensions “ruggedness” and “excitement,” captures a pe rsonality characterized as tough, outdoorsy, sociable, energetic, and active. Table 4-3 presents the means of actua l and ideal self-measurement of the respondent sample. The results illustrated a co nsistent profile of the respondents, as indicated college students. In terestingly, both actual and id eal respondent self-measures


53 indicate a dominance on the “competence” personality dimension. This, on the face, appears valid given that the respondents are college students, typi cally striving to be competent via education. Table 4-3. Respondent actual and ideal BPS self measures Actual Ideal N Mean Std. dev. N Mean Std. dev. Sincerity 272 3.77 .629 272 4.35 .620 Excitement 272 3.56 .668 272 4.25 .649 Competence 272 4.04 .646 272 4.79 .477 Sophistication 272 3.40 .851 272 4.49 .776 Ruggedness 272 3.03 1.12 272 3.89 1.02 Overall Competence/Sincerity Competence/Sophistication Overall, the samples’ ideal self was consistently rated higher on each BPS dimension. However, the second highest rated dimension for actual se lf was “sincerity,” while the second highest rated dimension fo r ideal self was “s ophistication”. This indicates that even though competence is mo st commonly used to describe both actual and ideal selves, that the aspirational ideal se lf of this sample of college students is sophistication rather than sincerity. Research Questions The analysis for each research question included both correlation analyses and stepwise regression analyses used to exam ine the linear relationship between BPC and UIC in brand attitude (Ab) models for Abercrombie & Fitch, Birkenstock, Banana Republic, and Nike. For each portion of analysis the independent variables were the congruity indicator scores (See Appendix B fo r the means and standard deviations of these variables), representi ng the calculated distance m easure between respondents’ actual or ideal self-image and (1) bra nd personality, or (2) brand user imagery.


54 Importantly, when more than one indepe ndent variable entered a regression model, the researcher addressed the impact of multicollinearity via diagnostic statistics. Tolerance and variance inflati on factor (VIF) values indi cate how much a particular independent variable is explai ned by other variables in a re gression model. Tolerance is the level of variance accounted for by a particular independent variable not accounted for by other variables. For diagnostic procedur es, tolerance level is “one minus the proportion of a variables varian ce explained by other variables, ” and hence, the closer to 1.0 the lower the level of multicollinearity (Hare, Anderson, Tatham & Black, 1998). The VIF value is the reciprocal of the tolerance level and researchers look for low values (i.e., close to 1.0) to indicate low intercorrelati on between independent variables in multiple regression analysis. A final note regarding the following se lf-brand congruity analysis, negative correlations between self-brand congruity indicators and bran d attitude scores, and the corresponding negative regression coefficients indicate a positive relationship between higher congruity and the dependent measure (i.e., brand attitude). This is because according to theory, as the congruity score goes down, which indicates a closer distance between self-image and brand-image (i.e., highe r congruity), brand attitude scores should increase, which becomes more positive. Actual-BPC vs. Actual-UIC The first research question addressed th e relationship between “actual” brand personality-self congruity (actual-BPC) and “actual” brand user image-self congruity (actual-UIC) as independent explanatory variables in brand attitude models for the four brands analyzed. Tables 4-4 th ru 4-11 display the results of the correlation and stepwise


55 regression analyses for the brands Abercrom bie & Fitch, Birkenstock, Banana Republic, and Nike. Abercrombie & Fitch For the brand Abercrombie & Fitch, the co rrelation matrix sh own in Table 4-4, illustrates that there was a signi ficant, moderate relationship ( r = .458) between this brand’s actual-BPC and actual-UIC indicator s. Additionally, for Abercrombie & Fitch, both actual-UIC ( r = -.698) and actual-BPC ( r = -.448) were signifi cantly correlated with the attitude measure at the .01 level. The negative correla tions indicate that as the congruity score decreases, that is as self and brand perceptions are closer (i.e., higher self-brand congruity) bra nd attitude increases. Table 4-4. Abercrombie & Fitch correla tion matrix: actual BPC and actual UIC Variables Actual BPC Actual UIC Brand Attitude Predictors Actual BPC 1 Actual UIC .458** 1 Dependent Brand Attitude -.448** -.698** 1 ** Correlation is significant at the .01 level (two-tailed) Stepwise multiple regression analysis, displayed in Table 4-5, employed Abercrombie & Fitch’s actual-BPC and actualUIC indicators as independent variables and brand attitude as the dependent variable , resulting in two models with a slight increase in variance explained. Table 4-5. Abercrombie & Fitch stepwise regression: actual BPC and actual UIC Unstand. Coefficients Standard. Coefficients Collinearity Stats. Model B Std. Error Beta T Sig. R2 Tol VIF 8.044 .216 37.179 .000 1 Constant A & F – Actual UIC -.973 .064 -.698 -15.202.000 .487 1.00 1.00 8.748 .306 28.583 .000 -.869 .071 -.624 -12.302.000 2 Constant A & F – Actual UIC A & F – Actual BPC -.139 .044 -.162 -3.194 .002 .508 .790 .790 1.266 1.266


56 The first model incorporated the actual-UIC (R2 = .487) indicator, explaining about 49 percent of variance in attitude towards the brand. The second model integrated actual-BPC, however the variance explained (R2 = .508) was increase d by only about 2 percent, indicating that brand user imag e congruity (actual-UIC) accounted for the greatest amount of variance explained in the dependent measure for Abercrombie & Fitch. Because the second model incorporated tw o independent variables, collinearity statistics provide diagnostics regarding the impact of intercorrelation between these two variables on the final regression variate. The tolerance (.790) and VIF (1.266) values indicate that the regression variate is slightly affected by multicollinearity, since both statistics are close to 1.0, t hough not at a degree to affect the interpretation of the final model. Birkenstock Illustrated in Table 4-6, actual-BPC and actual-UIC were significant, yet moderately related ( r = .383) for the brand Birkenstock, while each congruity indicator showed moderately strong rela tionships to brand attitude. Table 4-6. Birkenstock correlati on matrix: actual BPC and actual UIC Variables Actual BPC Actual UIC Brand Attitude Predictors Actual BPC 1 Actual UIC .383** 1 Dependent Brand Attitude -.429** -.534** 1 ** Correlation is significant at the .01 level (two-tailed) Employing Birkenstock’s actual-BPC and actual-UIC in a stepwise regression analysis resulted in two models displayed in Table 4-7. While moderate, the variance explained (R2 = .285) in attitude towards Birk enstock by actual-UIC was about 28


57 percent in the first model. The second model integrated the actual-BPC indicator and increased the variance explained (R2 = .345) in the dependent measure by about 6 percent. Table 4-7. Birkenstock stepwise regression: actual BPC and actual UIC Unstand. Coefficients Standard. Coefficients Collinearity Stats. Model B Std. Error Beta T Sig. R2 Tol VIF 7.150 .297 24.098 .000 1 Constant Birk – Actual UIC -.753 .084 -.534 -9.004 .000 .285 .854 1.172 8.076 .358 22.573 .000 -.611 .087 -.433 -7.029 .000 2 Constant Birk – Actual UIC Birk – Actual BPC -.209 .049 -.264 -4.276 .000 .345 .854 1.172 For the second model that included each independent variable, both the tolerance (.854) and VIF (1.172) collinearity statis tics are close to 1.0, indicating that multicollinearity has minimal impact on the regression variate. Banana Republic In comparison to the first two brands analyzed, for the brand Banana Republic, illustrated in Table 4-8, the relationship be tween actual-BPC and actual-UIC was not as strong ( r = .263), yet significant at the .01 level. Table 4-8. Banana Republic correla tion matrix: actual BPC and actual UIC Variables Actual BPC Actual UIC Brand Attitude Predictors Actual BPC 1 Actual UIC .263** 1 Dependent Brand Attitude -.180** -.588** 1 ** Correlation is significant at the .01 level (two-tailed) In regards to the relationsh ip to brand attitude, actual -UIC displayed a moderately strong, negative relationship ( r = -.588) with attitude to wards Banana Republic, while actual-BPC displayed a weak negative relationship ( r = -.180), both sign ificant at the .01


