Individual Differences in the Centrality of Visual Produact Aesthethic

8/12/2019 Individual Differences in the Centrality of Visual Produact Aesthethic

1/16


2/16

551

2003 by JOURNAL OF CONSUMER RESEARCH, Inc. Vol. 29 March 2003

All rights reserved. 0093-5301/2003/2904-0008$10.00

Individual Differences in the Centrality of VisualProduct Aesthetics: Concept and Measurement

PETER H. BLOCHFREDERIC F. BRUNELTODD J. ARNOLD*

This research conceptualizes and develops a scale to measure individual differ-ences in the centrality of visual product aesthetics (CVPA), defined as the levelof significance that visual aesthetics hold for a particular consumer in his/her re-lationship with products. Three related dimensions of product aesthetics centralityemerged from the research: value, acumen, and response intensity. A series ofeight studies provided evidence that the CVPA measure possesses satisfactoryreliability and validity. Additionally, this research illuminates important differencesbetween high and low CVPA consumers in product-design-related evaluations and

behaviors and provides suggestions for future research employing the scale.

The capacity of advanced aesthetic experienceis not common to all. . . . Both the capacityfor it and the interest in developing it are veryunevenly distributed. [Osborne1986, p. 119]

Most products today rarely break and tend to do whatthey promise. Therefore it is not surprising then thatconsumers increasingly make brand choices based on aes-thetic value and distinctiveness of visual design (Dumaine

1991; Schmitt and Simonson 1997). Visual aesthetics alsoare growing in prominence for an ever-wider selection ofproducts. Vegetable peelers, wireless phones, car-washingbuckets, and lawn tractors are all being designed with at-tention to the aesthetic value of their appearance. Althoughattempts to produce goods with attractive forms are nothingnew, we are today seeing a widespread emphasis on productdesign unmatched since the art deco era of the 1930s (Loewy1951; Nussbaum 1999).

Product design is a broad term that includes a considerablerange of engineering-related attributes such as ergonomics,

*Peter H. Bloch is professor of marketing at the College of Business,University of Missouri, Columbia, 65211 ([email protected]). Frederic

F. Brunel is assistant professor of marketing, School of Management, Bos-ton University, Boston, 02215 ([email protected]). Todd J. Arnold is assistantprofessor of marketing at Washington State University, College of Businessand Economics, 2710 University Drive, Richland, 99352 ([email protected]). This research was partially funded by the Research Coun-cil and College of Business Summer Research Program at the Universityof Missouri and the Boston University, School of Management JuniorFaculty Development and Research Fund. The authors wish to thank theeditor, the associate editor, and the three reviewers for providing invaluablefeedback on previous versions of this article. The authors also wish toacknowledge the support of the Corporate Design Foundation and theresearch assistance of Matthew Filosa, Paul Berger, and C. B. Bhattacharya.

production efficiency, strength, recyclability, and distributionease, as well as aesthetics (Bloch 1995; Davis 1987). Whilenot minimizing the importance of these other design elements,the scope of this article is limited to visual product aestheticsor those characteristics that create a products appearance,such as materials, proportion, color, ornamentation, shape,size, and reflectivity (Lawson 1983). Although product aes-thetics may include a range of important nonvisual elements,as in the case of music, foods, or fabrics, the focus here will

be limited to visual aesthetics because of marketplace prom-inence and relevance to the widest assortment of goods. Inaddition, inferences about other sensory experiences are oftenbased on product appearance. For example, the look of aleather car seat may suggest softness and the dark color of achocolate may indicate a semisweet taste.

Visual aesthetics influence consumers perceptions in sev-eral ways. First, superior designs distinguish products fromcompetitors and help gain recognition in a crowded mar-ketplace (Bloch 1995; Dumaine 1991; Schmitt and Simon-son 1997). Visual aesthetics also have a symbolic functionthat influences how a product is comprehended and eval-uated. Images of elegance, ease of use, youthfulness, du-rability, and innovativeness all may stem from choices mar-

keters make in developing the appearances of new products(Forty 1986). Finally, product appearance represents thecentral channel for the formation of consumer/productrelationships (Hollins and Pugh 1990). It is the first thingabout a product that connects with a potential buyer, andregardless of product class, judgments follow from this sen-sory experience.

The twentieth centurys most well known industrial de-signer, Raymond Loewy (1951), recognized that some con-sumer segments are more design oriented than others when


3/16

552 JOURNAL OF CONSUMER RESEARCH

he differentiated between users and aesthetes. It is importantto understand and measure these individual differences re-lating to design for several reasons. First, individual differ-ences in responsiveness to visual aesthetics may underlie anumber of other well-established consumer behavior vari-ables such as product involvement, brand loyalty, materi-

alism, innovativeness, self-image congruence, choice, andusage behavior. For example, one could test whether in-novators possess high aesthetics centrality due to desires forthe most visually advanced and distinctive products avail-able to purchase. Building on the work by Fournier (1998),one could also speculate that high aesthetics centrality un-derlies consumer relationships with and loyalty to marketersknown for superior design executions (e.g., Apple, Braun,Target, and Bang & Olafsun).

Explication of the aesthetics centrality concept is poten-tially important in understanding consumer decision pro-cesses. In particular, aesthetics centrality may determine themanner in which product aesthetics are evaluated and usedin arriving at a purchase decision. Using an ElaborationLikelihood framework, high centrality levels should lead tothe processing of product aesthetics in a systematic andelaborate fashion. Low centrality would be associated withperipheral processing. This is an important considerationbecause central/systematic processing is cognitively basedand produces more enduring outcomes than does peripheralprocessing (Eagly and Chaiken 1993). Studies of visual aes-thetics centrality also may enhance our understanding of theprice consumers are willing to pay for a product. Lichten-stein, Bloch, and Black (1988) found that product involve-ment served as a countervailing force to price consciousnessin explaining acceptable price levels. It is expected thataesthetics centrality may act in a similar fashion, but across

a range of product categories.Finally, measuring the centrality of visual aesthetics willallow marketers to identify receptive segments and makemore efficient resource commitments to design. Investmentin design can be better matched to the anticipated targetmarket to avoid either over- or underspending on design.Sales forecasts also might become more accurate if the cen-trality of product aesthetics to a market can be assessed andconsidered in light of the design characteristics of a newproduct. Understanding of and the ability to assess productaesthetics centrality will also allow researchers to achievegreater insight into differences in product reactions that mayexist across cultures, nations, and time (e.g., Galbraith, Sig-uaw, and Lim 1995). This can guide marketers in more

effectively modifying goods for export and in response tochanges in the product life cycle.

The present research has two objectives: first, to providea conceptual foundation for understanding the centrality ofvisual product aesthetics and, second, to develop a scale tomeasure differences in this centrality across consumers. Theliterature concerning individual differences in the centralityof aesthetic characteristics is first reviewed and linked to aconceptualization of the centrality variable. This is followedby a report of eight studies that develop a scale measuring

this construct. Finally, the article discusses the implicationsof the results for future research and for design decisionmaking by marketers.

