Citation
The influence of unity and prototypicality on aesthetic responses to new product designs

Material Information

Title:
The influence of unity and prototypicality on aesthetic responses to new product designs
Creator:
Veryzer, Robert W
Publication Date:
Language:
English
Physical Description:
vii, 171 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Aesthetics ( jstor )
Bathrooms ( jstor )
Beauty ( jstor )
Clothing ( jstor )
Consumer research ( jstor )
Experiment design ( jstor )
Marketing ( jstor )
Personnel evaluation ( jstor )
Product design ( jstor )
Questionnaires ( jstor )
Consumers -- Attitudes ( lcsh )
Dissertations, Academic -- Marketing -- UF
Marketing thesis Ph. D
New products ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1993.
Bibliography:
Includes bibliographical references (leaves 164-170).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Robert W. Veryzer.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
0030259818 ( ALEPH )
30541242 ( OCLC )

Downloads

This item has the following downloads:


Full Text









THE INFLUENCE OF UNITY AND PROTOTYPICALITY ON
AESTHETIC RESPONSES TO NEW PRODUCT DESIGNS
















By

ROBERT W. \TERYZER, JR.











A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1993























Copyright 1993

by

Robert W. Veryzer, Jr.













ACKNOWLEDGEMENTS


A number of people deserve credit for their role in helping me to accomplish the goal of finishing this dissertation. First and foremost is Dr. J. Wesley Hutchinson, the chairman of my committee, to whom I am extremely grateful. The time and effort that he has contributed to this endeavor are greatly appreciated. This work would not have been possible without his benevolent guidance.

I am very grateful to Dr. Richard Lutz, Dr. Chris Janiszewski, and Dr. David Mick for their. guidance in refining my ideas and for the encouragement that they have given me. I would also like to thank

Dr. John Lynch, Jr. for his guidance on the statistical analyses and Dr. Jonathan Hamilton for contributing an economist's point of view.

I am indebted to my parents, Robert and Marion Veryzer, for their love and support, and for always being there when I need them.

Finally, I would like to thank my brother David who has been a tremendous source of inspiration to me.







TABLE OF CONTENTS



ACKNOWLEDGEMENTS .................................... iii

A B STRA CT ............................................... vi

CHAPTERS

I INTRODUCTION ................................. 1

II CONCEPTUAL BACKGROUND ON THE
FACTORS THAT INFLUENCE AESTHETIC
RESPON SES ................................... 6

Introduction .................................... 6
Factors Affecting Aesthetic Responses ................ 9
Visual Organization Principles ...................... 9
Prototypicality ................................. 14
Derived Responses .............................. 21

III EXPERIMENT 1 ................................ 28

O verview ..................................... 28
Stim uli ...................................... 28
Experimental Design ............................ 39
Experimental Procedure .......................... 40
Results and Discussion .......................... 41
Sum m ary ..................................... 81

IV EXPERIMENT 2 ................................ 84

O verview ..................................... 84
Experimental Design ............................ 84
Experimental Procedure .......................... 86
Results and Discussion .......................... 87
Sum m ary ..................................... 101

V EXPERIMENT 3 ................................ 104

O verview ..................................... 104
Stim ulus M aterials ............................. 105
Experimental Design ............................ 106

iv







Experimental Procedure .......................... 108
Results and Discussion .......................... 108
Sum m ary ..................................... 114

VI GENERAL DISCUSSION ......................... 116

APPENDICES

A LIST OF PRODUCTS EXPLORED FOR USE AS
STIM U LI ...................................... 127

B STIMULUS SETS ............................... 129

C DIAGRAM OF EXPERIMENT 1 .................... 157

D DIAGRAM OF EXPERIMENT 2 .................... 159

E STRONG AND WEAK PRODUCT DESCRIPTIONS ..... 161 REFERENCE LIST ........................................ 164

BIOGRAPHICAL SKETCH .................................. 171













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


THE INFLUENCE OF UNITY AND PROTOTYPICALITY ON
AESTHETIC RESPONSES TO NEW PRODUCT DESIGNS By

Robert W. Veryzer, Jr.

August 1993

Chairman: Dr. J. Wesley Hutchinson Major Department: Marketing

This dissertation investigates the influence of the aesthetic aspects of product appearance on consumers' product preferences and product evaluations. Two factors, unity and prototypicality, are identified and discussed as being the principal factors that influence consumers' aesthetic responses to product designs. The research provides a basis for a theory of consumer aesthetics and has implications for new product development, product quality, and marketing strategy.

The dissertation postulates that consuim ers' responses to product designs are influenced by the design's consistency with the visual organization principle of unity (i. e., a congruity among the elements of a design) and its level of prototypicality (i. e., familiarity; shared features with the category


vi








schema). Furthermore, it is suggested that aesthetic responses influence consumers' attitudes toward products, perceptions of product quality, and price expectations.

Three experiments are conducted in order to examine the role of unity and prototypicality in influencing aesthetic responses and product perceptions. The first study examines the relationship between the two factors (i. e., unity and prototypicality) and consumers' responses to product designs. This study provides evidence that there is a positive effect of unity on aesthetic response and that the favorable aesthetic response generated by consistency with the unity principle also influences "non-aesthetic" product perceptions. The study also finds that the unity factor provides a better explanation of aesthetic response than does prototypicality.

The second and third experiments examine whether or not the context of the evaluation situation (i. e., presence of other similar products, written product descriptions) moderates the influence of unity. These studies find that even though the evaluation context can reduce the magnitude of the unity effect, the unity effect is quite robust and is pervasive.

Taken together, the results of the three experiments indicate that the unity visual organization principle is a very important design factor and that aesthetic responses can have a significant impact on consumers' responses to product designs. The results highlight the relationship between product aesthetics (i. e., design) and consumer behavior.



vii













CHAPTER I
INTRODUCTION


There is a growing recognition that product design, and particularly aesthetic aspects of product design, is emerging as a key marketing element (Kotler and Rath 1984). Throughout the 1980s companies updated their manufacturing methods, quality control, distribution networks, customer service, and labor/management relations in order to compete with other foreign and domestic manufacturers. Today, as more and more manufacturers are able to achieve similar levels of price, quality, reliability, and technology, product appearance is increasingly being acknowledged as the major difference around which consumers can exercise a choice (Oakley 1990, p. 4).

Product design, or "competitive aesthetics" as it has been called (Reed 1990), is gaining recognition as a strategic activity that companies can use to gain a sustainable competitive advantage (Kotler and Rath 1984; Whitney 1988). It is being used for every sort of product from Apple computers to La Croix sparkling water cans. As Bruce Nussbaum pointed out in Business Week (June 17, 1991, p. 62): "Recently, business has grown increasingly aware that design sells. U. S. Companies, in particular, are rediscovering that good design translates into quality products, greater market share, and heftier profits." Although the technical aspects of products remain vitally important,

I







2
more and more manufacturers of all kinds of products are having to come to terms with the reality that product appearance is often the major factor influencing (consciously or unconsciously) what people buy and how much they are willing to pay (Oakley 1990, p. 5).

The influence of aesthetic factors on product preferences and perceptions is an important but often neglected area of study in consumer research. The design of products inherently involves aesthetics. The aesthetic aspects of a product give rise to the registering of affect or pleasure due to the conscious or unconscious influences of the characteristics of the product (Holbrook and Zirlin 1985). The response arising from the interaction between the aesthetic aspects of an object and the perceiver of the object has been termed the "aesthetic response" (Olson 1981). There is a growing recognition among marketing researchers that aesthetic responses can significantly affect consumer behavior and thus aesthetics is gaining recognition as an important marketing variable (Kotler and Rath 1984; Wallendorf 1980). The influence of aesthetics is increasingly being acknowledged as an important part of new product development (Whitney 1988), marketing strategy (Kotler and Rath 1984), product quality (Garvin 1984; Zeithaml 1988), product differentiation (Dickson and Ginter 1987), and competitive advantage (Holt 1985; Kotler and Rath 1984).

Despite the growing awareness of the significant influence that the aesthetic aspects of products can have on product preferences and perceptions,







3

surprisingly little in the way of design/aesthetic theory has been offered that aids in our understanding of how aesthetic responses are formed (Berlyne 1974, p. 5; Gorski 1987, p. 68; Holbrook and Zirlin 1985; Pye 1978, p. 88). Even though disciplines concerned with basic research such as experimental aesthetics remain dormant or in a "protracted infancy" (Berlyne 1971), other disciplines such as industrial (product) design look to them to address the lack of design theory (Gorski 1987, p. 68). The end result is that managers continue to exhibit much "unease" when it comes to making decisions about design and managing design projects (Oakley 1990, p. 7). This is particularly problematic for managers charged with transferring new technology out of the laboratory and into the market. It is at this point that the appearance of a product plays a crucial role in communicating the products identity and use to consumers.

Although consumer research seems to be ideally suited for the study of aesthetic response due to its unique combination of scientific research methods and its tangible research context (i. e., consumer product focus), consumer researchers have taken only tentative steps toward exploring aesthetics and its relationship to consumers' behavior. Much of this work has been concerned with debating the definition and scope of consumer aesthetics rather than examining the nature and influence of aesthetics and aesthetic responses (Holbrook and Zirlin 1985). Thus, aesthetic research has proceeded without the benefit of a conceptual foundation (Holbrook and Zirlin 1985; Olson 1981).








4

Consumer researchers have not developed one, and they have not found one in any of the other disciplines to adopt as a starting point for building aesthetic theory.

This dissertation attempts to address this void in aesthetic theory by examining the influence of the aesthetic aspects of product appearance on consumers' product preferences and perceptions. This research represents important steps toward understanding the implications of aesthetics and product design for consumer research and the formulation of a theory of consumer aesthetics.

In Chapter II the two principal factors, unity and prototypicality, that influence consumers' aesthetic responses to a product's design are discussed and the supporting research is reviewed. Several hypotheses are developed that involve the relationship of the factors to aesthetic and derived responses.

Chapters III and IV each present the method and results of an experiment designed to investigate the influence of the unity and prototypicality factors on aesthetic and derived responses. Chapter III also includes a detailed discussion of how the stimuli employed in all of the experiments were developed.

Chapter V presents the method and results of an experiment that focuses on derived responses.

Chapter VI provides a general discussion of the findings and implications of these results for research that examines the influence of







5

product aesthetics on consumer behavior. The managerial implications of these findings are also discussed.













CHAPTER II
CONCEPTUAL BACKGROUND ON THE FACTORS THAT INFLUENCE AESTHETIC RESPONSES


Introduction

Even though there is a growing awareness of the important role that aesthetics play in influencing product preferences, our understanding of aesthetic responses is extremely limited. Our primitive understanding of aesthetic response phenomena may be due to the highly fragmented approach that has characterized its study. Aesthetics have been studied in a number of fields including philosophy, art history, psychology, experimental aesthetics, industrial design, and more recently consumer behavior. Each field has contributed to our comprehension of aesthetics and yet, little progress has been made in understanding the specifics of aesthetic response (Berkowitz 1987; Wohlwill 1981). Within the many conceptualizations of aesthetic response that have emerged from these different fields of study, there is little agreement and limited insight regarding why a particular object is perceived as pleasurable or beautiful while another is viewed as unattractive. The question concerning what makes an object aesthetically pleasing, regardless of whether it is a painting or a portable stereo, has received diverse and typically vague answers from the disciplines that have studied it. While a lack of convergence is not


6








7

uncommon in interdisciplinary research, the different orientations of the disciplines that are or have been concerned with aesthetic phenomena and the vagueness of the theories that have been offered to explain aesthetic responses have severely inhibited progress toward gaining an understanding of aesthetic responses.