58 level. Again, in comparison to the previous two brands, actual-BPC did not display as strong of a relationship to the atti tude measure for Banana Republic. Stepwise multiple regression analysis resulted in a single model and included actual-UIC for Banana Republic; illustrated in Table 4-9. Compared to the previous two brands, actual-UIC accounted for a modera te percentage of variance explained (R2 = .346) in attitude towards Banana Republic, a bout 35 percent, while actual-BPC did not add to the variance explained in this mode l and was excluded. The negative regression coefficient indicates that as self-brand UIC increases, that is higher congruity between self and brand user imagery, brand att itudes should increase by about .773 on the dependent measure. Table 4-9. Banana Republic stepwise regression: actual BPC and Actual UIC Unstandardized Coefficients Standardized Coefficients Model B Std. Error Beta t Sig. R2 7.534 .199 37.885 .000 1 Constant BR – Actual UIC -.773 .069 -.588 -11.241 .000 .346 Nike Similar to the brand Banana Republic, the correlation analysis for the brand Nike, displayed in Table 4-10, demons trated a moderately weak ( r = .268) relationship between actual-BPC and actual-UIC. In regards to attitude towards Nike, both actual-BPC ( r = .216) and actual-UIC ( r = -.424) were significantly rela ted to the dependent measure, with actual-UIC somewhat stronger. Table 4-10. Nike correlation ma trix: actual BPC and actual UIC Variables Actual BPC Actual UIC Brand Attitude Predictors Actual BPC 1 Actual UIC .268** 1 Dependent Brand Attitude -.216** -.424** 1 ** Correlation is significant at the .01 level (two-tailed)


59 The stepwise regression resulted in a si ngle attitude model for Nike with the actual-UIC congruity indicator. The e xplained variance is moderate (R2 = .180), indicating that “actual ” self-brand congruity of either type does not provide a good predictor of attitude towards Nike. Table 4-11. Nike stepwise regr ession: actual BPC and actual UIC Unstandardized Coefficients Standardized Coefficients Model B Std. Error Beta t Sig. R2 7.262 .177 41.032 .000 1 Constant NIKE – Actual UIC -.501 .069 -.424 -7.301 .000 .180 Ideal-BPC vs. Ideal-UIC Similar to the first research question, th e second research question focused on the self-brand congruity between respondents’ “ideal self” perceptions and brand image facets user imagery and brand personality. The go al of this analysis phase was to explore the relationship between “ideal ” brand personality-self congruity (ideal-BPC) and “ideal” brand user image-self congruity (ideal-U IC) when employed in the same model as independent variables used to predict brand attitudes. Abercrombie & Fitch The correlation analysis for the brand Ab ercrombie & Fitch, displayed in Table 412, shows a moderately strong association be tween ideal-UIC and BPC. Both ideal-BPC ( r = -.409) and ideal-UIC ( r = -.694) were significantly rela ted to brand attitude. The UIC indicator illustrated a stronger relationship with the dependent measure. Note the negative correlation, indicative of a positive linear rela tionship between brand attitude and higher ideal-self brand congruity of both type s (i.e., UIC and BPC) for this brand.


60 Similar to the analysis for the actual se lf-brand congruity indi cators, both ideal self-brand congruity indicator s (i.e., UIC and BPC) when regressed against attitude towards Abercrombie & Fitch, resulted in two models (Table 4-13) with the ideal-UIC indicator accounting for the larger percentage of variance explained in positive brand attitude. Table 4-12. Abercrombie & Fitch correla tion matrix: ideal BPC and ideal UIC Variables Ideal BPC Ideal UIC Brand Attitude Predictors Ideal BPC 1 Ideal UIC .401** 1 Dependent Brand Attitude -.409** .694** 1 ** Correlation is significant at the .01 level (two-tailed) The first model included the idea l-UIC indicator, accounting for (R2 = .480) about 48 percent of the explained variance, while the second model incorporated the ideal-BPC indicator with a slight increase in vari ance explained by approximately 2 percent. Multicollinearity diagnostic values, tolera nce (.859) and VIF (1.191), indicate minimal affect of collinearity betw een ideal-UIC and ideal-BPC for the second model. Table 4-13. Abercrombie & Fitch stepwise regression: ideal BPC and ideal UIC Unstand. Coefficients Standard. Coefficients Collinearity Stats. Model B Std. Error Beta t Sig. R2 Tol VIF 7.739 .200 38.733 .000 1 Constant A & F – Ideal UIC -.880 .059 -.694 -15.034.000 .480 1.00 1.00 8.449 .299 28.285 .000 -.801 .063 -.632 -12.763.000 2 Constant A & F – Ideal UIC A & F – Ideal BPC -.109 .035 -.156 -3.150 .002 .498 .859 .859 1.191 1.191 Birkenstock Likewise, for the brand Birkenstock, the correlation analysis, shown in Table 414, illustrates a positive relationship between ideal-BPC and ideal-UIC ( r = .417) at the .01 level. Correlations with attitude toward s Birkenstock showed that ideal-BPC is


61 moderately related ( r = -.437) to the dependent measure, while ideal-UIC ( r = -.583) resulted in a somewhat stronger relationship. Table 4-14. Birkenstock correlation matrix: ideal BPC and ideal UIC Variables Ideal BPC Ideal UIC Brand Attitude Predictors Ideal BPC 1 Ideal UIC .417** 1 Dependent Brand Attitude -.437** -.583** 1 ** Correlation is significant at the .01 level (two-tailed) The stepwise regression for Birkenstock illu strated in table 4-15, as expected from the correlation analysis resulted in a twomodel solution with the ideal-UIC indicator accounting for most of the explaine d variance in brand attitude (R2 = .339). The negative regression coefficients indicate an increase in brand attitude as the distance between ideal self and brand decreases. Similar to the br and Abercrombie & Fitch, when both ideal UIC and BPC entered into the same model, the co llinearity statistics indicated a minimal impact of multicollinearity on the regression variate and variance explained in attitude towards Birkenstock. Table 4-15. Birkenstock stepwise re gression: ideal BPC and ideal UIC Unstand. Coefficients Standard. Coefficients Collinearity Stats. Model B Std. Error Beta T Sig. R2 Tol VIF 7.157 .265 27.050 .000 1 Constant Birk – Ideal UIC -.769 .075 -.583 -10.211.000 .339 .826 1.211 7.947 .328 24.262 .000 -.640 .080 -.485 -7.980 .000 2 Constant Birk – Ideal UIC Birk – Ideal BPC -.157 .041 -.235 -3.867 .000 .385 .826 1.211 Banana Republic Shown in Table 4-16, the correlation ma trix for the brand Banana Republic illustrates that ideal-BPC and ideal-UIC are pos itively correlated, yet the strength of the relationship is moderate ( r = .236). In contrast to the pr evious two brands, UIC and BPC