INDIVIDUAL DIFFERENCES IN THE

CENTRALITY OF VISUAL PRODUCTAESTHETICS

For our purposes, the centrality of visual product aes-thetics (CVPA) is defined as the overall level of significancethat visual aesthetics hold for a particular consumer in his/her relationships with products. The CVPA represents a con-tinuous individual difference variable that may range fromnear zero to very high levels where visual aesthetics dom-inate a consumers acquisition and usage of goods. TheCVPA is also considered to be a general consumer trait.That is, consumers exhibiting higher CVPA are expected tohave greater than average concern for visual aesthetics in-dependent of product category or setting. The CVPA has

the potential to affect the weight that visual aesthetics holdin a purchase decision as well as preferences for brands andproducts that satisfy aesthetic needs. In nonpurchase con-texts, CVPA also may influence usage levels, duration ofownership, display, word of mouth, and product mainte-nance activity of consumers. Note that CVPA captures thegeneral significance or importance of visual product aes-thetics rather than preferences for or attitudes toward a par-ticular aesthetic style. High CVPA may be associated withpositive reactions to products that vary in their amount ofornamentation and overt styling cues.

Based on past research and our theorizing, CVPA en-compasses four related facets or dimensions: (1) the valuea consumer assigns to product appearances in enhancing

personal and even societal well being, (2) acumen, or theability to recognize, categorize, or evaluate product designs,(3) the level of response to visual design aspects of products,and (4) the determinancy of visual aesthetics in affectingproduct preferences and purchase satisfaction. These fourfacets of CVPA are described in more detail below.

Personal and Social Value of Design

One element comprising CVPA is the perceived value ofvisual product aesthetics as a means of enhancing qualityof life, both personally and for society in general. On anindividual level, consumers with high CVPA believe thatencounters with beautiful objects positively influence the

quality of their daily lives or allow them to satisfy higherlevel needs (Yalch and Brunel 1996). The rewards they re-ceive from the aesthetic properties of products are recog-nized and held dear. Such consumers may define themselvesin part by the value that design plays in their life, and theymay see themselves as connoisseurs who get substantialbenefit from owning beautiful objects (Csikszentmihalyi andRobinson 1990).

The value dimension in part captures the tendency forbeautiful objects to be deemed sacred by consumers.


4/16

PRODUCT AESTHETICS CENTRALITY SCALE 553

Belk, Wallendorf, and Sherry (1989) introduced the con-cept of sacred possessions where particular objects arehighly valued by a consumer and are treated in a reveren-tial manner as extensions of the self. These authors notedthat one way an object may become sacred is throughquintessence (see Cornfield and Edwards 1983). Quintes-

sential objects have variously been described as unique,magically desirable, wonderful, authentic, and unequivo-cally right. Sexton (1981) profiled and photographed a setof 130 products that are frequently described as quintes-sential such as Chris-Craft wooden speedboats and Eameschairs. He attributed their quintessence to superior designattributes. It may be argued that high CVPA consumerssacralize products based on their superior designs or quin-tessence.

Consumers with high visual aesthetics centrality are alsolikely to believe that fine design is valuable to society gen-erally and to believe that the quality of life for everyoneis affected by the quality of the designed environment.Charles, Prince of Wales, may be considered to exemplify

high CVPA when he passionately comments on the negativeimpact poor architectural design has on British society(Charles 1989). Individual differences in support for publicworks of art also may be explained by variations in thissocial value dimension.

Acumen

Another proposed facet of CVPA is that of acumen (Bloch1995). Acumen reflects an ability to recognize, categorize,and evaluate product designs and is expected to vary withina population (Osborne 1986). Berlyne (1971) argued thatsome people are endowed with more aesthetic taste than

others. In an early writing on the subject, Ortega Y Gasset(1925) noted that art could only be comprehended by peoplepossessed of the peculiar gift of artistic sensibility. Csiksz-entmihalyi and Robinson (1990) extensively discussed whatthey call the good eye or a gift for analyzing the visualarts. They argued most people in our culture are not awareof the range and intensity of the enjoyable experiences avail-able to them through the sense of vision (p. 2). The authorsexplain this shortcoming by noting that some people arevisual and some peoples abilities lie in other forms of in-tellectual pursuit (p. 125). In a similar vein, several con-sumer researchers (Childers, Houston, and Heckler 1985;Holbrook 1986) have indicated that some consumers favorvisual over verbal processing and that highly visual indi-

viduals may give greater weight to aesthetic elements inmaking product choices than do less visual processors.

Level of Response

Checkbook in pocket, we went back to thecanoe store ready to buy. But then I acciden-tally fastened my eye on an irresistible objecthanging from one of the display racks. It wasa brand new, but timelessly traditional, all-

wood canoe with a canvas hull covering. . . .I almost got dizzy looking at it. [Egan2000]

Throughout human history, beautiful objects have had thecapacity to generate significant responses among consumers.Ingarden (1983) theorized that reactions to the design of anobject vary from practical to the aesthetic, with the latterhaving the potential to be quite strong. In Rooks (1987)work on impulse buying, he indicated that impulse buyingfrequently involves products that have strong aesthetic orstyling elements and that stimulus characteristics may pro-duce this type of buying. Thus, individual differences inlevel of responses to design aesthetics may be one under-pinning of impulse buying. Responses to design aestheticscan take several forms.

A number of scholars have described the affective reactionspeople have to the design aesthetics of objects (Bloch 1995;Veryzer 1993). Csikszentmihalyi and Robinson (1990) char-acterized aesthetic responses as a state of intense enjoymentcharacterized by feelings of personal wholeness, a sense of

discovery, and a sense of human connectedness (p. 178).The literature has also suggested an attributional processwhere initial reactions to design are primarily emotional withcognitions following in an attempt to analyze ones feelings(Csikszentmihalyi and Robinson 1990; Durgee 1988). Thereare also behavioral responses to product designs such as mov-ing closer to the object, extended viewing of its appearance,touching of its surface, and ultimately its acquisition (Bloch1995; Csikszentmihalyi and Robinson 1990).

Responses to the design aesthetics of objects have valenceas well as intensity. Certain forms or designs will generatepositive responses in a particular consumer while othersevoke negative reactions. An unappealing design may en-gender feelings of dislike, critical arguments as to the prob-

lems with the design and even disgust.

Design Determinancy

The fourth component of CVPA concerns the extent towhich product aesthetics are used as evaluation criteria. Con-sumers with high visual aesthetics centrality should preferproducts with superior designs and admit to the influenceof design on forming product and brand preferences. Visualappearance is actively considered in comparing products andis a key determinant of purchase satisfaction.

For some consumers, design quality may not be consid-ered at all, is not recognized, or is simply deemed unim-portant. For others, there may be some interest in product

appearance, but these individuals may not be willing to tradeother product benefits such as low prices to obtain aestheticimprovements. With high levels of CVPA, however, con-sumers not only use visual appearance in comparing prod-ucts, but also are willing to expend resources in order toobtain fine design. These resources may be financial or tem-poral in the form of effort to locate and obtain goods withsuperior design aesthetics. These product-related aesthetesperceive a relatively high payback in purchase satisfactionin return for any additional cost they may incur.


5/16


Measurement of Individual Differences in Design-Related Traits

Given the existence of individual differences in CVPAand their relevance to a variety of consumer behavior con-structs, researchers need a means to assess these differences.