Some aesthetic response theories, which seem to view aesthetic response as idiosyncratic, maintain that their are no laws or principles of aesthetics (Mothersill 1989), while others suggest that inner tendencies of the visual system result in laws that govern perception and thereby influence aesthetic response (Katz 1950; Koffka 1935). There are aesthetic theories that point to fashion trends or the influence of culture as the determinant of systematic aesthetic responses for all classes of products (Pleydell-Pearce 1970; Sproles 1981). There are views of aesthetic response that would suggest that the consumer's preference for a specific product was determined by a desire for unity in variety (Auld 1981; Berlyne 1971; Lauer 1979). Other views maintain that preference is related to prototypicality (Loken and Ward 1990; Nedungadi and Hutchinson 1985). There are also those that periodically suggest that the term "aesthetic response" applies only to works of art (Holbrook 1981), only to be contradicted by those who maintain that all objects have an aesthetic component (Berlyne 1974; Wallendorf 1980). Even the artists and product designers who determine the forms of the objects seem to have trouble agreeing on the basics of design (Lauer 1979) and as yet "have not formulated








8

what they know" (Pye 1978, p. 11). In fact, the well worn maxim that "form follows function" has even been challenged (Lewalski 1988; Pye 1978). Interestingly enough, designers are beginning to look to the disciplines of experimental psychology and consumer behavior in order to understand the cognitive condition of design (Zaff 1987).

The lack of understanding concerning what makes an object aesthetically pleasing is particularly evident in the area of new product development. The issue of aesthetic response and the factors that influence it are frequently overlooked in many discussions of the new product development process. Much of the work in this area seems to sidestep these considerations by either ignoring the role of industrial design (i. e., the process of shaping or giving form to goods that are to be mass produced) in the new product development process or subsuming the industrial design function under the engineering function (e. g., Gruenwald 1985; Urban and Hauser 1993). In either case, the result is that the role of product appearance in the success of a new product is not explicitly acknowledged or addressed. Given the increasing importance of product aesthetics this omission is a rather serious deficiency. Fortunately, the distinct role of industrial design and its relationship to the engineering and marketing functions is beginning to receive the attention that it merits (e. g., Lorenz 1986). However, most of the work that addresses the role of industrial design in the new product development








9

process stops short of identifying and investigating the factors that influence aesthetic responses to products.



Factors Affecting Aesthetic Response

Aesthetic response is a complex phenomenon that is not yet well understood. As with any complex phenomenon it is likely to involve a number of factors; however, prior research seems to suggest that there are two principal factors: the visual organization principle of unity and prototypicality, that may significantly influence aesthetic responses. Some of the significant prior research that suggests that these factors may play an important role in influencing aesthetic responses will now be reviewed. The influence of aesthetics/aesthetic response on non-aesthetic perceptions (i. e., derived responses) will also be discussed.



Visual Organization Principles

The greatest void in aesthetic theory and research (and the greatest opportunity) concerns the identification of specific factors that systematically influence aesthetic responses. Although very general "rules" or "principles" have sometimes been offered (e. g., "unity in variety" -- Hutcheson 1725), the rules are usually vague and unspecified. Even though aestheticians seem to rely on principles such as "unity" and "balance" and aesthetics research has frequently employed scales that attempt to measure dimensions such as








10
disorderly/orderly, weak/powerful, somber/bright, etc. as they pertain to aesthetic stimuli (e. g., paintings), the specific contributions of these dimensions to aesthetic response are not yet well understood. Although principles such as "unity" and "balance" may be understood in terms of visual organization principles such as the Gestalt laws of perception (Lewalski 1988; Pickford 1972, p. 31), no reported research has been found that relates Gestalt laws to general rules regarding aesthetic responses (see Veryzer 1993 for an exception).



Unity

Design Principles such as "unity" describe perceived spatial relations between the parts of a visual display. The design principle of unity refers to a congruity among the elements of a design such that they look as though they belong together or as though there is some visual connection beyond mere chance that has caused them to come together (Lauer 1979). The tendency to perceive groupings of constitutive elements in certain ways or as integrated entities is an important aspect of perception. These tendencies are described by Gestalt laws (Koffka 1935) and design principles (Ching 1979; Lauer 1979). Although design principles are more general than the Gestalt laws of perception, the two sets of rules of perception are related and in some instances a set of Gestalt laws may be used to describe a design principle. For example, the Gestalt laws of proximity (i. e., elements that are closest to each







11
other tend to form groups), similarity (i. e., elements that are similar tend to form groups), and common destiny (i. e., parts of a figure that have a common destiny tend to form units) may be viewed as ways to achieve unity (Katz 1950; Lauer 1979). According to the Gestalt psychologists beauty is dependent on the degree to which an object displays relations consistent with the Gestalt laws of organization. Koffka (1935) clearly suggested this when he discussed how violations of such laws as "good continuation" and "good shape" are not only felt as violations, they conflict with our feeling of "fit" and "hurt our sense of beauty" (Koffka 1935, p. 175).

Although the visual organization principle of unity would seem to be a likely factor in influencing aesthetic responses, there is little in the way of research that relates visual organization principles (e.g., design principles, Gestalt laws) such as unity to aesthetic responses. There is, however, a limited amount of research that indicates that there are general guidelines or principles for combining visual elements in order to maximize aesthetic responses. A study by Bell, Holbrook, and Solomon (1991), which examined the impact of gestalt-like ensemble effects and the influence of personality factors on product evaluations, provides support for the view that unity may systematically influence aesthetic responses. In their study, subjects were asked to look at one of thirty-two color photographs containing various combinations of traditional and contemporary styles of five types of living-room furniture-- specifically, a chair, a table, a piece of sculpture, a floor lamp, and








12
a framed picture (each photograph contained all five types of living-room furniture but differed with respect to the mix of traditional and contemporary styles). Subjects rated the furniture in the randomly assigned photo in terms of perceived unity (i. e., a visual connection among elements that suggests that something beyond mere chance has caused them to come together), aesthetic response, social impression, general liking, and intention to own. Subjects also provided ratings on a number of items which measured personality and motivation variables. Bell et al. (1991), found that aesthetic response did, in

2
fact, depend on perceived unity (R = .05, p < .001) and that perceived unity depended on the product styles (R2 = .19, p < .001 with significant contributions of chair, table, and ensemble). These findings suggest that the principle of unity, which was operationalized in this study as the mix of the two styles shown in each photograph, does seem to influence aesthetic response.

The influence of unity has also been examined in the context of social information processing studies. For example, Lennon (1990) investigated the effects of clothing attractiveness on perceptions. In order to determine the degree to which people perceive others differentially as a function of the attractiveness of their clothing, slides of six different models in business attire, three wearing attractive clothing and three wearing unattractive clothing, were prepared. In the unattractive clothing condition, models wore garments and accessories that did not match either in color, style, or pattern. In the








13
attractive clothing condition, models wore clothing that was well matched and wore accessories to complement their clothing (two pilot studies were conducted to get a consensus regarding clothing attractiveness). Fifty-eight female subjects listened to a pre-recorded audio tape consisting of thirty suggestions relative to marketing a perfume. As a comment was heard, a slide of the woman purported to have made the comment was projected. Subjects rated the women on competence, work comfort (a measure of the extent to which the respondent would feel comfortable working with the woman shown), and sociability. A repeated measures analysis of variance revealed that there was a main effect for clothing attractiveness on perceived competence (F(1,58) = 52.14, p = .000). Models dressed in attractive clothing (M = 55.15) were perceived to be more competent than models dressed in unattractive clothing (M = 46.49). There was also a main effect for clothing attractiveness on perceived work comfort (F(1,58) = 24.18, p =.000). Respondents indicated that they would feel more comfortable working with attractively dressed models (M = 22.02) than with unattractively dressed models (M = 19.10). Finally, there was a main effect for clothing attractiveness on perceived sociability (F(1,58) = 5.28, p = .025). Attractively dressed models (M = 11.41) were perceived to be more sociable than those models who were unattractively dressed (mean = 10.86). These results provide further support for the view that configurations of aesthetic elements (i. e., visual organization rules such as unity, color







14

harmony, repetition, etc.) can significantly influence perceptions. The preceding discussion suggests the following hypothesis:


Hi: Aesthetic responses are more positive for objects (products) exhibiting high consistency with the visual organization principle of unity than they are for objects (products) that are not

consistent with this visual organization principle.



Prototypicality

Another factor that seems to exert an influence on aesthetic response is that of experience. Familiarity has been shown to lead to positive affect (Kunst-Wilson and Zajonc 1980). This would seem to suggest that more typical or familiar items should be better liked (Loken and Ward 1990). Typicality, or prototypicality, is concerned with the degree to which an object is representative of a category. A prototype can be defined as the central representation of a category, as possessing the average or mean value of the attributes of that category and as representing the averaged members of the class (Langlois and Roggman 1990; Rosch 1978). According to the prototypicality view, people respond most favorably to objects that are highly protypical and less favorably to objects that are less prototypical (Glass and Holyoak 1986, p. 170; Langlois and Roggman 1990). The prototypicality explanation of preference maintains that more prototypical examples tend to be better liked (Loken and Ward 1990; Nedungadi and Hutchinson 1985). Although prototypi-








15

cality theories were not developed to specifically address aesthetic issues, the absence of a theory of aesthetic response has led consumer researchers as well as others to rely on prototypicality as a default theory for explaining aesthetic response.

A number of explanations have been suggested for the relationship between prototypicality and preference. One explanation for the relationship between prototypicality and preference/attitude suggests that more prototypical items are more familiar and therefore better liked. Familiarity refers to either an item's meaningfulness (i. e., perceived knowledge about an item) or the frequency of exposure to the item (Loken and Ward 1990). Another

explanation suggests that more prototypical category members are preferred because they have more valued attributes. This explanation does not hold that prototypically per se leads to product preference, but rather maintains that as product categories evolve one or a few products tend to become market-share leaders because they have attributes widely desired by consumers who buy the product. Competitive brands are designed to appeal to the same segment(s) of consumers so they are similar in many ways to market leaders (Loken and Ward 1990). It has also been suggested that the link between prototypicality and preference may in part be due to the information theory notion of redundancy in that prototypes appear to be just those members of a category that most reflect the "redundancy structure" of the category as a whole (Rosch 1978, p. 37). The preceding discussion suggests the following hypothesis:








16

H2: Objects (products) that are more prototypical (i. e., have more shared features with the category schema) will receive more favorable aesthetic responses than objects that are less prototypical (U. e., fewer shared features with the category

schema).

Although prototypicality/familiarity seems to provide a satisfactory explanation of aesthetic response in some cases (e. g., Kunst-Wilson and Zajonc 1980; Loken and Ward 1990), such an explanation does not seem adequate in others (e. g., Meyers-Levy and Tybout 1989). In fact, in some cases it is the converse of prototypicality (i. e., novelty, distinctiveness) that seems to account for positive aesthetic response (Woll and Graesser 1982). Some research that examined novelty and complexity may provide insight into this apparent inconsistency. In a study that examined novelty ratings of simple and complex shapes, Eisenman (1968) found that more complex polygons (i. e., those with more sides) were rated as being more novel. In a related study, Berlyne (1970) examined the effect of repeated presentation on hedonic value for simple and complex (i. e., novel) patterns. Berlyne had subjects rate two simple and two complex patterns six times on a 7-point pleasingness scale. Between consecutive tests the subjects saw each of the patterns eight times without having to record a judgement. The results confirmed that ratings of complex (i. e., novel) patterns rose and then fell after reaching a maximum at the third test. Ratings of the simple (i. e., less novel) patterns, which were initially








17

higher than those for the more complex (novel) patterns, fell throughout the tests until they finally flattened out (see Figure II 1). This would seem to explain why (and roughly when) prototypicality will be liked better in some cases and novelty will be better liked in others.