62 are most distinct for this brand. Additi onally, both ideal congruity indicators are negatively correlated with attitude towards Banana Republic. However, for this brand, ideal-UIC is strongly related ( r = -.626) to brand attitude , while ideal-BPC displayed a moderately weak relationship ( r = -.242) with the dependent measure. Table 4-16. Banana Republic correlati on matrix: ideal BPC and Ideal UIC Variables Ideal BPC Ideal UIC Brand Attitude Predictors Ideal BPC 1 Ideal UIC .236** 1 Dependent Brand Attitude -.242** -.626** 1 ** Correlation is significant at the .01 level (two-tailed) The stepwise regression analysis fo r Banana Republic’s ideal congruity indicators, displayed in Table 417, resulted in a single model (R2 = .391) that included the ideal-UIC indicator, accounting for about 39 percent of the e xplained variance in attitude towards this brand. Likewise, the ne gative regression coefficient (b = -.761) indicates that as the distan ce between respondents’ ideal se lf-perceptions and brand user imagery decrease (higher congruity) positive brand attitudes increase. Table 4-17. Banana Republic stepwise regression: ideal BPC and ideal UIC Unstandardized Coefficients Standardized Coefficients Model B Std. Error Beta t Sig. R2 7.410 .173 42.908 .000 1 Constant BR – Ideal UIC -.761 .061 -.626 -12.394 .000 .391 Nike For the brand Nike, the correlation matrix, displayed in Table 4-18, illustrates that ideal-BPC and ideal-UIC are m oderately related (r = .403) at the .01 level, similar to the first two brands. In addition, both ideal-BPC ( r = .297) and ideal-UIC ( r = -.477) were significantly related to attitude towards Nike , with ideal-UIC resu lting in a stronger relationship to brand atti tude than ideal-BPC.


63 Similar to the brands Abercrombie & Fitch and Birkenstock, the stepwise regression analysis for the ideal congruity indi ctors and attitude towards Nike resulted in two models as displayed in Table 4-19. Table 4-18. Nike correlation ma trix: ideal BPC and ideal UIC Variables Ideal BPC Ideal UIC Brand Attitude Predictors Ideal BPC 1 Ideal UIC .403** 1 Dependent Brand Attitude -.297** -.477** 1 ** Correlation is significant at the .01 level (two-tailed) The first model included the ideal-UIC (R2 = .227) indicator and accounted for about 23 percent of the variance explained in attitude towards this brand. The second model for Nike included both ideal-UIC and ideal-BPC, with the addition of the idealBPC indictor there was a slight increase in the variance (R2 = .240) explained to about 24 percent. Table 4-19. Nike stepwise regr ession: ideal BPC and ideal UIC Unstand. Coefficients Standard. Coefficients Collinearity Stats. Model B Std. Error Beta t Sig. R2 Tol VIF 7.332 .163 44.890 .000 1 Constant NIKE – Ideal UIC -.566 .067 -.477 -8.457 .000 .227 .837 1.194 7.643 .223 34.296 .000 -.506 .073 -.427 -6.968 .000 2 Constant NIKE – Ideal UIC NIKE – Ideal BPC -.067 .033 -.125 -2.035 .043 .240 .837 1.194 In addition, the tolerance (.837) and VIF (1.194) values indicate that multicollinearity had only a slight impact on the second model and the interpretation of the final variate for Nike. Simultaneous Comparison of Congruity Types The objective of the third research questi on was to determine which of the selfbrand image congruity indicators (i.e., act ual-BPC, actual-UIC, ideal-BPC, and ideal-


64 UIC), or combination of, accounts for the most variance when each variable is employed in the same brand attitude prediction model. Si milar to the analyses for research questions one and two, first a correlation analysis fo r each brand illustrated the relationship between the four congruity indicators a nd brand attitude, followed by a stepwise regression analysis for each brand that indi cated the variance explained in the dependent measure by the linear combination of the self-brand congr uity indicators. Abercrombie & Fitch For the brand Abercrombie & Fitch, the co rrelation matrix displayed in Table 420, illustrates the relationship between the se lf-brand congruity indicators and brand attitude. Both the actual/ideal-BPC ( r = .815) measures had a strong relationship as did the actual/ideal UIC ( r = .837) measures. Table 4-20. Abercrombie & Fitch correla tion matrix: actual/ideal BPC and UIC Variables Actual BPC Ideal BPC Actual UIC Ideal UIC Brand Attitude Predictors Actual BPC 1 Ideal BPC .815** 1 Actual UIC .458** .420** 1 Ideal UIC .408** .401** .837** 1 Dependent Brand Attitude -.448** -.409** -.698** .694** 1 ** Correlation is significant at the .01 level (two-tailed) Actual-UIC was moderately related to both actual ( r = .458) and ideal ( r = .420) BPC measures, while similar results were indi cated for the ideal congruity indicators. In regards to attitude towards Abercrombie & F itch, both UIC measures resulted in stronger relationships with brand attit ude than did either of the BP C measures. In fact, similar results are illustrated throughout the an alysis for the next three brands. The stepwise regression for Abercrombie & Fitch, illustrated in Table 4-21, resulted in three models when all four congr uity indicators were included in the analysis.


65 The first model (R2 = .487) included the actual-UIC indicator, accounting for about 49 percent of the variance expl ained in attitude towards Abercrombie & Fitch. A second model (R2 = .546) incorporated the ideal-UIC indi cator, slightly increasing the variance explained in the dependent measure, while the third model (R2 = .546) included the actual-BPC measure and increased the va riance explained by about 6 percent. Table 4-21. Abercrombie & Fitch stepwise regression: actual/ideal BPC and UIC Unstand. Coefficients Standard. Coefficients Collinearity Stats. Model B Std. Error Beta t Sig. R2 Tol VIF 8.004 .216 37.179 .000 1 Constant A & F – Actual UIC -.973 .064 -.698 -15.202.000 .487 1.00 1.00 8.156 .210 38.924 .000 -.545 .112 -.391 -4.853 .000 2 Constant A & F – Actual UIC A & F – Ideal UIC -.465 .102 -.367 -4.550 .043 .528 .300 .300 3.332 3.332 8.808 .295 29.853 .000 -.463 .114 -.333 -4.081 .000 .284 3.524 -.449 .101 -.354 -4.466 .000 .299 3.341 3 Constant A & F – Actual UIC A & F – Ideal UIC A & F – Actual BPC .130 .042 -.151 -3.085 .002 .546 .788 1.269 Collinearity statistics demonstrated tw o important findings when employing each congruity indicator in the same regression model. First, as indicated by the low tolerance (.300) and high VIF (3.332) values for model two, there was a high le vel of collinearity between the two UIC indicators, while there wa s less collinearity between the actual UIC and BPC (i.e., tolerance = .788 and VIF = 1.269) indicators displayed in model three. Second, the remedy is to omit the highly correl ated UIC variables, which shows that the model displayed earlier in Table 4-6, provides a sufficient variate to represent the linear relationship between UIC and BPC when modeling attitude towards Abercrombie & Fitch.


66 Birkenstock Similar to the previous brand, the corr elation analysis displayed in table 4-22, shows strong relationships between actual and ideal-UIC measures ( r = .860) and between the actual and ideal-BPC measures ( r = .808), while the relationship across measures (e.g., actual-BPC and actual UIC) were significant, yet the strength of the relationships were moderate, ranging betw een .383 and .417. In regards to attitude towards Birkenstock, both UI C indicators (i.e., actual r = -.534 and ideal r = -.583) displayed slightly stronger relationships to the dependent measure than did the BPC indicators (i.e., actual r = .429 and ideal r = -.437). Table 4-22. Birkenstock correlation matrix: actual/ideal BPC and UIC Variables Actual BPC Ideal BPC Actual UIC Ideal UIC Brand Attitude Predictors Actual BPC 1 Ideal BPC .808** 1 Actual UIC .383** .374** 1 Ideal UIC .411** .417** .860** 1 Dependent Brand Attitude -.429** -.437** -.534** -.583** 1 ** Correlation is significant at the .01 level (two-tailed) The results of the stepwise regression fo r attitude towards Birkenstock illustrated in table 4-23, employing all four-congru ity indicators produce d different results compared to the previous brand (i.e., Ab ercrombie & Fitch). Two models resulted, however, the first model incorporated the ideal-UIC (R2 = .339) indicator rather than actual-UIC and accounted for about 34 percent of the variance in the dependent measure. The second model included the ideal-BPC indicator and increa sed the explained variance slightly by about 5 percent.