Although the literature contains no instrument specificallydeveloped to assess the centrality of any type of commercialaesthetics, there have been attempts to assess related vari-ables. For example, there are a number of tests used tomeasure art appreciation. Most of the research activity alongthese lines occurred between 1920 and 1940. Examples in-clude Thorndikes Test for Aesthetic Appreciation (1916)and the Child Test of Esthetic Sensitivity (1964). More re-cently, tests have been proposed by Goetz et al. (1979) andBamossy, Johnston, and Parsons (1985).

The basic format of these tests is to show subjects variousworks of art and then examine the overlap between asubjects preferences and those of established art experts.The level of correspondence with external standards of aes-

thetic value becomes a surrogate for art appreciation. Psy-chometric testing has been spotty for these instruments withvalidation typically focusing on the ability to discriminatebetween artists and nonartists. Although art-appreciationirmeasures bear some relevance to the design-acumen aspectof CVPA, the unwieldy format of these tests and their focuson famous works of art have meant limited usefulness inconsumer research.

Other instruments that bear some relationship to CVPAtake an information-processing approach by differentiatingpersons based on their propensity to employ verbal versusvisual styles of processing (Childers et al. 1985; Holbrook1986). For purposes here, one might expect that visual stylesof processing may be related to the centrality of visual aes-

thetics, and this idea is tested below in the empirical vali-dation component of the CVPA measure development.

Extant measures just noted offer important but tangentialinsights into the concept of visual aesthetics centrality. Toassess CVPA as conceptualized here, and to meet the needfor a measure that can be used to advance our understandingof product reactions, a scale-development project was un-dertaken. The empirical studies undertaken in the creationand psychometric testing of an instrument to assess CVPAare described next.

MULTI-STUDY SCALE-DEVELOPMENT

PROGRAM

The focus of the studies reported below was to developa reliable and valid scale to measure CVPA. It was also theintention to provide a measure that assessed such centralityin general, rather than for a single product class.

Study 1: Item Generation and Refinement

In order to generate the initial pool of items for the CVPAscale, two sources of information were employed: a reviewof relevant literature and depth interviews of individuals

believed to have above-average interest in design aesthetics.Interview subjects included three graduate students in designand one principal in a major product design firm. The in-formants commented on their reactions to product appear-ances, how they evaluate product designs, and the role thatwell-designed objects play in their lives. Each interview

lasted approximately one and one-half hours. Recurrentthemes discovered among the responses along with conceptsemerging in the literature were used in the conceptual ma-terial presented above and an initial pool of 86 scale items.

The initial 86 items followed the preliminary four-di-mensional conceptualization of CVPA: value (23 items),acumen (28 items), response (17 items), and determinancy(18 items). Although dimensions are specified, it is our con-tention that they occur together and can be combined toform an overall measure of CVPA. Subsequent to the con-ceptualization of these scale dimensions, a group of fivemarketing faculty and Ph.D. students at a large midwesternuniversity judged the content validity of the items. Each

judge was presented with a written definition of CVPA and

each of its dimensions coupled with an example scale item.Judges were then asked to assign each of the 86 items to adimension or to a none of these category, following meth-ods used by Bearden, Netemeyer, and Teel (1989) and Bear-den, Hardesty, and Rose (2001). Items that did not receiveconsistent classification by at least four of the five judgeswere eliminated. This analysis resulted in 52 remainingitems.

Next, similar to a procedure used by Bearden et al. (1989)and Zaichkowsky (1985), four more judges (faculty differentfrom the above respondents) were given a definition for eachdimension and asked to rate each statement as being clearlyrepresentative, somewhat representative, or not representa-tive of the dimension. Items evaluated as clearly represen-

tative by three judges and no worse than somewhat repre-sentative by a fourth judge were retained. This processresulted in a set of 34 items for further analysis.

Study 2: Item Reduction and Reliability Testing

This stage of the research involved scale purification andreliability testing for the 34 items. Data were collected viaa mail survey of a random sample drawn from current driv-ers license registrations in an industrialized midwesternstate. An initial mailing of 1,050 surveys was followed bya reminder letter and second copy of the questionnairemailed two weeks later. Usable responses were receivedfrom 318 individuals, resulting in an overall response rate

of 30%. Further, per Armstrong and Overton (1977), studyrespondents were divided into early (first tercile) versus late(last tercile) respondents to examine nonresponse bias. MeanCVPA scores for early (3.51) and late (3.48) respondentswere not statistically different, indicating nonresponse biasis not an issue with this data collection. Descriptively,slightly over half of the respondents were female (55%),with an average annual household income of between$35,000 and $50,000. The typical respondent had takensome college courses, with only 24% having a college or


6/16


TABLE 1

CONFIRMATORY FACTOR ANALYSIS MODEL FITCOMPARISONS

Model Chi-square df Chi-square differencea

Null 1,677.00 55 . . .1-factor 98.13b 44 1,578.87**3-factor uncorrelated 318.25 44 220.12**3-factor correlated 273.10 43 45.15**1-factor higher order 97.11b 43 175.99**

CFI NNFI NFI

1-factor .96 .95 .931-factor higher order .96 .94 .93

NOTE.dfis degrees of freedom, CFI is ComparativeFit Index, NNFI is Non-Normed Fit Index, and NFI is Normed Fit Index.

aChi-square differences represent comparisons of subsequent models (e.g.,null vs. 1-factor, 1-factor vs. 3-factor uncorrelated, etc.).

bThe chi-square difference between the 1-factor and 1-factor higher ordermodels is not significant

**p ! .01.

postgraduate degree. These demographic data follow veryclosely the 2000 Census norms for the state.

All items were scored using a five-point Likert formatwith higher scores representing greater levels of CVPA.Prior to data analysis, the Kaiser-Meyer-Olkin (KMO) testof sampling adequacy and the Bartlett test of sphericity were

used to determine the appropriateness of factor analysis(Kaiser 1974). The KMO level of .924 and the significanceof the Bartlett test indicated that factor analysis was appro-priate for the data (Kaiser 1974). As an initial reductionmethod, the correlation of each item with the score for thetotal set of 34 items was computed. Items that did not havecorrected item-total correlations above .40 were deleted.This resulted in the deletion of seven items. Exploratoryprincipal components factor analysis (varimax rotation) wasperformed next on the remaining set of 27 items. Ten itemsfailed to exhibit simple structure on any factor and weredeleted. Corrected item-total correlations were again cal-culated leading to the deletion of one additional item. Fol-low-up factor analyses on the remaining items suggested a

15-item scale with three dimensions.The first factor captured the value facet of visual aesthetics

centrality, the second captured acumen, while the third cap-tured response intensity. The determinancy dimension failedto materialize. Its constituent items were either deleted basedon criteria noted previously or loaded on another of the threedimensions. It is likely that the other three dimensions ef-fectively subsumed determinancy. That is, if one stronglyvalues product aesthetics and has a strong response to beau-tiful objects, it is likely that design will be an importantpurchase determinant. The exploratory results seemed rea-sonable and parsimonious, resulting in the introduction ofthe 15 remaining items to further structural testing throughconfirmatory factor analysis.