Mandler (1982) has theorized that the level of congruity between a product and a more general product category schema may influence the nature of information processing and thus product evaluations. Products that are moderately incongruent with their associated category schemes are said to stimulate processing that leads to a more favorable evaluation relative to products that are either congruent or extremely incongruent with the category schema. Mandler suggests that moderate incongruities are those that can be successfully and readily resolved by the processor.

Meyers-Levy and Tybout (1989) conducted a series of experiments to test Mandler's schema (in)congruity hypothesis. The general method, which was modified slightly over the course of the three studies that were conducted, consisted of presenting subjects with descriptions and samples of beverage/soft drink products and having them evaluate the products along dimensions such as appeal, taste, quality, interest in trial, etc. In their design, schema congruity and schema incongruity were manipulated by varying a single attribute in the product description (high preservatives vs. all natural ingredients) so as to alter the structural and descriptive congruence of the product description with the schema activated by the category label (beverage








18



4645

44

43 //
Mean 42 *O
Pleasingness 41 Rating 40


39 -S

38 37

1 2 3 4 5 6
Tests




Mean pleasingness ratings for complex (C) and simple (S) patterns in successive test. Consecutive tests were separated by eight presentations
of each pattern. (Source: Berlyne 1970)



FIGURE II 1 BERLYNE'S EXPERIMENTAL RESULTS








19
vs. soft drink) that was specified in the first sentence of the product description (e. g., subjects were told that they were going to be evaluating either a new "beverage" or a new "soft drink" and later they were told that the product had either "high preservatives" or had "all natural ingredients"). Across their three experiments Meyers-Levy and Tybout found support for Mandler's view that the process of responding to levels of schema congruity influences evaluations, and that moderate schema incongruity enhances evaluations. Moderate schema incongruity led to more favorable evaluations than either schema congruity or extreme schema incongruity. This discussion suggests the following hypotheses:


H& Moderate schema incongruity (i. e., distinctiveness, novelty) leads to a more favorable aesthetic response than does complete congruity between a product and its product class (i. e.,

prototypicality).


H4: Moderate schema incongruity leads to more favorable

aesthetic responses than does extreme schema incongruity.


It is also possible that people may prefer more novel products due to variety seeking (Holbrook and Hirschman 1982) or perhaps because of a product's salience relative to other products (Loken and Ward 1990; Woll and Graesser 1982).








20

H5: Objects (products) that are more novel (i. e., atypical) will receive more favorable aesthetic response ratings than objects

that are less novel U. e., more prototypical).


These hypotheses would predict that a product that exhibits a singular change from the category prototype should receive more positive aesthetic ratings than either the category prototype or a product that exhibits multiple changes G. e., extreme schema incongruity). While this result might be expected if familiarity is entirely a function of memory, it is possible that the unity design factor may influence feelings of "perceived familiarity." In an instance where two attributes of a stimulus were altered such that the stimulus was no longer prototypical but it did exhibit unity, the relational similarity (i. e., unity) to the category prototype may generate a sense of "perceived" familiarity (Goldstone, Medin, and Gentner 1991). This discussion suggests the following hypothesis:


116: Consistency with the visual organization principle of unity

is positively related to perceived familiarity.


In addition to the hypotheses that have already been presented concerning the influence of specific factors on aesthetic response, there are several other important hypotheses concerning the influence of the aesthetic responses that are fostered by these factors. These hypotheses are developed and presented in the section that follows.







21

Derived Responses

Attitude

Aesthetic responses seem to influence derived responses (i. e., nonaesthetic evaluations) although their influence is often attributed to other factors (e. g., Berkowitz 1987). One area that seems to be influenced by aesthetic responses is that of attitude. Bell et al. (1991) have suggested that aesthetic responses are key determinants of general liking. In their study, which was described earlier, they found evidence that general liking was influenced by aesthetic response (recall that aesthetic response in their study was influenced by perceived unity which was manipulated by means of the ensemble that was depicted).

In general, "attitudes" may be conceptualized as evaluative judgements. The term is usually used to refer to an individual's disposition to respond favorably or unfavorably to an object, person, institution, or event, or to any other discriminable aspect of an individual's world (AJzen 1989, p. 241). Ajzen (1989) has noted that affective reactions may feed into the overall evaluative response to an attitude object and thus may be at least partly responsible for the evaluative direction and intensity of a person's beliefs.

The influence of affect due to aesthetic factors U. e., happy or angry looking faces, taste, smell) on attitude formation and change was examined in a study by Edwards (1990). In two experiments, Edwards examined the hypotheses that the sequence of affect and cognition in an attitude's formation







22

is an important determinant of its subsequent resistance to affective and cognitive means of persuasion. Affect-based and cognition-based attitudes were induced and subsequently challenged by either affective or cognitive means of persuasion. The procedure used to create the two types of attitudes and the means of persuasion involved varying the sequence of affect and cognition while holding the content of communications constant. Edwards found that affect-based attitudes exhibited more change under affective means of persuasion than under cognitive means of persuasion. Cognition-based attitudes, on the other hand, exhibited equal change under both forms of persuasion. In addition, it was found that affect-based attitudes were expressed with greater confidence than their cognition-based counterparts. These findings demonstrate that aesthetic responses can play an important role in determining attitudes.


HT Attitudes (e. g., general liking) towards products are more favorable for products receiving more positive aesthetic ratings

than for products receiving less positive aesthetic ratings.



Perceived Quality

Another important area where aesthetic responses seem to play an important role in influencing non-aesthetic evaluations is that of product quality. Garvin (1984) has identified aesthetics as one of the eight dimensions of quality. Zeithaml (1988) has suggested that "intrinsic cues," which involve








23

the physical composition of the product (e. g., texture, color, flavor, etc.), are very important in signaling perceived quality to the consumer. In a study by Berkowitz (1987), consumers seemed to make unconscious inferences concerning freshness, taste, and quality based on the shapes of the products. Berkowitz examined consumer reaction to a food product -- frozen corn on the cob of two shapes (full ears with squared-off ends and full ears with untrimmed ends), in order to determine: (1) whether the shape of the product would influence preference; and (2) whether preference levels would vary with involvement and experience with the product category. The experimental design involved paired comparison tests at laboratory kitchens in enclosed malls and sequential monadic tests in subsequent home placements. Test panelists included 286 female homemakers of which 184 currently purchased the frozen variety of corn and 102 bought only the fresh variety. The findings showed a marked preference for the untrimmed shape. Preference ratios comparing preference scores for the rounded, untrimmed shape to those for the squared-off shape were:

Laboratory test -- 1.1: 1 frozen users; 2.0: 1 fresh only

Home placement -- 1.8: 1 frozen users; 2.2: 1 fresh only

The results were statistically significant at the .01 level. Nearly four out of five consumers said the reason for their choice in the home test was better taste, about half said the untrimmed was a more natural product, and half reported better texture. Visual appeal or a more pleasing shape, per se, were







24
very minor motivations. Ratings on ten attributes showed the basis of consumer preference in a more systematic way. Overall preference ratios were as follows: more like fresh -- 3.2; more natural -- 3.1; taste/flavor 2.3; quality -- 2.3; size -- 1.7; texture -- 1.5; shape -- 1.5; and appearance 1.1.

Panelists did not misperceive criteria like ease of preparation (1.0) and ease of holding (.8) for which shape had objectively little impact.

Berkowitz suggests that the data seem to indicate a chain of interrelated inferences which stem from the shape of the product rather than a single direct linkage. He notes that the findings indicate that an attribute communicated and presumably noticed may not be considered by consumers to be discriminating, but the attribute(s) that it triggers may be considered to be discriminating. Thus, the squared-off shape of the one test item may have fostered an association (or cued categorization) with processing or processed products while the more natural looking product may have been associated with freshness or "fun experiences such as summer family barbecues when fresh corn was served" (Berkowitz 1987, p. 559). Berkowitz' notion of "interrelated inferences" is similar to the concepts of perceptual categorization and perceptual inferences (Wilkie 1990). Wilkie (1990) has pointed out how these processes, which translate sensory inputs into a mental "identification" of a particular stimulus and develop (i. e., construct) beliefs concerning the stimulus based on other information such as stimulus properties, lead to








25
consumer inferences and thus play a major role in directing consumer behavior (Wilkie 1990, p. 267).

Evidence for the influence of aesthetic response on quality and ability/performance evaluations can also be found in the social perception literature. Landy and Sigall (1974) found significant main effects for writer attractiveness on evaluations of a writer and her work. Similarly, Lennon (1990) found a significant effect of clothing attractiveness (clothing and accessories that matched vs. clothing and accessories that did not match in color, style or pattern) on perceived competence. The findings of these studies suggest that the aesthetic aspect of products (i. e., objects) and the aesthetic responses that they give rise to may exert an influence on non-aesthetic aspects of products such as quality.


H8: Quality ratings are higher for products that receive more positive aesthetic ratings than for products that receive less

positive aesthetic ratings.



Price

The possibility that aesthetic response may influence perceived product quality raises some interesting questions with regard to the price-quality relationship. This relationship has been examined primarily in terms of price as cue to quality (Monroe 1973; Zeithaml 1988). Even though a positive priceperceived quality relationship does appear to exist, results of studies that have








26

examined the price-quality relationship have been somewhat mixed and the findings imply that price may not be the dominant cue in quality perception (Monroe 1977; Zeithaml 1988). Moreover, there seem to be cases where perceptions of (high) quality are formed without being diminished by (low) price and the quality perceptions subsequently influence price (e. g., the perceived quality of Japanese automobiles despite their initial low prices).

In cases where consumers initially do not have price information it seems likely that they might form impressions about a product based on nonprice information (e. g., physical composition of the product, packaging, brand name, etc.) and that these impressions could influence price expectations. Thus, it is conceivable that the same design factors that influence aesthetic responses and thereby perceptions of product quality may also influence price expectations.


H9: Price expectations (ratings) are higher for products that receive more positive aesthetic ratings than for products that

receive less positive aesthetic ratings.


Ultimately, the price expectations fostered by product design/aesthetics may play a role determining consumers' "price thresholds" (Monroe 1973) for a particular product within a category.

If indeed unity and prototypicality systematically influence aesthetic responses and these, in turn, influence product preferences and product








27
perceptions, then one would expect these design factors to influence the ratings of products in accordance with (the level of) their presence (or absence) in the products. Aesthetic responses would be expected to be more positive for objects (products) exhibiting high unity than they would for objects (products) that were not consistent with this visual organization principle. In general, prototypicality would be expected to lead to more favorable aesthetic responses than would atypicality; however, there is some question about the nature of this relationship with regard to the level of prototypicality (i. e., prototypical or moderately atypical) that maximizes positive aesthetic response. The unity and typicality factors would also be expected to influence non-aesthetic evaluations or derived responses such as attitude, perceived quality, and price expectations.