67 Table 4-23. Birkenstock stepwise re gression: actual/ideal BPC and UIC Unstandardized Coefficients Standardized Coefficients Model B Std. Error Beta t Sig. R2 7.157 .265 27.050 .000 1 Constant Birk – Ideal UIC -.769 .075 -.583 -10.211 .000 .339 7.947 .328 24.262 .000 -.640 .080 -.485 -7.980 .000 2 Constant Birk – Ideal UIC Birk – Ideal BPC -.157 .041 -.235 -3.867 .000 .385 In addition, both the actual UIC and BPC measures were excluded by the stepwise procedure from the final attitude model for Birkenstock. In regard s to the affect of multicollinearity in model two, since the re gression model did not change, that is incorporate any additional inde pendent variables compared to analysis reported in table 4-15 above, those results repr esent the relationship betw een these two independent variables. Banana Republic For the brand Banana Republic, the results of the correlation anal ysis displayed in table 424 showed somewhat different results compared to the other brands in the study. Table 4-24. Banana Republic correlati on matrix: actual/ideal BPC and UIC Variables Actual BPC Ideal BPC Actual UIC Ideal UIC Brand Attitude Predictors Actual BPC 1 Ideal BPC .773** 1 Actual UIC .258** .292** 1 Ideal UIC .189* .236** .800** 1 Dependent Brand Attitude -.176** -.242** -.588** -.626** 1 ** Correlation is significant at the .01 level (two-tailed) * Correlation is significant at the .05 level (two-tailed) For instance, similar to the other brands, the actual and ideal-BPC indicators were significantly related ( r = .773) as were the actual and ideal-UIC indicators ( r = .800) at the .01 level. In contrast to the other bra nds analyzed, the strength of the relationship


68 between the UIC and BPC measures was wea k, particularly between the ideal-UIC and actual BPC ( r = .189) indicators. In a similar manner, for Banana Repub lic, both UIC indicators showed much stronger relationships to brand attitude than did the either of the BPC indicators, in particular, the actual-BPC indi cator which resulted in a weak relationship (r = .176) with attitude towards Banana Republic. These resu lts are evident in the following stepwise regression analysis. Stepwise regression analysis employing each of the four congruity indicators for Banana Republic brand attitude , illustrated in Table 4-25, re sulted in two models with a slight increase in variance explained in the dependent measure. The first model incorporated the ideal-UIC (R2 = .391) indicator, accounting for about 39 percent of the variance. Table 4-25. Banana Republic stepwise regression: actual/ideal BPC and UIC Unstand. Coefficients Standard. Coefficients Collinearity Stats. Model B Std. Error Beta t Sig. R2 Tol VIF 7.410 .173 42.908 .000 1 Constant BR – Ideal UIC -.761 .061 -.626 -12.394.000 .391 .361 2.773 7.663 .190 40.238 .000 -.524 .101 -.431 -5.205 .000 2 Constant BR – Ideal UIC BR – Actual UIC -.320 .109 -.244 -2.947 .004 .413 .361 2.773 The second model incorporated the actual-UIC (R2 = .413) indictor, increasing the variance explained in brand attitude to about 41 percent. In addition, both of the BPC congruity indicators were excluded from the final model, indicating that user image congruity alone best predicts positive brand attitudes for Banana Republic. However, a low tolerance level (.361) and high VIF (2.773) indicates an impact of multicollinearity on model two and that correctiv e remedies should be taken be fore final interpretation.


69 Nike For the brand Nike, the correlation and regression analyses results produced results similar to the other three brands, with slight differences in which of the four indicators accounted for the most variance in the stepwise procedure. The correlation analysis, displayed in Table 4-26, like the other brands, showed strong relationships between the actual and ideal BPC measures as well as the UIC measures. Comparing across indicators, the actual-UIC measure was only moderately related to the actual-BPC ( r = .268) and ideal-BPC ( r = .285) indicators, while idea l-UIC displayed a slightly stronger relationship with actual-BPC ( r = .313) and ideal-BPC ( r = .403). Table 4-26. Nike correlation matrix: actual/ideal BPC and UIC Variables Actual BPC Ideal BPC Actual UIC Ideal UIC Brand Attitude Predictors Actual BPC 1 Ideal BPC .699** 1 Actual UIC .268** .285** 1 Ideal UIC .313** .403** .724** 1 Dependent Brand Attitude -.216** -.297** -.424** -.477** 1 ** Correlation is significant at the .01 level (two-tailed) Similar to the other three brands, the co rrelations between atti tude towards Nike and the four congruity indicators showed that both the UIC indicators are have a stronger relationship to the dependent measure than di d either of the BPC indicators. Both actualBPC ( r = -.216) and ideal-BPC ( r = -.297) showed a moderately weak association with brand attitude at the .01 level. In contrast, actual-UIC ( r = -.425) and ideal-UIC ( r = .478) displayed a slightly stronger, yet stil l moderate relationship to the dependent measure. The stepwise regression analysis incorpor ating all four-congruity indicators as predictors of Nike’s brand attitude score re sulted in three models, displayed in Table 4-


70 27. The first model incorporated the ideal-UIC indicator (R2 = .227), explaining about 23 percent of the variance in br and attitude. The second model incorporated the actual-UIC measure (R2 = .240) and increased the variance expl ained to about 24 percent. The third model incorporated the ideal-BPC indicator (R2 = .254) and increased the variance explained slightly more to about 25 percent, for a total increase in variance explained in the final model by about 3 percent. Note, for this stepwise regre ssion, the only congruity indicator excluded from each regression m odel was Nike’s actual-BPC indicator. Table 4-27. Nike stepwise regr ession: actual/ideal BPC and UIC Unstand. Coefficients Standard. Coefficients Collinearity Stats. Model B Std. Error Beta T Sig. R2 Tol VIF 7.332 .163 44.890 .000 1 Constant NIKE – Ideal UIC -.566 .067 -.477 -8.457 .000 .227 1.00 1.00 7.480 .178 42.077 .000 -.423 .096 -.357 -4.392 .000 2 Constant NIKE – Ideal UIC NIKE – Actual UIC -.196 .096 -.166 -2.038 .043 .240 .475 .475 2.104 2.104 7.796 .233 33.414 .000 -.361 .100 -.305 -3.604 .000 .433 2.309 -.198 .095 -.168 -2.075 .039 .475 2.104 3 Constant NIKE – Ideal UIC NIKE – Actual UIC NIKE – Ideal BPC .068 .033 -.126 -2.072 .039 .254 .837 1.194 For models two and three, multicollinearity diagnostics displayed similar results to the brand Abercrombie & Fitch when compar ing all four indicators in the same brand attitude model. For model two, the moderate ly low tolerance (.475) and high VIF (2.104) values are indicative of multicollinearity be tween the UIC indicators, while tolerance (.837) and the VIF (1.194) values for the ideal BPC indicator demons trated a lack of collinearity with this variable and the othe r independents. Hence, caution should be taken when interpreting the final model without corrective remedies.