Study 3: Confirmatory Factor Analysis

A second mail survey was conducted to generate data forconfirmatory factor analysis. Once again, a survey wasmailed to a random sample of adult consumers living in amidwestern state. An initial mailing of 520 surveys wasfollowed by a reminder letter and second copy of the ques-tionnaire mailed two weeks later. There were 136 usableresponses, resulting in an overall response rate of 26%. Thedemographics of the second sample were similar to thoseachieved in the first mailing. As with the first mailing, non-response bias did not appear to be a concern with the secondmailing.

A series of confirmatory models was examined and es-timated using EQS (Bentler 1992). Examination of itemreliabilities, Lagrangian multiplier tests, and tests of discri-minant validity from confirmatory factor analysis modelsfor both a one-factor second-order model (with three sub-dimensions) and a three-factor correlated model suggestedthe deletion of four additional items. In addition, concernsregarding high correlations among the three dimensions(i.e., value, acumen, and response intensity) prompted theexamination of several alternative factor structures. The

models estimated included: a null model, a one-factormodel for which all of the 11 remaining items were forcedto load on a single factor, a three-factor uncorrelated model,a three-factor correlated model, and a one-factor second-order model with three subdimensions.

As indicated in table 1, the one-factor second-order model

and the unidimensional model provided the best fit of thedata. The chi-square statistic of both models representedsignificant improvement over any of the competing mod-els. The chi-squares of the two models themselves, how-ever, were not significantly different. This, in combinationwith the high correlation among the second-order subdi-mensions (value and ; value and responseacumen p .93

; acumen and response ),intensity p .87 intensity p .88suggested that the appropriate representation of the CVPAscale is a unidimensional model. Such a conclusion wassubstantiated through replication of the confirmatory anal-ysis with the data collected in study 2.

In addition, the unidimensional model demonstratedsatisfactory fit of the data. The fit statistics were

, 41 degrees of freedom, ; Com-chi-square p 97.11 p !.01parative Fit Index ; Non-Normed Fit Index(CFI) p .96

; and Normed Fit Index . Al-(NNFI) p .95 (NFI) p .93though the chi-square statistic was significant, it was underthe limit of three times the number of degrees of freedomsuggested by Bollen (1989). The internal consistency of theitems, as determined through the sum of the loadings squareddivided by the sum of the loadings squared added to thesum of the item variance, was .89, indicating acceptablereliability of the scale (Fornell and Larcker 1981). Similarly,the corresponding coefficient alpha estimate of internal con-sistency reliability was .89. The 11 remaining items andtheir corresponding factor loadings are presented in table 2.The subdimension to which each item originally belonged

is retained to demonstrate the equality of representationamong the theoretical components of design aesthetics cen-trality.


7/16


TABLE 2

CONFIRMATORY FACTOR ANALYSIS RESULTS FOR CVPAITEMS

ItemSinglefactor

Value:Owning products that have superior designs makes

me feel good about myself. .77I enjoy seeing displays of products that have supe-

rior designs. .73A products design is a source of pleasure for me. .66Beautiful product designs make our world a better

place to live. .64Acumen:

Being able to see subtle differences in product de-signs is one skill that I have developed over time. .71

I see things in a products design that other peopletend to pass over. .64

I have the ability to imagine how a product will fit inwith designs of other things I already own. .62

I have a pretty good idea of what makes one productlook better than its competitors. .55

Response:Sometimes the way a product looks seems to reach

out and grab me. .75If a products design really speaks to me, I feel that

I must buy it. .56When I see a product that has a really great design,

I feel a strong urge to buy it. .55

NOTE.Single-factor model chi-squareis 98.125 with44 degrees offreedom;Comparative Fit Index p .96; Non-Normed Fit Index p .95; Normed Fit Indexp .93; Internal Consistency p .89.

The results of the confirmatory analysis indicate thatCVPA is comprised conceptually of three dimensions thatare useful in terms of understanding the construct but that

are unidimensional in scope and, hence, measurement. Asdiscussed by Carver (1989), in such instances it is appro-priate to combine the items tapping the dimensions of CVPAto arrive at a composite score that represents the generalhigh/low nature of the construct. As such, the compositescore of CVPA represents an individual difference in thesalience of visual product aesthetic among consumers.

Further, these results are in concert with previous workon experiential consumption (Holbrook and Hirschman1982; Holt 1995). Holt posited that under conditions ofconsumption for the experience, the product is consumedas an end in itself, rather than as a means to another end.In the present research, high CVPA consumers will likelysee a product with a beautiful design as worthwhile to con-

sume as an end in its own right. Holt also identified threecomponents that are necessarily associated with consump-tion for the experience: accounting, evaluating, and appre-ciating. He defined accounting as having the ability to applyan interpretive framework to make sense of the focal prod-uct. This component is similar to the acumen element ofour scale. Evaluating is allied with our value element beingdefined by Holt as applying ones interpretive frameworkto pass judgment on the focal object. His last component,appreciating, is described as an emotional reaction to the

focal object, being analogous to our response facet. Thusthe results of our analyses suggest that these three scalefacets logically occur together, with each individual elementbeing a necessary but not sufficient means to experientialconsumption and the centrality of visual product aesthetics(see Buchanan and Henderson 1992).

Study 4: Discriminant Validity and Relationshipswith Other Measures

For the purpose of assessing discriminant validity, a sur-vey containing the CVPA scale, and measures of four con-ceptually related constructs, was administered to 108 un-dergraduate business students (Campbell and Fiske 1959).Evidence of discriminant validity was provided through thetest recommended by Fornell and Larcker (1981), andthrough examination of chi-square difference tests betweenpairwise comparisons of the various scales (Anderson andGerbing 1988).

The first related measure employed here was the visualcomponent of Childers et al.s (1985) Style of Processing(SOP) scale, which captures an individuals propensity touse the visual modality of processing. Because individualswho score highly on CVPA may be more likely to processvisually, it was deemed appropriate to compare the twomeasures for purposes of establishing discriminant valid-ity. Second, the CVPA measure was assessed relative tothe Desire for Unique Consumer Products (DUCP) scale(Lynn and Harris 1997). The DUCP scale measures theextent to which consumers hold as a personal goal theacquisition of products that few others possess. One mightexpect consumers who score high on the CVPA scalewould be more likely to seek products that are unique, if

only for the appearance of the product. Finally, the rela-tionship between CVPA and consumer materialism wasinvestigated. Richins and Dawson (1992) argued that ma-terialism is comprised of three dimensions: acquisitioncentrality, acquisition as the pursuit of happiness, andpossession-defined success. It seems plausible that for con-sumers who score high on CVPA, the acquisition of aes-thetically pleasing products may become central to theirlives. In a similar manner, the high value of design aes-thetics for high CVPA consumers could produce a mind-set where feelings of satisfaction and well-being are notpossible without being surrounded by beautiful products.As indicated in table 3, each of the measures investigatedin combination with CVPA demonstrated an acceptable

level of reliability, with internal consistency scores rang-ing from .74 (DUCP) to .80 (success component of ma-terialism).