In the chapters that follow a series of three experiments that examine the hypotheses that have been presented here are discussed. Experiments I and 2 examine the influence of unity and prototypicality on aesthetic responses (i. e., H1, H2, H3 H4, and 115) and derived responses (i. e., H7, H8, and 119). The influence of unity on perceived familiarity is also examined (i. e., 116). Experiment 3 focuses on the influence of product aesthetics on the attitude and perceived quality derived responses (i. e., H7 and H8). In this experiment, additional information is presented during the evaluation task in order to examine whether or not the presence of the information moderates the influence of aesthetic response on the derived responses.













CHAPTER III
EXPERIMENT 1



Overview

The hypotheses developed in Chapter II concern the role of design factors in influencing aesthetic and derived responses. It was hypothesized that aesthetic and derived responses would be more positive for products exhibiting high unity than they would be for products that were not consistent with the unity visual organization principle. It was also suggested that although there is some question about the level of prototypicality that maximizes positive aesthetic response, prototypicality, in general, would be expected to lead to more favorable aesthetic and derived responses. This chapter first discusses the stimuli and methodology that were used in Experiment 1 to examine these hypotheses. The analyses and results of Experiment 1 are then presented and discussed.



Stimuli

In order to examine the hypotheses concerning the influence of unity and prototypicality stimulus sets were created that allowed these factors to be manipulated independently. This discussion will first describe how these


28








29
manipulations were accomplished. The manipulations will then be related to the basic design employed in the experiment. This will be followed by a general discussion of how the stimulus products employed in this experiment were selected and developed.



Stimulus Manipulations

The stimulus sets (i. e., design sets) were constructed by first selecting a prototypical product (i. e., product form) from a product category. Two prominent parts of the prototypical product were then selected for manipulation. Three variations of the prototypical product were then produced by altering either one or the other of the two selected features or both features. This produced a stimulus set consisting of four product variations (i. e., the original prototypical product variation, two variations that contained one altered feature, and one variation in which both features had been altered). The two product variations that shared one of the selected features with the prototypical variation (i. e., only one feature had been altered) were moderately atypical products. The product variation that did not share either of the two selected features with the prototype (i. e., both features had been altered) was the most atypical product of the set. Thus, within the stimulus set of four products three levels of prototypicality were represented (i. e., prototypical, moderately atypical, and extremely atypical).








30
The transformation of the two product features was done in a way that also manipulated each product variation's consistency with the unity visual organization principle. In the case of the prototypical variation the two selected features displayed a visual connection with each other (e. g., a repetition of the same shape). When one of the two selected features was altered it was done in a way that decreased unity (e. g., did not display a visual connection with other parts of the product) and therefore decreased the unity exhibited by the product variation. This was the case for each of the variations in which only one feature had been altered (i. e., the moderately atypical variations). In the case of the product variation where both features were altered unity was again achieved since the altered features displayed a visual connection to each other even though the features were very different from the features of the prototypical variation. Thus, changes in the two selected product features resulted in two levels of unity (unified and ununified), and three levels of prototypicality (prototypical, moderately atypical, extremely atypical).

An example of how unity and prototypicality were manipulated by transforming two of a product's features is shown in Figure III-1. In this example two features of a telephone (handset and base) are altered using the shape transformation in order to make the product either more or less prototypical and more or less unified. The upper left-hand cell of the figure contains the most prototypical (++) form of the product. The prototypicality of







31
Feature A
Base











L3

Feature B
Handset Unity/Prototypical (++) Non-Unity/Atypical (+-)














Non-Unity/Atypical (-+) Unity/Atypical (--)


(++) Prototypical Handset, Prototypical Base
(+-) Prototypical Handset, Atypical Base (-+) Atypical Handset, Prototypical Base
(--) Atypical Handset, Atypical Base


FIGURE III-1
EXAMPLE OF UNITY AND PROTOTYPICALITY MANIPULATIONS








32

all of the products that were used as stimuli was established through an examination of the products available on the market and was confirmed in pilot tests.1 This product form is also unified in that there is consistency or compatibility among the shapes of the base (Feature A) and the handset (Feature B). In this (prototypical) case the number of shared features with the category schema is high (positive) and unity or a visual connection or consistency among product features is also high (positive). In the upper righthand cell of Figure 111-i, the base (Feature A) of the product has been altered using a shape transformation so that it is no longer the same as that of the category prototype. This change has at the same time decreased the unity between the handset feature and the base feature. Thus, this version of the telephone product is atypical (+-) with respect to one of the two features being manipulated and is no longer unified. The same is true of the product in the lower-left cell of Figure III-i. In this case, the shape of the handset was altered in order to effect the manipulations of prototypicality and unity that would produce an atypical (-+) stimulus on one product feature and an "n unified" appearance. The base and handset features of the product in the lower-right corner of Figure III-1 have both been altered. In this case the product no longer shares the shape of either Feature A or Feature B with the


1 The aggregate mean ratings for the familiarity of the eleven stimulus
sets that were pilot tested were as follows: Unity/Prototypical (M = 6.46), Non-unity/Atypical (M = 4.86), Non-unity/Atypical (M = 4.96), and Unity/Atypical (4.76). Product variations were rated on 9-point semantic
differential scales with 9 being the most familiar.








33

category prototype. Thus, this product is quite atypical (--) with respect to prototypicality. This version does, however, exhibit unity since there is a visual connection or consistency among the two product features. As can be seen in the figure, this manipulation of the two product features results in three levels of prototypicality and two levels of unity being produced.

Four possible patterns of results are shown in Figure 111-2. These patterns have been labeled to indicate the explanation of aesthetic response that each supports. The unity explanation of aesthetic response would be supported by results showing that both unity conditions were more highly rated than both of the non-unity conditions (Figure JJI-2a). Such results would suggest that a product variation's consistency with the unity visual organization principle positively influenced subjects' aesthetic responses. The predictions of the schema incongruity hypotheses are the opposite of those for the unity explanation of aesthetic response. The schema incongruity hypotheses predict that moderately atypical stimuli will be preferred to stimuli that are prototypical or extremely atypical (Figure IJI-2b). The prototypicality explanation of aesthetic response would be supported by the results depicted in Figure III-2c. Here, the more prototypical a product variation is the higher its ratings. A pattern of results exactly opposite those for the prototypicality explanation would suggest an effect of novelty (Figure JJJ-2d). In this way, the basic 2(Feature A) x 2(Feature B) design allows the hypotheses presented earlier to be examined.







34




hifty Sch a Incolgndty
Feature A Feature A


Feature B ......... Feature B



(a) (b)










ProttyPicality Novelty
Feature A Feature A

+ 0 0
Feature B Feature B
0 0 +

(c) (d)






+ indicates the highest (i. e., most positive) rating(s) for a product version contained in the stimulus set.
- indicates the lowest rating(s) for a product version contained in the set.
0 indicates a rating between the highest and lowest ratings.

FIGURE 111-2
POSSIBLE PATTERNS OF RESULTS








35

Stimulus Development

The construction of stimuli that are to be used in research that examines aesthetic influences requires a great deal of care. Many aspects of an object's appearance have the potential to affect aesthetic responses. The influence of aesthetic aspects of a stimulus object (e. g., a product or a picture of a product) such as color, perspective, shading, etc. that could affect an aesthetic response must be eliminated or controlled. In addition to the problems and limitations inherent in controlling for extraneous aesthetic influences, the construction of visual (i. e., pictorial) stimuli for aesthetic research is further complicated by the fact that it is usually difficult to precisely determine the strength of variables of interest (Nunnally 1981).

Although there are quite a number of studies that have manipulated prototypicality (e. g., Hutchinson and Alba 1991), there is little in the way of precedent for constructing stimuli that simultaneously exhibit different levels of prototypicality and unity. The construction of sets of stimuli for this experiment necessitated an exploration of the ways that unity and prototypicality could be manipulated across a range of products. There were four principal requirements that guided the development of the stimulus sets. The first requirement, which pertained to the selection of particular products, was that the product class had to have a strong category prototype. In order to investigate the hypotheses concerning prototypicality it was necessary for there to be a strong prototypical product design (i. e., form, configuration) for








36
the product category. A second requirement that directed the selection of products for the stimuli sets concerned the (non)existence of a product that was atypical but unified. Product categories that contained instances of products that were atypical (i. e., novel) and unified were considered problematic because in such cases people may be more favorably disposed U. e., receptive) toward atypical (but unified) product designs due to prior experience with or knowledge of the product. If this were the case it could lead to results that overstated the influence of the unity design factor. This is especially important since Experiment 1 effectively provides a theory test between a prototypicality explanation of aesthetic response and a unity explanation. A third

requirement involved product conduciveness to the manipulation of unity. In order for a product to exhibit unity U. e., a visual connection among elements, repetition of form or pattern) or disunity the product had to have parts that could be perceived and manipulated separately. The fourth requirement that guided the development of the stimulus sets concerned the medium used to create the stimuli and stimulus communicability. A medium was needed that would allow the creation and presentation of stimulus products that were drastically different from the products that currently existed. This made it difficult to construct stimuli by altering (either photographically or using computer scanned images) existing products or pictures of existing products because such modification often results in introducing "aesthetic confounds"







37
(e. g., inconsistent perspective) and unintended degradation (e. g., blurry edges, cloudy surfaces) into a stimulus.2

The need for a method that afforded the construction of previously untreatedd" (i. e., nonexistent) products that were markedly different from existing products and the need to control for extraneous aesthetic influences led to the use of line drawings. The use of drawings does entail a trade-off of realism for "producibility" and greater experimental control. While this reduction in realism is unfortunate, it is not uncommon for research that is conducted in a laboratory setting. The use of drawings was also more practical for the purpose of reproducing the stimuli for inclusion in the booklets used in the experiment. This method of constructing stimuli necessitated the selection of products that could be clearly communicated through simple line drawings.

A number of products in a wide variety of product categories were examined in catalogs and in stores in order to determine their suitability for use as stimuli in this experiment. An initial series of product. "studies" (i. e., sketches) was done to explore ways of manipulating products along the unity and prototypicality dimensions. A number of products that seemed to meet the four requirements were then selected for further development in product



2 The possibility of modifying computer scanned images was explored.
This approach proved to be unsatisfactory for the purposes of creating the stimuli for this particular experiment because the radical changes that had to be made to the computer scanned images introduced aesthetic confounds into the stimuli and often resulted in severe
degradation of the image.








38

studies. A list of these products is presented in Appendix A. These product studies led to the identification of three basic ways to transform the features of the products so as to simultaneously accomplish the prototypicality and unity manipulations. These transformations involved altering product features by changing the shape of parts of the product, adding texture to parts of the product, or adding trim to part of the product. Additional product studies were then undertaken to determine which products could be manipulated using each of the three transformations (shape, trim, and texture). These studies involved producing four versions of each product (for each of the three types of transformations) by systematically transforming two product features in order to accomplish the unity manipulation (i. e., a visual connection among product features) at a particular level of prototypicality (i. e., shared features with the category prototype).

In order to increase the generalizability of the results and demonstrate the robustness of the effects of interest, nine product categories were selected for use as stimuli. These nine product categories were: alarm clocks, bathroom scales, dressers, flashlights, hair dryers, lamps, refrigerators, telephones, and television remote controls. This resulted in a total of twenty-seven replications (nine product categories x three types of transformation). Each of these products had to be produced in four versions (Unity/Prototypical +, Nonunity/Atypical +-, Non-unity/Atypical -+, and Unity/Atypical .--). Thus, the








39

entire stimuli set contained one hundred and eight drawings of products. Examples of these stimuli sets are presented in Appendix B.