71 Summary of Results A total of twelve stepwise multiple regression analyses addressed the research objectives and questions. The first four comp ared actual-UIC and actual-BPC indicators’ relationship to attitude towards the brands Abercrombie & Fitch, Birkenstock, Banana Republic, and Nike. Overall, actual-UIC indicat ors consistently accounted for the most variance explained in atti tude towards the brand (Ab). The second set of four stepwise regressions compared ideal-UIC and ideal-BPC indicators’ re lationship to brand attitude across the same set of brands , illustrating similar results found for the actual congruity indicators. Table 4-28. Summary of varian ce explained in brand attit ude by congruity indicators Ab Abercrombie & Fitch Ab Birkenstock Ab Banana Republic Ab Nike Actual-UIC .487 Excluded .022 .013 Ideal-UIC .041 .339 .391 .227 Actual-BPC .018 Excluded Excluded Excluded Ideal-BPC Excluded .046 Excluded .014 The third set of regression analyses incor porated all four cong ruity indicators as independent variables compared to the brand attitude scores. This set of regressions provided insight into which combination of self-brand indicators accounted for variance explained in the dependent measure. Table 4-28, illustrates a summary of the variance explained by each self-brand congruity indicat or that resulted from the third set of regression analyses, including th e variables excluded by the stepwise procedures. From this summary table, it is evident that id eal-UIC consistently accounts for the most variance in the dependent measure, while the BPC indictors do add to the variance explained for three of the four brands analyz ed, particularly Birken stock. The researcher employed tolerance and variance inflation factors (VIF) values to examine the degree of


72 multicollinearity and determine its affect on th e final regression model and interpretation when multiple independent variables entered a model; implications are addressed in the following discussion chapter. The following two summary tables (i.e., Table 4-29 and 4-30) illustrate the correlations between variables for each bra nd. Displayed in table 4-29, the first column after the brand names displays the correlati ons between actual-BPC and actual-UIC. The strongest relationship between these two constructs is for the brand Abercrombie & Fitch (.458) and the brand Birkenstock (.386). In contrast, the relationship between BPC and UIC is weakest for the brands Banana Republic (.263) and Nike (.268). Table 4-29. Summary table: actual congruity indicators Brand Actual BPC/UIC Correlation Actual BPC/Ab Correlation Actual UIC/Ab Correlation Abercrombie & Fitch .458 -.448 -.698 Birkenstock .386 -.429 -.534 Banana Republic .263 -.180 -.588 Nike .268 -.216 -.424 The third and fourth colu mns summarize the correlations between Actual BPC/brand attitude (Ab) and Actual-UIC/brand attitude (Ab) for each brand. Actual-BPC displayed the strongest rela tionships with the brand Ab ercrombie & Fitch (-.448) and Birkenstock (-.429) and weakest relationship with the brands Banana Republic (-.180) and Nike (-.216). In comparison, Actual-UIC di splayed moderately strong relationships with brand attitude (Ab) for each brand, particularly the brand Abercrombie & Fitch (.698).


73 Likewise, table 4-30 summarizes simila r results for the “ideal” congruity indicators. The first column shows the correlation between ideal BPC and UIC, the strongest relationship is fo r the brand Birkenstock ( r = .487). Table 4-30. Summary table: ideal congruity indicators Brand Ideal BPC/ UIC Correlation Ideal BPC/ Ab Correlation Ideal-UIC/ Ab Correlation Abercrombie & Fitch .401 -.409 -.694 Birkenstock .487 -.437 -.583 Banana Republic .236 -.242 -.626 Nike .403 -.297 -.477 In contrast, the weakest relationship between ideal BPC and UIC was for the brand Banana Republic (.236). In regards to brand attitude (Ab), ideal-BPC had a moderately strong relationship to attitude towards Abercrombie & Fitch (-.409) and Birkenstock (-.437). The ideal-UIC congruity indicator resulted in the strongest relationships with each brand, in particular the brands Abercrombie & Fitch (-.694) and Banana Republic (-.626). The following chapter synthesizes the a bove results in a discussion regarding implications in relation to the theoretical framework presented in chapter two, and implications for communication professionals in relevant fiel ds. In addition, limitations to the study’s design and their a ffect on the results are overviewed, as well as areas for future research and a brief conclusion providing closure to the present study.


74 CHAPTER 5 DISCUSSION AND CONCLUSION The self-brand congruity m odel hypothesizes that favorab le brand attitudes should result as the congruence increases between consumer self-image and a brand’s image. Researchers typically operationalize self -brand congruity with indicator scores representing the measured distance between self and brand image. The latter, brand image, is traditionally captured in respondent s’ perceptions of the stereotypical brand user, referred to as brand user imagery. D ecades of studies have conceptualized brand image via the “user imagery” construct, while other brand image constructs received little if any empirical consideration in th e domain of self-congruity theory. Constructs are the building blocks of theory. Encourag ing congruity researchers to examine different brand image constructs will enhance the useful ness of the self-brand congruity model. Because the primary purpose of the self-brand congr uity model is to predict consumer brand attitudes and since br and image represents the cognitive structure of brand attitudes (Fishbein & Ajzen, 1975), the author argues a broader conceptualization of brand im age enhances the application of self-congruity in modeling brand attitudes. The present study cataloged a conceptual and empirica l comparison of two types of self-brand congruity, based on distin ct brand image comp onents [i.e., Brand Personality Congruity (BPC) and Brand User Imagery Congruity (UIC)]. From survey data, the researcher determined UIC and BP C for four publicly consumed brands and conducted statistical comparisons illustrating the nature of their relationship when used to


75 model brand attitudes. Implications for theo ry, professionals, and the discipline of mass communication follow, as well as limitati ons imposed on the study conclusions and suggestions for future research. Theoretical Implications This study’s primary empirical goal was to compare brand user imagery based congruity and brand personality based congrui ty within the framework of self-congruity theory. The first two objectives isolated respondents’ actual and ideal self-image dimensions, thereby allowing the researcher to control the modera ting effects of selfimage and directly compare UIC and BPC. The third research objective simultaneously examined UIC and BPC with different sel f-image dimensions, enabling a comparison within and across self-brand congruity types. To achieve the research objectives, stepwise multiple regression procedures and related sta tistics illustrated the linear relationship between UIC and BPC, and their individual contributions modeling brand attitudes. Research Question One The objective of the first research was to compare actual-UIC and actual-BPC, that is, the congruence between respondents’ actual self-image and the discussed brand image facets across different brands. For each brand, results indicated a significant, yet moderately weak association between the two congr uity types. The strongest relationship between UIC and BPC was for the brand Aber crombie & Fitch; the weakest relationship was for Banana Republic. In general, this shows that the rela tionship between the “actual” congruity indicators differs across brands. In addition, there was a significant relati onship between brand attitude and both congruity types for each brand. However, actual -UIC displayed a stronger relationship to the attitudinal measure for each brand, partic ularly with attitude toward Abercrombie &


76 Fitch. In comparison, actual-BPC had the strongest relations hip with attitude towards Birkenstock, and weakest with atti tude towards Banana Republic. Evident from these results, actual-UIC accounted for the most variance in brand attitude regression models. However, for two brands (i.e., Abercrombie & Fitch and Birkenstock), actual-BPC was added to the at titude model by the stepwise procedures, accounting for a portion of variance explained in favorable brand attitudes. Collinearity diagnostics showed slight deviations from acceptable levels in the two models that included both UIC and BPC, indicating a low impact of multicollinearity on the stability of the coefficient of determination (R2) and statistical independence of the explanatory variables. Therefore, in regards to the first re search question, the results provided two important theoretical implicati ons. First, the statistical ev idence was not strong enough to suggest complete independence between UI C and BPC. Rather, the low correlations between the constructs was indicative they are complementary, but not linear dependent to the degree that affects their individual contributions modeling brand attitudes. Second, actual-BPC adds to the predic tion of favorable brand attit ude, at least for particular brands. This, in itself, suggests differences across brand types between UIC and BPC. In this regards, respondents rated Nike and Banana Republic highest on the attitude measure, both of which were not strongly relate d to “actual BPC.” This raises the issue of why actual-BPC was more predic tive of favorable brand att itudes for the lower evaluated brands. Research Question Two The second goal of analysis was to exam ine the relationship between “ideal-UIC” and “ideal-BPC,” that is, the congruence be tween respondents’ idea l self-image and the