Table 3 also shows that CVPA is, as anticipated, positivelycorrelated with each of the measures discussed. Each cor-relation (phi estimate) between CVPA and the other re-spective constructs was significant at the .01 level. This isnot surprising, given the strong product-oriented com-monalities of the measures. Given that several of the cor-relations exceeded .40, further analysis was conducted to


8/16


TABLE 3

CORRELATIONS AND RELIABILITY ESTIMATES FOR KEY VARIABLES

CVPAVisual

processing DUCP

Materialism

CentralitySuccess Happiness

CVPA . . .Visual processing .34** . . .DUCP .53** .14 . . .Materialism:

Success .58** .14 .34** . . .Happiness .34** .11 .32** .52** . . .Centrality .50** .02 .47** .55** .24* . . .

Internal consistencya .89 .78 .74 .80 .79 .76

NOTE.CVPA p centrality of visual product aesthetics scale; DUCP p Desire for Unique Consumer Products scale.aDetermined through two-factor model comparisons involving CVPA and each respective construct.*p ! .05.**p ! .01.

ensure that discrimination existed among the tested scales.Specifically, following the recommendations of Fornell andLarcker (1981), the variance-extracted estimates for eachpair of constructs were examined to ensure that they ex-ceeded the squared phi estimate between the constructs. Thevariance-extracted estimates ranged from a low of .34(DUCP) to a high of .47 (happiness component of materi-alism), indicating discrimination between CVPA and eachrespective scale. The variance-extracted estimate for CVPAwas .43. Finally, the discriminant validity of the factors wasassessed by testing whether, for every pair of factors, a two-factor model fit significantly better than a one-factor model(Anderson and Gerbing 1988). In all cases, the two-factormodel fits significantly better than the one-factor model withthe smallest chi-square difference of 74.15 ( ) wit-p !.01

nessed in relation to the DUCP scale. Thus, it was concludedthat the CVPA measure is clearly discriminated from theother scales examined in this study.

Data from study 3 were also used to gain further insightinto the construct validity of the CVPA scale as well as toallow further examination of the scale interrelationshipsnoted above. Respondents to the study 3 survey were pre-sented with a list of 30 products and asked to indicate thoseproducts for which design or appearance was very importantto them in making purchases.1 Respondents could select asmany or as few product categories as they wished. Accord-

1The product list was generated through an open-ended survey givenpreviously in a separate pretest to 104 undergraduate and graduate studentsat a large university in the northeastern U.S. Respondents were asked to

list up to six products for which product appearance was very importantto them in making choices and six products for which appearance was notat all important. This list was supplemented with products shown on nav-igation buttons at the Wal-Mart and QVC home-shopping Web sites. Theproducts used in the validation of the CVPA scale were: clothing/shoes,personal computers, lawn mowers, clothing accessories/belts, furniture,power tools, bedding/linens, small kitchen appliances, gardening tools,eyeglasses, large appliances, hand tools/hardware, watches/jewelry, auto-mobiles, vacuum cleaners, perfume/cosmetics, cameras, barbecue grills,dinnerware/china, televisions, space heaters/fans, toothbrushes, VCRs,pens/stationary, camping/outdoor gear, telephones/cell phones, bicycles,sporting goods, audio equipment, and cookware.

ingly, scores could range from zero to 30, depending on thenumber of products selected. It was presumed that higherscores on the CVPA scale would be associated with a largernumber of selected products. For validation purposes, thenumber of products indicated was regressed on the CVPAscale, Childers visual processing measure, and the DUCPscale in order to examine the unique contribution of eachconstruct to explaining designs usage as a purchase cri-terion.

As table 4 indicates, CVPA, as well as the other twomeasures were all significantly correlated with the numberof product classes indicated, although CVPA was the onlyconstruct statistically significant at the .01 level ( ,rp .29

), with SOP and DUCP being statistically significantp !.01at the .05 level. Results indicated that the instruments ex-

plained unique portions of the variance in the number ofproducts selected. That is, the tolerance scores for CVPA(.96), visual orientation (.93), and DUCP (.96), and the lowVariance Inflation Factor scores (1.039, 1.078, and 1.05,respectively) indicated that multicollinearity is not an issuesuggesting that the independent variables are contributingindependently to the prediction of the dependent variable(see Neter, Wassermann, and Kutner 1989). Overall, the find-ings provide preliminary evidence for the nomological va-lidity of the CVPA scale.

Study 5: Social Desirability Testing

Due to the recent media coverage given to product de-

sign aesthetics (e.g., Nussbaum 1999), it was consideredprudent to test the CVPA measure for susceptibility tosocial desirability bias. Social desirability was measuredwith the 20-item impression management (IM) scale de-veloped by Paulhus (1993). The IM scale and CVPA itemswere administered to 53 undergraduate and graduate stu-dents at a midwestern university. The correlation betweenthe 11-item CVPA and Paulhus measures was .07 (NS),suggesting that social desirability bias was not a significantproblem for the CVPA measure.


9/16


TABLE 4

TEST RESULTS WITH BREADTH OF USAGE OF DESIGN AS A PURCHASE CRITERION

Independent variable BetaCorrelation with number of

products indicated

Results with number of product categoriesselected as dependent variable:a

CVPA scale .31* .29*SOP/visual scale .27* .19**DUCP scale .26** .14**

NOTE.CVPA p centrality of visual product aesthetics; SOP p Style of Processing; DUCP p Desire for Unique ConsumerProducts.

aAdjustedR2 p .19,F(3, 142)p 6.19,p ! .001.*p ! .05.**p ! .01.

Study 6: Known-Groups Validation

When naturally existing groups should meaningfully dif-fer on a scale in question, a known-groups validation test

is appropriate (Lastovicka et al. 1999). To conduct such atest, a sample of design professionals taken from a subscriberlist for a design-related magazine (@Issue) was comparedto the random samples drawn from the first and secondgeneral mail survey administrations (i.e., studies 2 and 3,respectively). Based on the training, experience, and careerdemands typical of these individuals, it was expected thatdesign professionals would have high interest in design andconsider visual aesthetics to be highly central. Two hundredfifty design professionals received a mail survey similar tothat sent to the general random sample and a reminder sur-vey two weeks later. Responses were received from 62 in-dividuals, resulting in a response rate of 25%. The meanfor the 11-item CVPA scale was markedly higher for the

design professionals than for the general population in bothinstances (4.26 vs. 3.51 with study 2 data;2 4.26 vs. 3.44with study 3 data; both differences significant, ). Thisp !.01successful known groups test provides additional evidencefor the CVPA scales validity.

Study 7: Experimental Validation of the CVPAScale

As further validation, a study was designed to test ifCVPA levels influence reactions to product designs that dif-fer in aesthetic quality. While we expected a main effectwhere the superior design execution generates more positivereactions overall, our chief hypothesis is that high CVPA

consumers should be more discriminating than low CVPAconsumers in their reactions to designs with high versus lowaesthetic content, as stated in the following hypothesis:

H1: For high CVPA consumers, the difference betweenreactions to the low and high aesthetic productswould be greater than the difference between the

2Across all general population research applications of CVPA, the scalemean has ranged from a low of 3.31 to a high of 3.6. The standard deviationof the scale has ranged from a low of .59 to a high of .62.

corresponding reactions of low CVPA consumers.This difference will be observed for four types ofresponses: (a) aesthetic appraisal of the product,(b) attitude toward the product, (c) purchase in-

tention, and (d) price willing to be paid for theproduct.