Experimental Design

The overall design of the experiment is a 2(Feature A) x 2(Feature B) x 9(Products) x 3(Version) x 2(Order) mixed factorial design. As was pointed out in the preceding discussion the manipulation of features A and B result in the manipulation of prototypicality and unity. The twenty-seven replicate sets of stimuli were organized into three questionnaire versions. Each of the three questionnaire versions contained all nine of the product categories (each product category was made up of a set of four variations of the product -- i. e., Feature A x Feature B) but for only one of the three transformation types (shape, trim, or texture). So for example, questionnaire version one contained the telephone stimulus set (set of four product variations) that were altered using the trim transformation; questionnaire version two contained the telephone stimulus set altered using the texture transformation; and questionnaire version three contained the telephone stimulus set altered using the shape transformation. While each questionnaire contained each of the nine product categories, in three of the nine cases the trim transformation was utilized, in three of the nine cases the texture transformation was utilized, and in three of the nine cases the shape transformation was utilized. A diagram of this design is presented in Appendix C. Thus, Feature A, Feature B,








40

Products, and Transformation were all within-subjects factors. Particular transformations of the nine product categories (i. e., stimulus sets) were contained in the questionnaire Versions. Questionnaire Version and Order were between-subjects factors. The order in which the stimulus products (stimulus sets) were presented was reversed for half of the subjects.



Experimental Procedure

One-hundred and ninety-seven volunteer subjects enrolled in the Introductory Marketing course at the University of Florida participated in this experiment.3 Subjects received extra credit for their participation. The subjects were run in groups of 10-20 participants. The stimulus materials were contained in a booklet. The introductory page informed the subjects that they would be shown drawings of products and asked to evaluate the appeal of the product ideas based on their appearance. The subjects were told that the purpose of the study was to obtain consumers' reactions to products that companies were considering for introduction. It was also explained to the subjects that the products they were going to be evaluating were in the early stage of the product development process and for this reason the drawings were in a very rough (unfinished) form. Subjects were then told that each of the versions of a product performed equally well and that they were to rate all


3 Power estimates using pilot data and the procedure suggested by Cohen
(1977, pp. 364-379) for factorial designs indicated that this sample size
would be sufficient for estimating effect sizes at the p < .05 level.







41
four of the product versions that were shown on each page. Subjects were then allowed to proceed through the task of rating the products at their own pace. The subjects rated each product design on 9-point semantic differential scales that measured aesthetic response (beautiful/ugly), familiarity (familiar/unfamiliar), attitude (like/dislike), quality (high quality/low quality), and price (high price/low price).4 Following the rating tasks, subjects were asked to write about how they determined the ratings that they gave to the proposed products. The entire procedure took appro3dmately one-half hour to complete.



Results and Discussion

The manipulation check of the effect of a change in either Feature A or Feature B on aesthetic responses found that there was, in fact, a reduction in ratings of the appearance of the products (compared to those of the prototypical product version) as measured on the semantic differential scale anchored by Beautiful[Ugly. This is to say that in the case of the product variations that were altered so that one feature (either Feature A or Feature B) was no longer prototypical or unified the effect of the change was a reduction in positive aesthetic response. The average effect of a change in Feature A on aesthetic responses was -1.24. Likewise, the average effect of a change in Feature B on



4 The direction of some of these scales was reversed in order to reduce
possible response bias on the part of subjects.








42
aesthetic response was a -1.34. This pattern was consistent across most of the sets of stimuli for each product transformation (i. e., stimulus sets).5 The effects of a change in Feature A or Feature B from the category prototype are shown in Table III-1. In twenty-one of the twenty-seven stimulus sets the change of either Feature A or Feature B resulted in a lower beauty rating for the product variation as compared with the rating for the category prototype. In two stimulus sets the effects of a change in the product features were mixed. That is, a change in one feature (either A or B) had a negative effect on aesthetic response but a change in the other feature (B or A) had a positive effect (see the lamp/trim and lamp/texture stimulus sets in Table III-1). In only four of the twenty-seven stimulus sets were the effects of a change in Feature A and Feature B both (separately) in the opposite (i. e., positive) direction (see the bathroom scale/trim, bathroom scale/texture, TV remote control/trim, and the hair dryer/trim stimulus sets in Table III-1). Overall, the manipulation check demonstrates that for the majority of the stimulus sets a reduction in favorable aesthetic response does occur when the products are altered in such a way as to be both less prototypical and less unified than the category prototype. It should be noted that there is variation in the reduction in favorable aesthetic response across the twenty-seven stimulus sets. This is to be expected given the difficulties of precisely estimating the strength of each


5 "Stimulus set" will be used throughout this discussion to refer to one of
the transformation types (shape, trim, or texture) of a product category
(e. g., telephone/shape transformation set).








43
TABLE III 1
MANIPULATION CHECK FOR THE EFFECT OF A CHANGE
IN FEATURE A OR B ON BEAUTY RATINGS


Product Feature A Feature B
1. Bathroom scale
Sh -0.14 -0.17
Tr 1.00 0.14
TX 1.77 0.86

2. TV Control
Sh -0.73 -0.22
Tr 0.65 0.97
Tx -0.98 -0.84

3. Flashlight
Sh -1.83 -1.16
Tr -3.04 -3.55
TX -2.36 -3.31

4. Lamp
Sh -0.16 -0.14
Tr 0.02 -1.28
TX -0.07 1.06

5. Refrigerator
Sh -4.00 -3.60
Tr -1.09 -1.12
Tx -3.80 -3.64

6. Telephone
Sh -2.60 -1.01
Tr -0.28 -1.25
Tx -0.56 -0.99








44

TABLE III 1 -- continued
MANIPULATION CHECK FOR THE EFFECT OF A CHANGE
IN FEATURE A OR B ON BEAUTY RATINGS


Product Feature A Feature B

7. Hair dryer
Sh -1.59 -3.72
Tr 0.33 0.33
Tx -3.58 -3.77

8. Dresser
Sh -0.13 -0.54
Tr -1.88 -1.75
TX -2.07 -1.20

9. Clock
Sh -3.33 -3.00
Tr -1.43 -1.44
Tx -1.60 -1.72

Mean effect of change
across products 1-1.24 -1.34


manipulation (i. e., change in Features A and B via the transformation type) as it is adapted for each of the stimulus sets.



The Influence of Unity and Prototypicality on Aesthetic and Derived Responses

The critical test concerns what happens to aesthetic responses in the case where both Feature A and Feature B are altered so that they are less prototypical but more unified. The prediction consistent with the "prototypicality't hypothesis is that in such a case the aesthetic response ratings should go down since the resulting stimulus is quite atypical. The








45
"unity" hypothesis, however, suggests that if the changes in the features are made in a way that increases unity then aesthetic responses should be higher in this case (as opposed to the non-unity cases) since the product version exhibits consistency with the visual organization principle of unity. This is in effect a theory test (for a linear model) which corresponds to a test of the Feature A x Feature B interaction. This test can be extended to the

investigation of the influence of unity and prototypicality on derived responses.



Analyses

The data were analyzed using mixed ANOVA designs with Version and Order being treated as between-subjects factors and Feature A, Feature B, and Products being treated as within-subjects factors. Transformation was also treated as a within-subjects factor because each subject saw products that had been altered using the three different transformation types.

Subjects' product ratings on the five 9-point semantic differential scales were treated in three different ways in order to compute the dependent measures that were analyzed. The first approach was to perform the analysis directly on subjects' ratings of the products. A second approach involved computing the linear contrast that reflected the interaction of Feature A and Feature B (i. e., the main effect of unity) for each design set. Under this approach difference or "interaction" scores were formed by adding subjects' ratings for the Unity/Prototypical and Unity/Atypical conditions and








46
subtracting from this the ratings for both of the Non-unity/Atypical conditions. The resulting score provides a straight-forward measure of the main effect of unity. The third way of forming the dependent measures involved a modification of the interaction score. This variation entailed subtracting the product rating of the lowest rated Non-unity/Atypical condition from the product rating for the Unity/Atypical condition. This score allows a non-linear model of prototypicality to be examined (see Figure 111-3). If prototypicality dominates unity then product versions that are more similar to the category prototype should receive higher ratings than less typical product versions (i. e., the minimum rating for either of the Non-unity/Atypical conditions should always be greater than the rating for the Unity/Atypical condition since the former always has more shared features with the category prototype than does the latter).

Both the unity linear score (UL) and the unity nonlinear score (UN) are needed in order to get a (more) complete sense of the effects of unity and prototypicality. While the UL score does provide an indication of the main effect of unity, it does not examine a (decreasing) non-linear model of prototypicality. The UN score does examine a non-linear model of prototypicality; however, the UN score reflects some random variation in the subjects' ratings (i. e., the lowest rated Non-unity/Atypical condition is always used to construct the UN score). Parallel analyses were conducted on each of the three variations of dependent measures (i. e., ratings scores, linear unity









47

Linear Model Aesthetic (or Dervied) Response













(++) (+-) (-+) (--)

Unity/Prototypicality





Nonlinear Model Aesthetic (or Derived) Response












I I


Unity/Prototypicality





FIGURE III-3
LINEAR AND NONLINEAR "UNITY" SCORES







48

scores, and nonlinear unity scores) in order to insure that the results reported were not simply an artifact of a particular analysis scheme.



Hypothesis Testing

The critical test for examining the hypotheses concerning the influence of unity and prototypicality on aesthetic and derived responses involves the Feature A x Feature B interaction. In this design, the Feature A x Feature B interaction is the main effect of unity. This interaction "tests" the competing predictions of the prototypicality and unity hypotheses. If the Unity/Prototypical (++) products of this design receive higher ratings than the Non-Unity/Atypical (+-) and Non-unity/Atypical (-+) products and these in turn are more highly rated than the Unity/Atypical (--) products then the prototypicality explanation of aesthetic (and derived) response would seem to be supported (Hypothesis 2). If, however, the Unity/Prototypical (++) products and the Unity/Atypical (--) products were both rated significantly higher than the Non-unity/Atypical (+-, -+) products then support for the unity hypothesis would be indicated (Hypothesis 1). The schema incongruity explanation of aesthetic response would be supported if the Non-unity/Atypical (+-, -+) products were rated significantly higher than the prototypical products (++) and/or the extremely atypical products (--). If aesthetic response was found to be positively related to atypicality than this would suggest a novelty effect.








49
The means for the product beauty ratings for each of the twenty-seven stimulus sets are presented in Table 111-2. Across the twenty-seven stimulus sets the products that did not exhibit unity (M = 4.06) were rated lower than the products that did exhibit unity (M = 5.12). The analysis of variance presented in Table 111-3 indicates that this difference which is captured in the Feature A x Feature B interaction is significant F (1,184) = 338.16, p < .0001. This supports the unity hypothesis which predicted that aesthetic responses are more positive for objects (products) that exhibit high consistency with the visual organization principle of unity than they are for objects (products) that are not consistent with the unity visual organization principle.