77 two brand image dimensions across different br ands. Similar to the “actual” self-brand congruity indicators, there was a moderately weak linear association between these two congruity types, particularly weak for Ba nana Republic, while slightly stronger for Birkenstock. Pertaining to the association to brand at titude, ideal-UIC consistently displayed the strongest relationship to favorable attit udes compared to idea l-BPC. Nevertheless, ideal-BPC had a slightly str onger relationship to brand atti tude across brands compared its counterpart “actua l-BPC.” This consistency require s further investigation, with a variety of different brands. As expected, the stepwise regression mode ls first incorporated ideal-UIC for each brand, while ideal-BPC s lightly increased the variance explained in attitudes for all brands except Banana Re public. For the attitude models with two indicators, diagnostics indicated low impact of multicollinearity on the final regression models. Therefore, in regards to the second re search question, the present evidence indicates ideal-UIC and BPC ar e not perfectly independent. More accurately, they are somewhat complementary, though not linear de pendent when modeling brand attitudes; similar to the nature of the “actual” congruity indicators. In other words, the empirical evidence supports the argued c onceptual distinctions betw een the self-brand congruity types and showed that ideal-BPC is more often a stronger predictor of positive brand attitudes than actual-BPC, most evident wh en comparing within congruity types. Research Question Three The third research question addressed the relationship between each congruity indicator when compared simultaneously. This allowed an examination of the relationships within (e.g., actual-UIC and ideal-UIC), a nd across (e.g., actual-UIC and


78 actual-BPC) self-brand congruity types, i ndicating which indicator type(s) best-model favorable brand attitudes. As anticipated, w ithin self-brand congruity types there was a strong linear relationship between ideal and actual indicators, wh ile across types the relationship was typically w eak, as demonstrated above. Overall, the comparison of brand attitude with each congruity indicator showed that UIC, either actual or ideal, displayed a st ronger relationship to br and attitude than did BPC indictors. When compari ng within congruity types, di fferences were evident across the analyzed brands. For the BPC indicat ors, ideal-BPC gene rally had a stronger association to brand attitude than actual-BPC, while the same held true for both the ideal and actual-UIC indicators. Interestingly, wh en employing all four indicators in the stepwise procedures, the manner in which si gnificant congruity i ndicators entered the regression model illustrated differences across brands and self-brand congruity types. Because of strong correlations within congr uity types, for this portion of analysis, the existence of multicollinearity was most problematic. When both UIC and/or BPC indicators entered a model, this limited the ab ility to differentiate the independent effects of the explanatory variables among strongl y related congruity indicators when interpreting the coefficient of determination (i.e., R2). The researcher remedied this by eliminating the highly intercorrelated congrui ty indicator that en tered the regression model in later stages. The logi c in doing this is that the stepwise regression procedure first incorporates the indepe ndent variable with the st rongest relationship to the dependent. Hence, the variable entered first naturally accounts for the most variance, and the highly correlated independent variable entered in later steps was determined redundant.


79 The stepwise procedures for Abercrombie & Fitch first incorporated actual-UIC into the regression model, ideal-UIC entere d next, followed by ideal-BPC. Since actualUIC entered the model first and because id eal and actual-UIC are intercorrelated, the researcher removed the ideal-UIC indicator, while the ideal-BPC di splayed little impact of multicollinearity and remained in the model. In comparison, the final attitude model for Nike was a mirror image of Abercrombie & Fitch because the stepwise procedure first incorporated the ideal-UIC indicator, followe d by the actual-UIC i ndicator, and then ideal-BPC. Actual-UIC was removed, while the impact of multicollinearity for ideal-BPC was not significant and it remained in th e final model. For Birkenstock, ideal-UIC entered the attitude model in the first step, followed by ideal-BPC; neither actual-self congruity indictor, either UIC or BPC, entere d the model. Banana Republic differed from the other three brands in that both BPC i ndicators were excluded from each step by the stepwise procedure, while ideal-UIC entered first and actual-UIC was eliminated because of multicollinearity. Theoretical Summary In regards to the self-b rand congruity framework a nd the modeling of selfexpressive brand attitudes, BPC provides an additional brand image construct useable in the operationalization of the congruence phe nomenon. Similar to th e traditional user imagery based self-brand congruity, BPC te nds to differ based on respondents’ selfimage dimension; at times actual-BPC best predicts brand attit udes, while in other instances ideal-BPC drives favorable attit udes. When BPC was compared to UIC, the user imagery type congruity consistently had the strongest relationshi p to brand attitude. This suggest that not only should congruity researchers explore different symbolic


80 aspects of a brands image, but those that study the construct brand personality need to account for the user imagery construct. In light of this, self-c ongruity theory provides a framework for studying the construct brand personality a nd its relationship to brand at titudes. The brand personality construct, recently conceived by communication scholars does not belong to a particular theoretical domain. Incorporating this bra nd image construct into the self-congruity model, as a basis for modeling brand att itudes, provides researchers a theoretical framework to study brand personality. Professional Implications Communicating a clear and c onsistent brand image is central to establishing a product’s marketplace position. Communication prof essionals in the fields of advertising and public relations are regular ly employed to plan the strategic foundations for future branding efforts. To this end, brand pers onality profiles are regularly part of communication platforms used to guide creativ e message development. A prerequisite to the development of effective brand messages is an understanding of what drives target audiences’ likes and dislikes, dr iving strategist to identify th e most important benefits to imbue in a brand image. Since ultimately brand attitudes are a functi on of salient brand image associations, the self-brand congruity m odel provides a framework for comparing different brand image facets and determining which relate to favorable brand attitudes. Although this was the first study of its kind, the implications raise awareness th at brand personality and user imagery need to be examined independently or at least both in cluded in the planning process to determine their importanc e in a brand’s overall identity.


81 Central to developing and maintaining a long-term brand image strategy is an understanding of the dyadic relationships fo rged between consumers and brands. The nature of congruity theory focuses on this relationship and the appl ication of self-brand congruity research contributes to the comp rehensiveness of a consumer centered brand strategy, particularly for brands that are c onsumed for symbolic and social reasons. From the results of this study, implications regarding positioning strategies for the analyzed brands are evident. For example, Nike should be positioned towards target consumers’ ideal self-motivations, as a brand th at promises the user an expression of their aspirational selves that is both “rugged” and “exciting” in nature. Banana Republic never related to brand personality dimensions, rather ideal user imagery should be at the center of this brand’s positioning strategy; the imag e of the type of person the consumer wants to project will drive consump tion. For the brand Birkenstock, both ideal user imagery and brand personality should be part of the pos itioning strategy, or on the other hand, strategy for this brand should position away from the actual user imagery of “earthy” and “downto-earth” and closer to the ideal brand pers onality of “sincerity” and “competence.” Only for Abercrombie & Fitch should the brand posit ion towards consumers’ actual self-image motives with associations such as “competence” and “sincerity” highlighted in communications. Limitations and Future Research This study was exploratory in nature and presented an original comparison using the newly developed construct referred to as BPC. Several limitations of the study are identified that restrict the conclusions that can be made, though they do not preclude the value of incorporating a more encompassi ng conceptualization of the brand image construct in the self-congruity model. In stead, via their acknowledgement, areas for