To test the proposed hypotheses, a factorial design2# 2was used. The product design manipulations were within-subject (subjects saw color pictures of two design executionsfor the same product class: one with low aesthetic contentand one with high content; no product features informationwas given). The CVPA was a between-subjects factor. Theposition in which the products appeared was counterbal-anced. A within-subject design was necessary to obtain com-parative evaluative judgments and to simulate cross-overtesting procedures often used in real-world tests (Tapp

1998).One hundred and ninety undergraduate business studentsat a large northeastern university participated in this studyfor course credit. Toasters were selected as product stimulifor this study because of relevance to student subjects andvariations in design aesthetics without commensurate vari-ations in other product features. These last two reasons arecritical for our investigation as we wished to isolate theeffect of aesthetics from function. Based on qualitative pre-testing, two toaster designs were selected as representing ahigh and a low level of product design aesthetics (see ap-pendix). Both of the photos were digitally altered for con-sistency and to hide brand identification. Toaster A was ratedin the pretest as representing a high aesthetics design, and

toaster B as a low aesthetics design.Both products were featured on the same page (position

counterbalanced). After viewing the pictures, subjects com-pleted a set of seven-point product aesthetics appraisal itemstaken from Hirschman (1986). Subjects also indicated theiroverall attitude toward each design and their purchase in-tention. For each of these three reaction variables, meanswere computed across constituent items to produce an over-all score that could range from one to seven with higherscores representing more favorable reactions. Finally, prices


10/16


FIGURE 1

MODERATING EFFECT OF DESIGN AESTHETICS CENTRALITY ON PRODUCT DESIGN EVALUATIONS

willing to be paid for each product design were capturedthrough open-ended questions.

An overall CVPA score was computed for each subject.The mean CVPA score for this sample was 3.45 (SD p

). To test our hypotheses, the sample was divided into.469terciles by CVPA score and a series of two-way analysesof variance was conducted using a design. In these2# 2analyses, the product design manipulation (low vs. high aes-

thetics design) was a within-subject factor and CVPA (highp top tercile of scores on the total 11 item scale, Mp

; low p bottom tercile of scores, ) as a be-3.90 Mp 2.96tween-subjects factor. Confirming the product aesthetics ma-nipulation, we found that subjects perceived the two prod-ucts to be aesthetically different ( ;M p 3.20 M pLA HA

; , ). Further, and as ex-5.10 F(1, 134) p 110.23 p !.001pected, the higher aesthetics product led to more favorableproduct attitudes ( ; ;M p 4.60 M p 5.44 F(1, 134) pLA HA

, ), higher purchase intentions ( ;32.38 p !.001 M p 4.53LA

; , ) and willingnessM p 5.03 F(1, 133) p 5.99 p p .016HAto pay a higher price ( ; ;M p $24.08 M p $37.32LA HA

, ).3 We also found a significantF(1, 129) p 119.12 p !.001multivariate effect of product aesthetics across the four de-pendent measures that were just mentioned (F(4, 124) p

, ). Results are shown graphically in figure 1.66.26 p !.001As hypothesized, a multivariate test showed a signifi-

cant interaction effect of CVPA and product version with

the four dependent measures ( , ).F(4, 124) p 3.47 p p .01Univariate tests were also analyzed. First, we found thatthere was a significant interaction of product aestheticswith CVPA on aesthetic appraisals (panel A: F(1, 134) p

, ). Even though both low and high CVPA12.53 p p .001respondents evaluated the high aesthetics product as more

3For price, the degrees of freedom are different because four subjectshad either incomplete answers or aberrant answers ( or outlierprice p 0of more than 3 standard deviations).


11/16


12/16


FIGURE 2

COMPARISON OF CONJOINT DESIGN STANDARDIZEDWEIGHT ESTIMATES FOR LOW AND HIGH CENTRALITY OF

VISUAL PRODUCT AESTHETICS RESPONDENTS

mean CVPA was 3.59 ( ). A tercile split on scoresSD p .611for the 11-item CVPA scale was used to identify low (bottomthird) versus high (top third) CVPA groups.

The first step involved estimating the entire model, ir-respective of CVPA. A least square regression was run withpurchase intention as the dependent variable and all the

manipulated factors as predictors. The overall model wassignificant ( , ) and explainedF(7, 1,724) p 118.08 p !.00132.5% of the variance. Each dummy variable was significantas well. As expected, increased power ( ,b p .41 tp

, ), memory ( , , ) and20.55 p !.001 b p .17 tp 8.67 p !.001warranty ( , , ) contributed tob p .24 tp 12.26 p !.001higher purchase intentions, and increased prices (b $28 p

, , ; b , ,15 tp 613 p !.001 $35 p.26 tp10.59) contributed to lower purchase intentions. With re-p !.001

spect to aesthetic quality, the middle-level product elicitedslightly lower purchase intentions ( , ,b p .05 tp 2.06

) than did the low aesthetic quality option, somewhatp !.05contradicting our pretest results. However, the high aestheticquality product did generate an increase in purchase inten-

tions ( , , ).4b p .13 tp 5.28 p !.001The next step involved studying the weight of product

aesthetics in the formation of purchase intentions for lowversus high CVPA respondents. To do so, we ran a secondmodel where dummy terms were used to capture the dif-ferences in aesthetic weights for high versus low CVPArespondents across the three product designs. To decomposethese effects and to capture the six weights, we used a totalof five dummy variables. This new model was significant( , ) and explained 33.2% of F(10, 1,090) p 54.16 p !.001the variance. The result replicated the findings of our firstmodel but also showed that for all three of the product designexecutions, the importance placed on aesthetics in the for-mation of purchase intentions was greater for high CVPA

respondents than for low CVPA respondents (see fig. 2 forthe aesthetic weights for each groups). Specifically, therewas an increase of .103 ( , ) for the betatp 3.17 p !.005weight of the low aesthetic quality product option, .084( , ) for the medium product, and .082tp 2.58 p !.01( , ) for the high aesthetic quality product.tp 2.73 p !.001

These results provide additional evidence of the validityand predictive ability of the CVPA scale. Again, it shouldbe noted that these findings might be considered rather con-servative given the product stimuli used in the study.

DISCUSSION

Findings from ResearchMarketplace responses clearly indicate strong differences

in consumers concern with product appearance or design.

4On closer examination, it appears that one of the reasons that the apriori lower and middle aesthetic scales lead to results that are contraryto expectations is that, for the low CPVA respondents, there is not asignificant difference for the impact of aesthetics of these two products ontheir purchase intention ( , ) . However, and as expected,tp.862 p p .389high CPVA respondents are able to attribute different weights to eachdesign.