The significance of the Factor A x Factor B interaction (i. e., main effect of unity) was also tested in parallel analyses that utilized the linear unity scores (UL) formed from subjects' ratings of all four versions of each stimulus set [i. e., (Unity/Prototypical + Unity/Atypical) (Non-unity/Atypical + NonUnity/Atypical)] and the nonlinear unity scores (UN) [i. e., (Unity/Atypical) minimum (Non-unity/Atypical)]. The analysis of variance tables for each of these approaches are shown in Table 111-4 and Table 111-5. The unity hypothesis (i. e., Feature A x Feature B interaction) is tested by determining whether or not the intercept is zero. That is, if the intercept were actually zero, what would the probability be of obtaining, by chance alone, a value as large or larger than the one actually obtained? The results reported in Table









50
MEAN BEA TABLE III 2
MEA BEUT RAING AD UJNITY SCORES

Product (++) (- -)(1. Bathroom Scale
Sh 5.30 5.16 5.13 5.18 0.85 0.88***
Tr 4.42 5.42 4.56 5.81 1.38***
TX 4.03 5.80 4.948108 .38**

2. TV Control 48 .1-.90'3
Sb 4.82 4.09*a 4.60 4.95 0.51** 1.1*
Tr 3.98 4.63 4.95 6.2414 22***
Tx 5.36 4. 45 .80.84** 1.3***

3. Flashlight .8* 45* 5.804* 136*
Sh 6.31 3.95*** 3.00*** 3.05*** 1.20*** 0.55**
Tr 5.57 3.74*** 4.41*** 3.97*** 0.70*** 0.61*
TX 5.89 2.85*** 2.34*** 5.15* 2.94**-* 3.00***

4. Lamp
Sb 5.64 5.66 4.36*** 4.36** 0.00 0.39*
Tr 4.65 4.58 5.71 5.92 0.14 1.42***
Tx 5.00 4.84 4.86 6.05 0.68** 2.08***

5. Refrigerator
Sh 6.00 2.00*** 2.40*** 3,95*** 2.80*** 2.20***
Tr 5.31 4.22** 4.19** 5.70 1.28*** 2.03***
Tx 6.03 2.23*** 2.39*** 3.34*** 2.34*** 1.32***

6. Telephone
Sh 5.30 2.70*** 4.29** 3,48*** 09** Q9*
T r 5.4 5.1 6+ .1 ** 4.18*** 0.15 0.21+
Tx574 58* 4,75*** 6.02 0.90)*** 1.41***








51


TABLE III 2 -- continued
MEAN BEAUTY RATINGS

Product (++) (+-) (-+) (--) U U


7. Hair dryer
Sh 6.57 4.98*** 2.85*** 3.30*** 0.90*** 0.61**
Tr 4.64 4.98 4.98 5.42 0.03 0.92***
Tx 6.33 2.75*** 2.56*** 5.17** 3.06*** 2.94***

8. Dresser
Sh 5.08 4.95 4.54* 5.66 0.62*** 1.80***
Tr 5.44 3.56*** 3.69*** 6.61 2.41*** 3.26***
Tx 5.64 3.57*** 4.44** 4.54** 0.99*** 1.42**

9. Clock
Sh 5.72 2.39*** 2.72*** 4.75** 2.65*** 2.58***
Tr 4.92 3.49*** 3.48*** 4.26** 1.07*** 1.03***
Tx 5.17 3.57*** 3.45*** 5.00 1.56*** 1.71***
Totalc 5.34 4.10*** 4.01*** 4.90*** 1.05*** 1.48***

a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
b product/transformation pair.
Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.








52
TABLE III 3
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE RATING SCORE APPROACH

Source DF Type III SS F-Value P > F
Between Suhiect Effects
Version 2 42.9350 1.63 0.1991
Order 1 13.7790 1.05 0.3080
Version Order 2 107.6630 4.08 0.0184
Error Sub (Version Order) 184 2425.7720
Within Subiect Effects
Product 8 1081.1417 40.78 0.0001
Trans 2 165.0833 24.91 0.0001
Prod Version (Residual) 14 566.1260 12.02 0.0001
Prod Order 8 199.3217 7.52 0.0001
Trans Order 2 5.4222 0.82 0.2500
Prod Version Order (Residual) 14 43.1955 0.93 0.2500
Error Sub (Product) 1472 4878.0299
Feature A 1 124.2720 23.97 0.0001
A Version 2 252.2238 24.33 0.0001
A Order 1 10.6083 2.05 0.1542
A Version Order 2 6.1115 0.59 0.5556
Error Sub (A) 184 953.7624
Feature B 1 52.8499 11.86 0.0007
B Version 2 114.6225 12.86 0.0001
B Order 1 21.7458 4.88 0.0284
B Version Order 2 17.8531 2.00 0.1378
Error Sub (B) 184 819.8276
Feature A Feature B 1 1941.8813 338.16 0.0001
A B Version 2 67.4647 5.87 0.0034
A B Order 1 3.1291 0.54 0.4614
A B Version Order 2 5.6170 0.49 0.6140
Error (A B) 184 1056.6273
Prod A B 8 933.9437 53.26 0.0001
Prod A B Version 16 787.4382 22.45 0.0001
Prod A B Order 8 94.0783 5.37 0.0001
Prod A B Version Order 16 35.9705 1.03 0.4255
Error Sub (Prod A B) 1472 3226.3444








53

TABLE III 4
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE LINEAR SCORE APPROACH

Source DF ] Type III SS F-Value P > F
Between Subject Effects
Intercept 1 7767.5254 338.16 0.0001
Version 2 269.8589 5.87 0 0034
Order 1 12.5163 0.54 0.4614
Version Order 2 22.4679 0.49 0.6140
Error Sub (Version Order) 184 4226.5090
Within Suhiect Effects
Product 8 3735.7748 53.26 0.0001
Trans 2 549.8955 31.36 0.0001
Prod Version (Residual) 14 2599.8573 21.18 0.0001
Prod Order 8 376.3133 5.37 0.0001
Trans Order 2 40.2259 2.29 0.2500
Prod Version Order (Residual) 14 103.6562 0.84 0.2500
L Error Sub (Prod) 1472 12905.3777

111-4 and Table 111-5 clearly indicate support for the unity hypothesis. The intercept using the linear unity score (UL) as the dependent measure was significantly different from zero F (1,184) = 338.16, p < 0.0001.6 The intercept as tested using the nonlinear unity score (UN) approach was also significantly different from zero F(1,190) = 473.48, p < 0.001.7 The fact that there is a small, but significant t (190 = 6.04, one-tailed p < .001) difference between the means of the Unity/Prototypical (++) products (M = 5.34) and the



6 The results for the linear unity scores are identical to the results for the
ratings score approach.
7 The difference in the denominator degrees of freedom are due to missing
data that caused several subjects to be eliminated from the "unity score"
analysis.








54

TABLE III 5
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
NONLINEAR UNITY SCORE APPROACH


Source DF -_Type_ III SS F-Value P > F
Between Subjiect Effects ___________ _____ ____Intercept 1 3873.5708 473.48 0.0001
Version 2 T340.8709 20.83 0.0001
Order 1 3.2588 0.40 0.5287
Version Order 2 j 1.3466 0.08 0.9210
Error Sub (Version Order) 190 15416 Within Subject Effects _____ ___Product 8 298.1844 9.67 0.0001
Trans 2 63.3658 8.21 0.0001
Prod Version 14 484.9146 8.98 0.0001
Prod Order 8 123.7531 4.01 0.0100
Trans Order 2 41.8300 5.42 0.0100
Prod Version (Residual) 14 75.8172 1.40 0.2500
Error Sub (Prod) 1520 5861.4233 1_____ ____Unity/Atypical (--) products (M = 4.90) suggests that prototypicality does have

some effect on aesthetic responses.

The analysis of variance presented in Table 111-3 as well as those shown

in Table 111-4 and Table 111-5 also show that there were a number of factors and interactions among factors that resulted in significant effects in this design. Most of these significant effects can be attributed to differences in the strength of the manipulations used to create each stimulus set and the differences across product categories. The effect of Product was significant, F (8,1472) =40.78, p < .0001, as was the effect of Transformation type F (2,1472) 24.91, p < .0001. Since stimulus sets (i. e., Transformation types) are contained in each of the versions of the questionnaire the effects of these







55

factors are also indicated in the Product x Version (Residual) interactions.8 The differential strength of the manipulations across stimulus sets was further examined in order to ensure that the effects were in the right direction across the twenty-seven stimulus sets. This was done by conducting t-tests using the flunity" scores for each of the stimulus sets in order to assess the degree to which the effect of unity was significant and in the proper direction across the twenty-seven stimulus sets despite the differences in the strength of the manipulations. The analysis for the linear unity scores is shown in Table 111-6 (the unity scores were also presented in Table 111-2). In twenty-one out of the twenty-seven stimulus sets the linear unity scores are significantly different from zero (p < .05; p < .01 for nineteen of the stimulus sets). In six cases the interaction scores were not significant and in only one case was a score negative and significant (i. e., in the "wrong" direction).9 The findings are similar when the analysis is conducted using nonlinear unity scores. This analysis is presented in Table 111-7. When the differential strength of the


8 In this design the main effect of Transformation type is examined using
a Latin square type orthogonal fraction of the complete five factor design. This is consistent with the treatment suggested by Winer, Brown, and Michels (1991, pp. 706-711) for Latin squares and related
designs.
9 This negative case occurred for the texture manipulation of the
bathroom scale stimulus set. All three of the bathroom scale product/cases were found to be somewhat problematic because this product was not easily divided into two separate parts (e. g., telephone stimulus handset, base). This was particularly true for the texture manipulation of the bathroom scale (see bathroom scale/texture stimuli
in Appendix B).







56

TABLE III 6
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES LINEAR UNITY SCORES


U Score for
Product -eauty S. E. t-test
1. Bathroom scale
Sh 0.8335 0.2032 0.8201
Tr 0.1212 0.1449 1.6726*
Tx -0.8923 0.4442 -4.0172

2. TV Remote Control
Sh 0.5076 0.3853 2.6345**
Tr 0.4384 0.3913 2.2410*
Tx 0.8409 0.4711 3.5696**

3. Flashlight
Sh 1.2032 0.3527 6.8233***
Tr 0.6969 0.3547 3.9298***
Tx 2.9394 0.4888 12.0267***

4. Lamp
Sh 0.0000 0.2412 0.0000
Tr 0.1385 0.2787 0.9938
Tx 0.6770 0.4214 3.2124**

5. Refrigerator
Sh 2.7955 0.4302 12.9961***
Tr 1.2770 0.4308 5.9281***
Tx 2.3359 0.4761 9.8119***

6. Telephone
Sh 0.9000 0.4309 4.1773***
Tr 0.1485 0.2062 1.4401
Tx 0.9015 0.4160 4.3341***








57

TABLE III 6 -- continued
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES LINEAR UNITY SCORES


U Score for
Product -3eauty S. E. t-test

7. Hair dryer
Sh 0.8940 0.4167 4.2902***
Tr 0.0303 0.3029 0.2000
Tx 3.0539 0.5295 11.5353***

8. Dresser
Sh 0.6137 0.3105 3.9526***
Tr 2.4077 0.4661 10.3311***
Tx 0.9925 0.5844 3.3966***

9. Clock
Sh 2.6462 0.4438 11.9243***
Tr 1.0682 0.4164 5.1308***
Tx 1.5531 0.3905 7.9536***

*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.

manipulations is examined by looking at the difference between the Unity/Atypical product and the lowest rated Non-Unity/Atypical product for each stimulus set, all of the scores are positive and the scores for twenty-five of the stimulus sets are significant at the p < .05 level (twenty-two of these are significant at the p < .01 level or higher). These analyses suggest that even though there are differences in the strengths of the manipulations due to transformation type and product class, the effect of unity is relatively consistent (i. e., significantly positive) across both products and transformations.