82 advancement in future research are recomme nded, enhancing the contributions of this study and maximizing the potential of incorpor ating a more holistic conceptualization of the brand image construct in se lf-brand congruity research. There are at least two reasons the use of the brand personality scales and techniques used to calculate BPC type se lf-brand congruity limited this study. The BPS scales were not originally de veloped to assess respondent se lf-image or for gauging selfbrand congruity. Further care must be taken in replication and constr uct validation studies to address this issue. Additionally, as ar gued by congruity theorist Sirgy (1997), the technique of calculating c ongruity scores using the same scales to measure both respondent self-image and brand image pe rceptions are hampered by methodological flaws (e.g., see Chapter 2 for a discussion of these flaws). These same flaws may have affected the present results. To address the above issue, future research would benefit from the development of a “global” type BPC measure similar to the user-image measure or other innovative techniques that would aid in the measurement of the BPC construct. With the advent of such a measure or technique, future resear chers can conduct a variety of studies that would incorporate the BPC construct into the self-congruity framework and more accurately study its relationshi p to the traditional UIC type self-brand congruity. Another limitation of the pr esent study was the lack of any moderating variables incorporated into the design. Previous studies have demonstrated several key variables moderate the relationship between user-image ry type congruity and brand attitudes. Research needs to determine if such variables also affect BPC type congruity in a similar manner. For example, the variable, “self-monitoring behavior” has been demonstrated to


83 play a significant role in the self-brand congruity model (Aaker, 1999). Does selfmonitoring behavior moderate BPC type se lf-brand congruity and UIC in the same manner? Researchers should address this same question for a number of other moderating variables such as brand/produc t type and different responde nt segmentation variables. For several reasons, respondent characteris tics and differences not addressed in this study may play an important role in the self-brand congruity phenomenon. The present sample was limited to a gender imbalanced (i.e., mostly females) college student population. The sample represents top echelon st udents that tend to focus on their idealist selves and are also heavy consumers of “badg e” brands. Studies need to address different consumer segments, from diverse populati on stratum (e.g., older adults and different ethnicities), in order to advance the self -congruity model. In particular, cultural differences may be a significant factor in better understandi ng this phenomenon. The present sample was limited to an Anglo-Eu ropean, self-individua listic culture, from which individuals’ self values are expressed in terms of “I.” On the other hand, self-brand congruity may function differently in a co llectivist cultu re, where self values are predominately expressed in term s of “we” rather than “I.” In regards to the designed study, three additional limitations are evident that future research can address. First, the study was limited to a convenience sample of college students, restricting th e conclusions to this sample. Future research should utilize probability sampling techniques and examine di fferent population segments as discussed. Second, this study analyzed four brands fr om a limited product category range; several studies are necessitated to examine a dive rse set of brands and catalog differences between BPC and UIC. Third, future research can advance the present research findings


84 by applying different statistical techniques such as structural equation modeling to depict the relationship between vari ables and cluster analyses to examine differences in consumer segments. As a final note, this study and related rese arch stream is restricted by a lack of theoretically developed and agreed upon definitions of “brand image” and related constructs such as brand personality. The research community needs conceptual development in this area to provide a paradigmatic foundation to consistently operationalize brand image and streamline rese arch, maximizing theoretical development in regards to the self -brand congruity model and the relate d structure of brand attitudes. Conclusion As with many explorations of discove ry, scientific endeavors are seldom conclusive, rather they offer insights into ne w beginnings. The nature of this dissertation is rather similar, new insights garnered into the nature of self-brand congruity from the original conception of the research problem through the final data analysis. The author focused attention on an area researcher s have bypassed in regards to the conceptualization of brand image and the affect on consumer attitude s. In framing selfbrand congruity theory in this way, new ideas and concepts for the management of strategic brand communications and research were advanced, while opening doors for the agenda of future brand image researchers. In conclusion, ultimately, creative ideas are the key to innovation in scien tific endeavors and no matter the results; an original contribution of new knowledge i nvigorates a discipline and is powerful in its own right.







PAGE 100


PAGE 101


PAGE 102


PAGE 103


PAGE 104

92 APPENDIX B DESCRIPTIVE STATISTICS: INDEPENDENT VARIABLES Independent Variable Mean Standard Deviation UIC Indicators Actual UIC Abercrombie & Fitch 3.16 1.19 Ideal UIC Abercrombie & Fitch 3.15 1.31 Actual UIC Birkenstock 3.40 1.09 Ideal UIC Birkenstock 3.36 1.16 Actual UIC Banana Republic 2.68 1.03 Ideal UIC Banana Republic 2.54 1.12 Actual UIC Nike 2.40 .98 Ideal UIC Nike 2.25 .98 BPC Indicators Actual BPC Abercrombie & Fitch 7.34 1.95 Ideal BPC Abercrombie & Fitch 8.78 2.39 Actual BPC Birkenstock 6.66 1.97 Ideal BPC Birkenstock 7.80 2.31 Actual BPC Banana Republic 6.50 1.74 Ideal BPC Banana Republic 7.75 2.10 Actual BPC Nike 6.24 1.79 Ideal BPC Nike 6.57 2.13

PAGE 105

93 LIST OF REFERENCES Aaker, D.A. (1996). Building strong brands. New York, NY: The Free Press. Aaker, J.L. (1997). Dimensions of brand pe rsonality. Journal of Marketing Research, 34, 347. Aaker, J.L. (1999). The malleable self: The ro le of self-expression in persuasion. Journal of Marketing Research, 36, 1, 45. Aaker, D., & Biehl, A. (1993). Brand equity in advertising. Hillside, NJ: Lawrence Erlbaum. AMA WWW dictionary of marketing terms. (n.d.). Retrieved June 4, 2005, from APA online glossary. (n.d.). Retrieved March 2005, from Assael, H. (1995). Consumer behavior and marketing behavior . Cincinnati, OH: South-Western College Publishing. Batra, R., Myers, J. G., & Aaker, D. A. (1996). Advertising management. (5th ed.). Upper Saddle Rive r, NJ: Prentice Hall. Belk, R.W. (1988). Possessions and the extended self. Journal of Consumer Research, 15, 139. Birdwell, A.E. (1968). A study of influe nce of image congruence on consumer choice. Journal of Business, 41, 76 – 88. Bower, B. (1999). When stones come to life: Researchers p onder the curious human tendency to view all sorts of things as alive. Science News, 155, 360 – 362. Boyer, P. (1996). What makes anthropomo rphism natural: Intuitive ontology and cultural representations. Journal of Royal Anthropology Institute, 2, 83 – 97. Broniarczyk, S.M., & Alba, J.W. (1994). The importance of the brand in brand extensions. Journal of Marketing Research, 31, 214 – 228.

PAGE 106

94 Burke, B. (1994). Position, personalit y, not price, should frame consumer messages. Brandweek, 35, 36. Davis, J.J. (1997). Advertising research: Theory and practice . New Jersey: Prentice-Hall. Dolich, I.J. (1969). Congruence relationshi ps between self-images and product brands. Journal of Marketing Research, 6, 80. Erickson, M.K. (1996). Using self-congruity and ideal congruity to predict purchase intention: A european persp ective. Journal of Euro-Marketing, 6, 41. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research . Reading, MA: Addison-Wesley. Franken, R. (1994). Human motivation (3rd ed.). Pacific Grove, CA: Brooks/Cole. Gould, S.J. (1991). The self-manipulation of my pervasive perceived vital energy through product use: An introspective-praxis pers pective. Journal of Consumer Research, 18, (September), 194-207. Graeff, T.R. (1996). Using promotional messages to manage the effects of brand and self-image on brand evaluati ons. Journal of Consumer Marketing, 13, 4. Grubb, E.L., & Grathwhohl, H.L. (1967). Consumer self-concept, symbolism, and market behavior: A theoretical approach. Journal of Marketing, 31, 22. Grubb, E.l., & Hupp, G. (1968). Perception of self, generalized stereotypes, and brand selection. Journal of Marketing Research, 5, 58. Hair, J.F., Anderson, R. E., Tatham, R. L ., & Black, W.C. (1998). Multivariate data analysis . (5th ed). Upper Saddle River, NJ: Prentice Hall. Hastak, M., & Olson, J.C. (1989). Assessing the role of brand-related cognitive responses as mediators of communicati on effects on cogniti ve structure. Journal of Consumer Research, 15 , (March), 444. Hogg, M.K., Alastair, J.C., & Keeling, K. (2000). The impact of self-monitoring on image congruence and product/brand evaluation. European Journal of Marketing, 34, 5/6, 641. Hong, J.W., & Zinkhan, G.M. (1995). Self-conc ept and advertising effectiveness: the influence of congruency, conspic uousness, and response mode. Psychology and Marketing, 12, 53.