All consumers care about what they receive from a productduring its acquisition and use. The CVPA reflects an en-during concern with the aesthetic benefits provided by aproduct. However, these consumer differences have not beenwell illuminated in past research, nor has there been a meansto measure them. The intention of this research was to con-ceptualize and develop a multi-item scale that captures in-dividual differences in the centrality of visual product aes-thetics. The present research would not only fill gaps in ourunderstanding of consumer preferences and reactions to de-sign, but also has the potential to further our understandingof a number of seminal consumer behavior concepts suchas brand loyalty, involvement, materialism, and innovative-

ness.A series of eight studies, integrating a range of meth-

odologies, suggest that CVPA is both a theoretically andmanagerially relevant construct and that the resulting scalepossesses acceptable reliability and validity. In particular,validity was successfully demonstrated through the testingof convergence and discrimination (both within the scaleand among published scales), known-groups comparisons,experimental and conjoint procedures, and replicable nom-ological analysis. All of the studies undertaken here em-ployed recognized psychometric procedures and standardssimilar to those reported in other recent scale-developmentwork (e.g., Bearden et al. 1989, 2001; Lastovicka et al. 1999;Richins and Dawson 1992). This research has broadened

our understanding of aesthetics in consumer behavior gen-erally as well as bringing needed attention to individualdifferences in reactions to design.

Based on the conceptual and empirical material presentedearlier, CVPA emerged as a unidimensional construct thatis comprised of elements drawn from three domains thatoccur in concert: the perceived value attached to superiorproduct design executions, abilities in understanding andevaluating product design, and the level of response to prod-uct aesthetics. Given the results for the scale reported above,


13/16


consistent scores across the three dimensions are expected.However, it should be noted that relatively high levels ofoverall CVPA could derive from different patterns of re-sponse among the three dimensions. For example, a con-sumer may strongly value product aesthetics while not re-porting high levels of acumen.

High CVPA consumers consider aesthetics to be impor-tant for a wider range of product categories than do con-sumers scoring low on visual aesthetics centrality. Further,the importance placed on product aesthetics in the formationof purchase intentions is greater for high CVPA respondentsthan for those with low CVPA. It should be recognized,however, that the objective of this research was to exploreand measure only one aspect of consumer-related aesthetics:the visual appearance of products. We recognize that designand aesthetics are also important for other commercial el-ements and for other sensory channels. We selected visualproduct aesthetics for this research due to their significancein the marketplace and in contemporary discussions of de-sign. In addition, visual aesthetics are relevant to the widest

set of product classes. Whereas sound, scent, and taste maybe critical for a few products, visual appearance is a keyfeature for nearly all goods.

Research Directions

There are also several areas of research expansion re-garding CVPA that appear profitable. In subsequent study,researchers could extend the concept of aesthetic centralityto other sensory experiences. One could explore whetherconsumers who are concerned with the visual aesthetics ofproducts also care more than average about aesthetics re-lating to sound, smell, or feel. It might also be the case thatvisual aesthetics are unique as a source of consumer cen-

trality. Determining whether concern with visual productaesthetics is allied with the centrality of visual aestheticsgenerally represents another research direction of promise.For example, do consumers with high CVPA also tend tocare about the appearance of store interiors or Web pagelayouts? Answering this question may require attention tothe impact of selective product involvements and role per-formance demands. For example, the male canoe buyerquoted earlier may find it culturally acceptable to express aconcern about the styling of the canoe, but not about theambience of a sporting gear dealership or the attire of thesales staff. It still seems likely, however, that a person withgreater abilities and interests regarding visual aesthetics willgive them relatively high weight wherever they may beencountered.

Future research can also address another limitation of thepresent research by using more disparate product categoriesas arenas of inquiry. It would be worthwhile to see if resultsreported above for consumer durables such as toasters alsoapply to nondurables, packaging, and to more experientialgoods such as restaurant food presentations. The CVPAscale may also have application in cross-regional or cross-

national research. Certain nations, cities, and, regions havereputations for the fine design of their products and the hightastes of their citizens. The CVPA scale could be used toempirically verify whether geographic areas in fact differin terms of the importance design plays in a persons life.Differences observed could be related to regional charac-

teristics to allow greater insight into the development ofCVPA. Furthermore, longitudinal research would allow usto determine whether a regions CVPA is evolving overtime. With the growing attention to design execution in moreand more product classes, one might expect to see an ele-vation of CVPA over the course of time.

It might also prove profitable to explore whether highCVPA consumers have a greater facility for making productand brand inferences based on aesthetic features. Becausethey care about product appearances, those with high CVPAmay retain aesthetically based product meanings that can berecalled during subsequent decision episodes. Those withlow CVPA are not expected to have the previously formedinferences at their disposal and will have to form them at

each purchase decision. The process of acquiring meaningand making judgments when exposed to the visual aspectsof an object will therefore be faster for those with highCVPA.

Other key consumer behaviors and attitudes also may befurther understood through the examination of CVPA. Forexample, our understanding of the antecedents of productinvolvement may be enhanced through investigation ofCVPA as an independent variable. Although consumer re-searchers have actively studied involvement for many years,we know very little about its genesis. It could be that manyinstances of product involvement are based on the responseor value that a consumer absorbs from the aesthetics of aproduct class. In other words, the aesthetic appeal of jewelry,

sports cars, or roses may be the force that creates highlyinvolved consumers for each of these categories. Thus, re-search could explore whether high CVPA consumers aremore likely to be involvement-prone generally due to therelatively high aesthetic payoffs many products providethem.

Similarly, the coinvestigation of visual aesthetics cen-trality and consumer impulse purchasing may be fruitful.Beatty and Ferrell (1998) examined multiple antecedents tothe felt urge to buy impulsively. Although not examined,CVPA may, among certain consumers, be an important an-tecedent to impulse purchasing as suggested by Rooks(1987) work. That is, following a strong response to theaesthetics of a product, it may be very likely that a consumerwill take an impulsive action. Assuming a design-orientedconsumer, such a relationship would lend credence to thestrategies of retailers that are increasingly adding merchan-dise with high aesthetic content. There appear to be manyavenues of future research related to the CVPA scale, boththeoretical and managerial. It is hoped that the ultimate valueof this scale will be demonstrated through its future appli-cation.


14/16


APPENDIX

FIGURE A1

PRODUCT STIMULI

[Received January 2001. Revised July 2002. David GlenMick served as editor and William O. Bearden served as

associate editor for this article.]

REFERENCES

Anderson, James C. and David W. Gerbing (1988), StructuralEquation Modeling in Practice: A Review and RecommendedTwo-Step Approach, Psychological Bulletin, 103 (3), 411423.

Armstrong, J. Scott and Terry S. Overton (1977), Estimating Non-response Bias in Mail Surveys, Journal of Marketing Re-search, 14 (August), 396410.

Bamossy, Gary, Marilyn Johnston, and Michael Parsons (1985),The Assessment of Aesthetic Judgment Ability, EmpiricalStudies of the Arts, 3 (1), 6379.

Bearden, William O., David M. Hardesty, and Randall L. Rose(2001), Consumer Self-Confidence: Refinements in Concep-tualization and Measurement, Journal of Consumer Re-search, 28 (June), 121134.


15/16


Bearden, William O., Richard G. Netemeyer, and Jesse E. Teel(1989), Measurement of Consumer Susceptibility to Inter-personal Influence, Journal of Consumer Research, 15(March), 473481.