58

TABLE III 7
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES
NONLINEAR UNITY SCORES


U Score for
Product Beauty S. E. t-test
1. Bathroom scale
Sh 0.8788 0.1985 4.4272***
Tr 1.3788 0.1988 6.9359***
Tx 0.4308 0.2816 1.5296+

2. TV Remote Control
Sh 1.4091 0.2653 5.3113***
Tr 2.2308 0.2287 9.7551***
Tx 1.3636 0.2981 4.5741***

3. Flashlight
Sh 0.5539 0.2352 2.3551**
Tr 0.6061 0.2678 2.2628*
Tx 3.0000 0.3260 9.2017***

4. Lamp
Sh 0.3939 0.1753 2.2479*
Tr 1.4242 0.2275 6.2592***
Tx 2.0769 0.2768 7.5047***

5. Refrigerator
Sh 2.1969 0.2800 7.8451***
Tr 2.0308 0.2685 7.5629***
Tx 1.3182 0.2663 4.9493***

6. Telephone
Sh 0.9394 0.2678 3.5073**
Tr 0.2121 0.1441 1.4725+
Tx 1.4091 0.2443 5.7688***








59

TABLE III 7 -- Continued
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES NONLINEAR UNITY SCORES

U Score for
Product Beauty S. E. t-test
7. Hair dryer
Sh 0.6061 0.2470 2.4536**
Tr 0.9242 0.2226 4.1516***
Tx 2.9385 0.3375 8.7074***

8. Dresser
Sh 1.8030 0.2416 7.4624***
Tr 3.2615 0.2375 13.7356***
Tx 1.4242 0.4108 3.4666**

9. Clock
Sh 2.5846 0.2556 10.1118***
Tr 1.0303 0.2471 4.1696***
Tx 1.7121 0.2068 8.2777***

*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.


While an interaction of order with the effect of unity (i. e., the Feature

A x Feature B interaction) was not indicated, order was found to interact with

factors such as Product. In part, the effect of order seems to be attributable

to subjects increasing acceptance of the use of the simplified (i. e., line

drawings without color) drawings to depict products. Such an explanation is

suggested by the fact that, in general, when the stimuli for a particular

product category were rated toward the end of the rating task they received

slightly higher ratings than when they were rated at the beginning of the








60
rating task. This was controlled for to some degree by reversing the presentation order of stimuli in the second order.



Perceived Familiarity

Hypothesis 6 predicted that consistency with the visual organization principle of unity was positively related to perceived familiarity. The means for the familiarity ratings for each of the twenty-seven stimulus sets are presented in Table 111-8. The analysis of variance tables for the rating score, linear unity score, and nonlinear unity score approaches are presented in Tables 111-9, 111-10, and III-11, respectively.

The analysis of variance for the subject ratings of familiarity (Table III9) indicates a significant unity effect for perceived familiarity E (1,181) = 254.47, p < .0001. The results for the analysis of variance using linear unity scores were very similar to those observed using subject ratings with the unity effect achieving the same level of significance F (1,181) = 254.47, p < .0001. The analysis of variance for the nonlinear score approach, which is probably the most appropriate analysis in this case since it removes the influence of the Unity/Prototypical cell of the design, also indicates a significant effect of unity on perceived familiarity F (1,188) = 10.65, p < .0013.10 These results suggest that unity can encourage feelings of perceived familiarity.


10 The pattern of interactions observed among Product, Transformation
type, and Order were the same as were discussed for the influence of
unity on aesthetic responses.










TABLE 111 61
MEAN FAMILIARITTRAOiN N UNITY SCORES


Product (++) (+-)
1. Bathroom scale U
Sh 7.62 7.14**a
Tr 7.06 6.97 5.40*** 5Z6*** 0.31**b 0.45*
TX 7.13 6.98 0.025 0.27

2. TV Remote Control 7.05 6.77 6.23* 6.10** 0.03 0.42
Sh 7:6529 5.08**
Tr 6 6.13* 5.41*** 4.94*** 0.97***
TX 6.51 0.65
7.35 6.89 0.53** 1.4,5***
3. Flashlight 5.17*** 5.32*** 5.86*** 1.31*** 1,00***
Sh 7.93 4.80***
Tr 7.49 416*** 4.05*** 1.54***
TX 5.52*** 5.79*** 5.49*** 0.84*** 0.27
7.49 4.52*** 4X*** 6.35*** 2.43*** 0.35
4. Lamp 2.21***
Sh 8.22 5.95*** 3.98*** 3.57*** 0.89*** -0.18
Tr 7.21 6.10*** 6.81* 6.60* 0.45**
TX 7.43 6,38** 6.18** 7.46 0.67**

5- Refrigerator 2.02***
Sh 7.97 3.16*** 3.25*** 3.84*** 2.68*** 0.92***
Tr 7.75 5.52*** 5.28*** 6.61***
TX 8.22 4.21** 4.29*** 1.88*** 1.74***

6. Telephone 4.52*** 2.16*** 0,67**
Sh 7.26 3.84*** 4.77*** 3.41*** 1.10***
Tr 7.16 6.19*** 5.46*** -0:3
TX 6.89 5.52*** 0.50***
6.24** 5.73*** 6.41* ().65*** 0 4**
0.82**








62

TABLE III 8 -- continued
MEAN FAMILIARITY RATINGS AND UNITY SCORES

Product (++) (+_) (-+) (--) U U


7. Hair dryer
Sh 8.05 6.48*** 4.46*** 4.40*** 0.83*** 0.21
Tr 7.05 6.38** 6.54* 6.44** 0.33* 0.52**
Tx 7.67 3.98*** 4.30*** 5.69*** 2.68*** 2.02***

8. Dresser
Sh 7.76 6.62*** 5.98*** 6.29*** 0.73** 0.56*
Tr 7.34 4.87*** 4.97*** 7.51 2.52*** 2.95***
Tx 7.43 5.68*** 5.63*** 6.37* 1.18*** 1.15***

9. Clock
Sh 7.28 4.51*** 4.75*** 6.43** 2.19*** 2.09***
Tr 7.33 5.68*** 5.78*** 6.08*** 0.97*** 0.77**
Tx 7.05 5.44*** 5.76*** 6.86 1.32*** 1.50**

Totalc 7.45 5.53*** 5.34*** 5.79*** 1.19*** 0.94***

a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each

b product/transformation pair.
Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.








63

TABLE III 9
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE RATING SCORE APPROACH


Source DF ] Type III SS I F-Value P > F
Between Sub U0
ject Effe-1-Version 2 73.7530 0.55 0.5789
Order 1 233.9050 3.48 0.0638
Version Order 2 8.6500 0.06 0.9377
Error Sub (Version Order) 181 12174.5330

-Within Subject Effects
Product 8 946.3903 20.96 0.0001
Trans 2 724.2835 64.16 0.0001
Prod Version (Residual) 14 677.6986 8.58 0.0001
Prod. Order 8 38.3970 0.85 0.5582
Trans Order 2 46.4860 4.12 0.0250
Prod Version Order (Residual) 14 189.0356 2.39 0.0100
Error Sub (Product) 1448 8172.4271
Feature A 1 1459.8545 259.23 0.0001
A Version 2 300.8827 27.71 0.0001
A Order 1 7.6979 1.37 0.2439
A Version Order 2 17.3149 1.54 0.2177
Error Sub (A) 181 1019.2983
Feature B 1 919.4761 185.88 0.0001
B Version 2 14.4650 1.46 0.2345
B Order 1 1.4952 0.30 0.5831
B Version Order 2 4.1657 0.42 0.6570
Error Sub (B) 181 895.3306
A B 1 2362.5960 254.47 0.0001
A B Version 2 69.0487 3.72 0.0261
A B Order 1 21.4075 2.31 0.1306
A B Version Order 2 22.0981 1.19 0.3066
Error Sub (A B) 181 1680.4642
Prod A B 8 549.3698 28.33 0.0001
Prod A B Version 16 434.3241 11.20 0.0001
Prod A B Order 8 109.7385 5.66 0.0001
Prod A B Version Order 16 52.1372 1.34 0. 1618
Error Sub (Prod A B) 1448 3510.0931








64

TABLE III 10
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
LINEAR UNITY SCORE APPROACH







Order 1 54.3667 8.05 005
Version Order 2 11.3613 0.84 042
Error Sub (Version Order) 188 1269.5378
Within Subiect Effects __Product 8 727.7061 20.07 000
Trans 2 661.0854 72.93 0.0001
Prod Version (Residual) 14 557.1115 8.78 0.0001
Prod Order 8 60.4733 1.67 0.1016
Trans Order 2 1.2788 0.14 0.2500
LProd Version Order (Residual) 14 84.6541 1.33 0.2500
Error Sub (Prod) 1504 6816.5729 ____ _____








65
TABLE III 11
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE NONLINEAR UNITY SCORE APPROACH


Source DF I Type III SS F-Valuej P > F
Between Suhiect Effects
Intercept 1 9450.3839 254.47 0.0001
Version 2 276.1947 3.72 0.0261
Order 1 85-6300 2.31 0.1306
Version Order 2 88.3924 1.19 0.3066
Error Sub (Version Order) 181 6721.8569
Within Sub:iect Effects
Product 8 2197.4792 28.33 0.0001
Trans 2 395.5843 20.40 0.0001
Prod Version 14 1341.7119 9.88 0.0001
Prod Order 8 438.9539 5.66 0.0001
ans Order 2 9.8527 0.51 0.2500
Prod Version Order (Residual) 14 198.696 1.46 0.2500
Error Sub (Prod) 1448 14040.3723


Derived Responses

The remaining three hypotheses concerned the effect of aesthetic

response on derived responses (i. e., attitude, quality, and price). It was hypothesized that aesthetic responses would influence derived responses and thus products that received more positive aesthetic ratings (i. e., products that exhibit unity) would also be rated higher with regard to liking (i. e., attitude), quality (i. e., perceived quality), and price (i. e., expected price) than would products that received less positive aesthetic ratings. The critical tests for examining these hypotheses are the same as the critical test for examining the influence of unity on aesthetic response except the dependent measures reflect







66

the construct of interest (e. g., attitude, perceived quality). Thus, if the products that received higher (more positive) aesthetic response ratings also receive significantly higher ratings with respect to scales measuring derived responses than this would provide at least some evidence of the broader influence that product aesthetics may exert on consumer behavior.