PAGE 107

95 Kapferer, J.N. (1992). Stra tegic brand management. New York, NY: The Free Press. Kardes, F. R. (1999). Consumer beha vior and managerial decision making . Reading, MA: Addison-Wesley. Kassarjian, H.H. (1971). Personality and consumer behavior: A review. Journal of Marketing Research, 8, (November), 409. Katz, D. (1960). The functional approach to the study of attitudes. Public Opinion Quarterly, 24, 163. Keller, K.L. (1993). Conceptualizing, m easuring, and managing customer-based brand equity. Journa l of Marketing, 57, 1, 1. Keller, K.L., & Aaker, D.A. (1992). The e ffect of sequential introduction on brand extensions. Journal of Marketing Research, 29, (February), 35. Landon, L.E. (1974). Self concept, ideal se lf concept, and consumer purchase intentions. Journal of Consumer Research, 1, 44. Levy, S.J. (1959). Symbols for sa le. Harvard Business Review, 37, 117. MacKenzie, S.B., Lutz, R.J., & Park, W.P. (1989). An empirical examination of the structural antecedents of attitude to ward the ad in an advertising pretesting context. Journal of Marketing. 53, 48. Malhotra, N.K. (1981). A scale to measure self-concepts, person concepts and product concepts . Journal of Marketing Research, 23 , (November), 456. Martin, W.S., & Bellizzi, J. (1982). An anal ysis of congruous relationships between self-images and product im ages. Journal of the Academy of Marketing Science, 10, 4, 473. McCrae, R.R., & Costa, P.T. (1985). Comparison of EPI and psychoticism scales with measures of the five-factor model of personality. Personality and Individual Differences, 6, 5, 587. McCrae, R.R., & Costa, P.T. (1987). Valid ation of the five-factor model of personality across instruments and observe rs. Journal of Personality and Social Psychology, 52, 1, 81. McCraken, G.S. (1986). Culture and consum ption: A theoretical account of the structure and movement of cultural meaning of consumer goods. Journal of Consumer Research, 13, (June), 71.

PAGE 108

96 Mitchell, A.A., & Olson, J.C. (1981). Are product beliefs the only mediator of advertising effect on brand attitudes? Journal of Marketing Research, 18, (August), 318. Ogilvy, D. (1963). Confessions of an advertising man . New York, NY: Ballantine Books. Ogilvy, D. (1983). Ogilvy on advertising. New York, NY: Vintage Books. Osgood C.E., Suci, G.J., & Tannenbaum, P.H. (1957). The measurement of meaning. Urbana, IL: University of Illinois Press. Parente, D., Bergh, V., Barban, A., & Marra, J. (1996). Advertising campaign strategy: A guide to marketing communication plans . Fort Worth, TX: The Dryden Press. Park, C.W., Jaworski, B.J., & Macinnis, D.J. (1986). Strategic brand concept-image management. Journal of Marketing, 50, (October), 135. Patterson, M. (1999). Re-appraising the c oncept of brand image. Journal of Brand Management, 6, 409. Phau, I., & Lau, K.C. (2001). Brand persona lity and consumer self-expression: Single or dual carriageway? J ournal of Brand Management, 8, 6, 428. Plummer, J.T. (1985). How personality makes a difference. Journal of Advertising Research, 24, 6, 27. Plummer, J.T. (2000). How personality makes a difference. Journal of Advertising Research, Nov-Dec , 79. Rogers, C.R. (1959). A theory of therapy, pers onality, and interpersona l relationships, as developed in the client-centered framew ork. In S. Koch (ed.), Psychology: A study of a science. (Vol. 3, pp. 184-256). New Yo rk, NY: McGraw-Hill. Ross, I. (1971). Self-concept and brand preference. Journal of Business of the University of Chicago, 44, 38. Rychlak, J.F. (1981). Introduction to personality and psychotherapy . Boston, MA: Houghton Mifflin Company. Schultz D.E., & Barnes, B.E. (1999). St rategic brand communication campaigns. Lincolnwood, IL: NTC Business Books.

PAGE 109

97 Shocker, A.D., & Srinivasan, V. (1979) . Multiattribute appr oaches for product concept evaluation and generation: A cr itical review. Jour nal of Marketing Research, 16, (May), 159. Sirgy, M.J. (1982). Self-Concept in cons umer behavior: A critical review. Journal of Consumer Research, 9, 287. Sirgy, M.J. (1986). Self-congruity: Toward a theory of personality and cybernetics . New York, NY: Praeger Publishers. Sirgy. M.J., Grewal, D., Mangelburg, T., Park, J., Chon, K., & Claiborne, C.B. (1997). Assessing the predictive validity of two methods of measuring self-image congruence. Academy of Marketing Science, Journal, 25, 3, 229. Sirgy, M.J., & Su, C. (2000). Destination imag e, self-congruity, and travel behavior: toward an integrative model. Journal of Travel Research, 38, 2, 340. Solomon, M.R. (1983). The role of products in social stimuli: A symbolic interactionism perspective. J ournal of Consumer Behavior, 10, 319. Spears, N., & Singh, S. (2004). Measuring at titude toward the brand and purchase intentions. Journal of Current Issu es and Research in Advertising, 26, 2, 53. Sutherland, J., Marshall, S., & Parker, B.T. (2004). Real, ideal, and undesired self concepts and their effects on viewer preferences: Who do you love? Paper Presented at the American Academy of Advertising , Baton Rouge, LA, March. Underwood, R., Bond, E., & Baer, R. (2001). Building service brands via social identity: Lessons from the sports market place. Journal of Marketing Theory and Practice, 9, 1, 1. Wee, T.T. (2004). Extending human pers onality to brands: The stability factor. Journal of Brand Management, 11, 4, 317. Wilkie, W.L. (1994). Consumer behavior (3rd ed.). New York, NY: Wiley Wylie, R.C. (1975). The self concept: A cr itical survey of pertinent research literature . (2nd ed.). Lincoln, NE: University of Nebraska Press. Ziegler, J.D., & Hjelle, L.A. (1992). Pe rsonality theories: Basic assumptions, research, and applications . New York, NY: McGraw-Hill Co. Zinkham, G.M., & Hong, J.W. (1991). Self-concept and advertising effectiveness: A conceptual model of c ongruency, conspicuousness, and response mode. Advances in Consumer Research, 18, 348.

PAGE 110

98 BIOGRAPHICAL SKETCH Brian Parker, presently an assistant profe ssor at Florida International University, has extensive field experience as a research analyst, media rela tions coordinator and production coordinator. With an academic background in psychology and mass communications, he earned a doc toral degree in mass communications at the University of Florida. His honors include membership in Phi Kappa Phi National Honor Society and Kappa Tau Alpha National Honor Society in Journalism and Ma ss Communications. His expertise and research interest ar e consumer behavior and branding.