Beatty, Sharon E. and Elizabeth M. Ferrell (1998), Impulse Buy-ing: Modeling Its Precursors,Journal of Retailing, 74 (Sum-mer), 169191.

Belk, Russell W., Melanie Wallendorf, and John F. Sherry (1989),The Sacred and Profane in Consumer Behavior: Theodicyon the Odyssey, Journal of Consumer Research, 16 (June),138.

Bentler, Peter M. (1992), EQS: Structural Equations ProgramManual, Los Angeles: BMDP Statistical Software.

Berlyne, Daniel. E. (1971), Aesthetics and Psychobiology, NewYork: Appleton-Century-Crofts.

Bloch, Peter H. (1995), Seeking the Ideal Form: Product Designand Consumer Response, Journal of Marketing, 59 (July),1629.

Bollen, Kenneth A. (1989), Structural Equations and Latent Var-iables, New York: Wiley.

Buchanan, Bruce and Pamela W. Henderson (1992), Assessingthe Bias of Preference, Detection, and Identification Measures

of Discrimination Ability in Product Design, Marketing Sci-ence, 11 (1), 6475.

Campbell, Donald T. and Donald W. Fiske (1959), Convergentand Discriminant Validation by the Multitrait-MultimethodMatrix, Psychological Bulletin, 56 (March), 81105.

Carver, Charles S. (1989), How Should Multifaceted PersonalityConstructs Be Tested? Issues Illustrated by Self-Monitoring,Attributional Style, and Hardiness, Journal of Personalityand Social Psychology, 56 (4), 577585.

Charles, Prince of Wales (1989), A Vision of Britain: A PersonalView of Architecture, London: Doubleday.

Child, Irwin L. (1964), Observations of the Meaning of SomeMeasures of Esthetic Sensitivity,Journal of Psychology, 57,4964.

Childers, Terry L., Michael J. Houston, and Susan E. Heckler

(1985), Measurement of Individual Differences in Visual ver-sus Verbal Information Processing,Journal of Consumer Re-search, 12 (September), 125134.

Cornfield, Betty and Own Edwards (1983), Quintessence: TheQuality of Having It, New York: Crown Publishers.

Csikszentmihalyi, Mihaly and Rick E. Robinson (1990), The Artof Seeing, Malibu, CA: J. Paul Getty Museum.

Davis, Marian L. (1987), Visual Design in Dress, EnglewoodCliffs, NJ: Prentice-Hall.

Dumaine, Brian (1991), Design That Sells and Sells and . . .Fortune (March 11), 8694.

Durgee, Jeffrey F. (1988), Product Drama,Journal of Advertis-ing, 17 (February/March), 4249.

Eagly, Alice H. and Shelly Chaiken (1993), The Psychology ofAttitudes, New York: Hartcourt, Brace, Javanovich College

Publishers.Egan, Peter (2000), Side Glances: Morgans, Old MGs and a

Wooden Canoe, Road & Track (November), 3335.Fornell, Claes and David Larcker (1981), Evaluating Structural

Equation Models with Unobservable Variable, Journal ofMarketing Research, 18 (February), 3950.

Forty, Adrian (1986), Objects of Desire, New York:Pantheon Books.Fournier, Susan (1998), Consumers and Their Brands: Developing

Relationship Theory in Consumer Research,Journal of Con-sumer Research, 24 (March), 343373.

Galbraith, Craig S., Judy A. Siguaw, and Hye-Hwa Lim (1995),

Aesthetic Consumption Differences in Asian-American andAnglo Subcultures: Implications for Segmentation and Po-sitioning Strategies,Journal of Non-Profit and Public Sector

Marketing, 3 (4/4), 324.Goetz, Karl O. et al. (1979). A New Visual Aesthetic Sensitivity

Test, Perceptual and Motor Skills, 49, 795802.Green, Paul E. and Yoram Wind (1975), New Way to Measure

Consumers Judgments, Harvard Business Review, 53(JulyAugust), 107117.

Hirschman, Elizabeth C. (1986), The Effect of Verbal andPictorialAdvertising Stimuli on Aesthetic, Utilitarian and FamiliarityPerceptions,Journal of Advertising, 15 (2), 2734.

Holbrook, Morris B. (1986), Aims, Concepts, and Methods forthe Representation of Individual Differences in Esthetic Re-sponses to Design Features. Journal of Consumer Research13 (December), 337347.

Holbrook, Morris B., and Elizabeth C. Hirschman (1982), TheExperiential Aspects of Consumption, Journal of Consumer

Research, 9 (September), 132140.Hollins, Bill and Stuart Pugh (1990), Successful Product Design,

London: Butterworths.Holt, B. Douglas (1995), How Consumers Consume: A Typology

of Consumption Practices, Journal of Consumer Research,22 (June), 116.

Huber, Joel C. (1997), What Have We Learned from 20 Years ofConjoint Research: When to Use Self-Explicated, GradePairs,Full Profiles, or Choice Experiments, in Sawtooth SoftwareConference Proceedings, http://sawtoothsoftware.com/techabs.shtml#whatwe.

Huber, Joel C., Dick R. Wittink, John A. Fieldler, and Richard L.Miller (1991), An Empirical Comparison of ACA and FullProfile Judgments, in Sawtooth Software Conference Pro-ceedings, http://sawtoothsoftware.com/techabs.shtml#anemp.

Ingarden, Roman (1983), Aesthetic Experience and Aesthetic Ob-ject, in Selected Papers in Aesthetics, ed. P. J. McCormick,New York: Catholic University of America Press, 107132.

Kaiser, H. F. (1974), Little Jiffy, Mark IV, Educational and

Psychology Measurement, 34, 111117.Lastovicka, John L., Lance A. Bettencourt, Renee Shaw Hughner,

and Ronald J. Kuntze (1999), Lifestyle of the Tight andFrugal: Theory and Measurement, Journal of Consumer Re-search, 26 (June), 8598.

Lawson, Bryan (1983),How Designers Think, Westfield, NJ: East-view Editions.

Lichtenstein, Donald R., Peter H. Bloch, and William C. Black(1988), Correlates of Price Acceptability, Journal of Con-sumer Research, 15 (September), 243252.

Loewy, Raymond (1951), Never Leave Well Enough Alone, NewYork: Simon & Schuster

Lynn, Michael and Judy Harris (1997), The Desire for UniqueConsumer Products: A New Individual Difference Scale,Psychology and Marketing, 14, 601616.

Neter, John, William Wasserman, and Michael H. Kutner (1989),Applied Linear Regression Models, Homewood, IL: Irwin.

Nussbaum, Bruce (1999), A Decade of Design: How Great Prod-ucts Can Boost the Bottom Line, Business Week(November29), 8488.

Ortega Y Gasset, J. (1925). The Dehumanization of Art, Princeton,NJ: Princeton University Press.

Osborne, Harold (1986), What Makes an Experience Aesthetic?inPossibility of the Aesthetic Experience, ed. Michael Mitias,Boston: Kluwer Academic Publishers, 117138.

Paulhus, Delroy L. (1993), The Balanced Inventory of Desirable


16/16

Individual Differences in the Centrality of Visual Produact Aesthethic

Documents

Transcript of Individual Differences in the Centrality of Visual Produact Aesthethic