The analyses and findings concerning attitude, perceived quality, and price expectations will be discussed together since the analyses employed were identical for all three and the results were for the most part parallel. The means for the liking, quality, and price ratings are presented in Table 111-12, Table 111-13, and Table 111-14, respectively. For the most part, the pattern of the means that was observed for the liking and quality derived responses was similar to the pattern observed in the case of aesthetic response. That is, both the Unity/Prototypical and the Unity/Atypical conditions received higher ratings than either of the two Non-unity/Atypical conditions. The pattern observed for the price ratings was somewhat different. The means for the price ratings across stimulus sets contained quite a number of instances where the Unity/Typical (++) product version was the lowest rated version (this was true for eleven of the twenty-seven stimulus sets). In twenty of the stimulus sets the Unity/Atypical (--) was the highest rated product version. One possible explanation for this is that subjects may attribute the unusual or novel appearance of the atypical product versions to expensive "designer










TABLE 111 12 67
MEAN LIKING -Jz

Product (++) RATINGS AND LTNITY SCORES
I. -Bathroom scale V
Sh
rfr 6.31 5.92**a 5.()6**
TX 4.89 5,91 5.22 4.92** 0.12
4.47 6.44 6.49 0.13+b 0. 3***
2. TV Remote Control 5.25 5.28 -0-97 L 71***
Sh 0 56+
Tr 5.37 4.58* 5.09
TX 4.33 4.73 5.44 5.20 0.40* 1.42***
5.97 4.61** 5.00** 7.og 0.62**
3. Flashlight 5.74 2.74***
Sh 7.08 1,68***
Tr 4.09***
6 3.9,5*** 3:,431::: 3.47***
TX 45 1.53***
6:726 2.92*** 2.46*** 4.26*** 0.96*** 0.78** 4. Lamp 5.48** 3.36*** 0.77**
Sh 3.23***
Tr 6.56 5.95* 4.34***
TX 5-08 4.80 4.53*** 0,37**
5.36 5.30 6.02 6.25 0.66**
4.98 6.7 0.26* 1.61**
5. Refrigerator 3 0.91** 2.58***
Sh
Tr 6.97 2 '06*** 2.46*** 3.98***
TX 6.13 4.14*** 4.20*** 3.25*** 2.14***
6.89 2.31*** 6.05 1,96*** 2.55***
6- Telephone 2.37*** 3.40*** 2.72*** 1.36***
Sh
Tr 5-98 3 '03*** 4.20***
TX 6.02 5-65+ 4429*** 3.63*** I.lg*** 0,92**
6.05 5.26** 4.40*** 0.26* 0.44**
4.97** 6.42 1.10***








68

TABLE III 12 -- continued
MEAN LIKING RATINGS AND UNITY SCORES

Product (++) (+-) (-+) (--) UL UN


7. Hair dryer
Sh 7.40 5.45 3.11 3.50 1.09*** 0.71**
Tr 5.22 5.37 5.14 5.74 0.24+ 1.14***
Tx 6.97 3.06 2.80 5.61 3.31*** 3.23***

8. Dresser
Sh 5.45 5.26 4.74 5.86 0.65** 1.85***
Tr 5.89 3.38 3.69 7.03 2.90*** 3.85***
Tx 6.13 3.84 4.68 4.71 1.18** 1.56**

9. Clock
Sh 6.44 2.58 2.81 5.13 3.10*** 2.75***
Tr 5.63 3.74 3.79 4.81 1.39*** 1.45***
Tx 5.69 3.85 3.51 5.75 2.03*** 2.38***

Totalc 5.97 4.38 4.21 5.24 1.30*** 1.73***

a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
b product/transformation pair.
b Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.









69

TABLE III 13
MEAN QUALITY RATINGS AND UNITY SCORES

Product (++) UT- UN) -1. Bathroom scale
Sh 5.28 5.03 5.25 5.39 0.6b 0.92***
Tr 4.53 5.90 4.65 6.11 0.05 1.0*
Tx 4.35 5.81 5.02 5.65 -0.44 0.95***

2. TV Remote Control
Sh 5.24 5.26 5.34 5.37 -0.00 0.68**
Tr 4.03 5.18 5.48 6.66 -0.03 1.77***
TX 5.41 5.18 5.23 5.66 0.31+ 0.77**

3. Flashlight
Sh 5.61 4.95**a 4.90* 5,02* 0.42* 0.70**
Tr 5.07 4.70+ 4.69* 5.00 0.34** 0,69**
Tx 5.42 4.26** 4.16** 5.68 1.24*** 1.83***

4. Lamp
Sh 5.59 5.43 5.03* 5.08+ 0.08 0.53**
Tr 4.39 4.84 5.53 5.82 -0.09 1.26***
TX 4.23 5.39 4.87 6.45 0.19 1.91***

5. Refrigerator
Sh 6.02 3.97*** 4.45*** 5.03** 1.33*** 1.24***
Tr 5.24 4.73* 4.58* 5.66 0.79*** 1.46***
Tx 6.36 4.74*** 4.85*** 5.25** 1.00** 0.85***

6. Telephone
Sh 5.55 4.71** 4.85** 4.77** 0.36* 0.37*
Tr 5.23 5.34 4.79* 4.56** -0.17 0.06
Tx 5.55 5.33 5.13* 5.89 0.53** 1fl1***








70

TABLE III 13 -- continued
MEAN QUALITY RATINGS AND UNITY SCORES

Product (++) (+) (-+) (--) UL UN

7. Hair dryer
Sh 6.39 5.39** 4.57*** 4.59*** 0.50** 0.48*
Tr 4.40 5.08 5.22 5.63 -0.18 0.82**
Tx 5.82 4.00*** 4.03*** 5.81 1.77*** 2.05***

8. Dresser
Sh 5.05 5.29 5.26 5.95 0.22* 1.32***
Tr 4.81 4.34* 4.32* 6.52 1.29*** 2.38***
Tx 5.18 4.70* 4.75 5.64 0.63** 1.44***

9. Clock
Sh 5.13 3.53*** 3.74*** 4.84 1.32*** 1.57***
Tr 4.64 4.10** 4.20* 4.61 0.43* 0.69**
Tx 4.87 3.98*** 4.19** 4.74 0.73*** 0.92***

Totalc 5.16 4.86*** 4.78*** 5.46 0.48*** 1.12***

a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each

b product/transformation pair.
Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.









71

TABLE III 14
MEAN PRICE RATINGS AND UNITY SCORES


Product (++) (+_) (-+) (--) U U

1. Bathroom scale
Sh 4.42 4.57 4.86 4.93 -0.04 1.02***
Tr 3.53 5.09 3.79 5.38 0.01 1.65***
Tx 3.66 5.74 4.38 5.36 -0.55 1.11**

2. TV Remote Control
Sh 4.38 5.22 5.08 5.56 -0.18 0.91***
Tr 3.80 4.89 4.84 6.56 -0.31 2.08***
Tx 4.62 4.92 4.85 5.55 0.21 1.03***

3. Flashlight
Sh 4.49 4.89 5.21 5.46 -0.10 0.91***
Tr 4.62 4.33+a 4.68 4.57 0.06 0.47*
Tx 4.34 4.14 4.17 5.09 0.55**b 1.17***

4. Lamp
Sh 4.45 4.95 5.10 5.27 -0.13 0.90***
Tr 3.53 4.52 4.95 5.28 -0.32 1.05***
Tx 3.80 4.82 4.49 5.74 0.07 1.54***

5. Refrigerator
Sh 5.58 4.81* 4.66** 5.89 0.99** 1.48***
Tr 5.07 4.84 4.93 6.25 0.72*** 1.69***
Tx 5.50 4.58** 4.87* 5.10 0.57** 0.77**

6. Telephone
Sh 5.26 5.11 5.03 5.84 0.44** 1.18***
Tr 4.85 4.97 4.25** 4.33 0.03 0.28*
5.63 5.34 5.01** 6.14 0.69** 1.30***








72

TABLE III 14 -- continued
MEAN PRICE RATINGS AND UNITY SCORES

Product (++) (+.) (-+) (--) U U

7. Hair dryer
Sh 5.17 4.58** 3.92** 4.28** 0.45** 0.75**
Tr 3.91 4.67 4.72 5.58 0.05 1.15***
Tx 5.03 4.44* 4.21** 5.87 1.08*** 1.82***

8. Dresser
Sh 4.30 4.67 5.39 5.98 0.11 1.62***
Tr 4.39 4.10 4.36 6.33 1.08*** 2.25***
Tx 4.65 4.42 4.80 5.32 0.31* 1.14***

9. Clock
Sh 4.61 3.09*** 3.31*** 4.39 1.24*** 1.51***
Tr 3.90 3.33** 3.52+ 3.92 0.46** 0.71**
Tx 4.09 3.20** 3.44** 3.94 0.68** 0.88***


Totalc 4.50 4.60c 4.55 5.33 0.29*** 1.19***

a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each

b product/transformation pair.
Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.








73

products" and thus relate novelty atypicalityy) in general to the higher prices associated with such products.

The analysis of variance results for liking, quality, and price are reported in Table 111-15 to Table 111-23. For all three of the derived responses there was a significant effect of unity. In the case of attitude (i. e., liking ratings) the effect of unity (i. e., Factor A x Factor B interaction) was highly significant across all three variations of the dependent measure: rating scores F (1,185) = 378.04, p < 0.0001; Linear unity scores F (1,185) = 378.04, p < .0001; and pure-unity score F (1,188) = 530.00, p < .0001. That aesthetic responses would have a strong influence on attitudes is not too surprising since many conceptualizations of attitude maintain that attitudes contain an affective component or that affective responses are antecedents of attitudes (Ajzen 1989; Edwards 1990). The effect of unity on (perceived) quality was not as strong as it was for liking but it was still significant except in the analysis that utilized nonlinear unity scores where it was marginally significant: rating scores F (1,179) = 83.54, p < .0001; linear unity scores F (1,179) = 83.54, p < .0001; nonlinear unity scores F (1,188) = 3.74, p < .0545. The effect of unity on ratings of price was significant across all the dependent measure variations: rating scores F (1,179) = 57.67; linear unity scores F (1,179) = 57.67, p <.0001; nonlinear unity scores F (1,188) = 398.57, p < .0001.

Once again, the patterns of interactions observed among Products, Transformation type, and Order for these derived responses were the same as








74

TABLE III 15
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE RATING SCORE APPROACH


Source DF Type III SS F-Value P > F
Between Suhiect Effects
Version 2 46-0270 1.66 0.1937
Order 1 27.3070 1.96 0.1627
Version Order 2 124-4730 4.48 0.0126
Error Sub (Version Order) 185 2571-1570
Within Subiect Effect.
Product 8 1223.6860 40.79 0.0001
Trans 2 166.7490 22.23 0.0001
Prod Version (Residual) 14 547.9980 10.44 0.0001
Prod Order 8 141.8833 4.73 0.0001
Trans Order 2 7.1283 0.95 0.2500
Prod Version Order 14 42.2361 0.80 0.2500
(Residual) 1480 5550.5546
Error Sub (Product)
Feature A 1 350.7149 48.87 0.0001
A Version 2 439.3120 30.61 0.001
A Order 1 21.8458 3.04 0.0827
A Version Order 2 14.4999 1.01 0.3662
Error Sub (A) 185 1327.7566
Feature B 1 136.4246 24.32 0.0001
B Version 2 154.4816 13.77 0.0001
B Orde 1 36.0534 6.43 0.0121
B Version Order 2 25.8407 2.30 0.1028
Error Sub (B) 185 1037.9498
A B 1 2954.8430 378.04 0.0001
A B Version 2 95.1822 6.09 0.0027
A B Order 1 3.3147 0.42 0.5157
A B Version Order 2 8.7269 0.56 0.5732
Error Sub (A B) 185 1446.0027
Prod A B 8 1289.1620 57.08 0.0001
Prod A B Version 16 895-3401 19.82 0.0001
Prod A B Order 8 81.3623 3.60 0.0004
Prod A B Version 16 74.3545 1.65 0.0509
1 Order 1480 4178.2312
L.Kr-ror Sub (Prod A B)








75

TABLE III 16
ANALYSIS OF VARIANCE FOR LIKING'DEPENDENT MEASURE LINEAR UNITY SCORE APPROACH

Source DF Type III SS F-Value P > F
Between Suhiect Effe-+Intercept 1 11819.3718 378.04 0.0001
Version 2 380.7288 6.09 0.0027
Order 1 13.2590 0.42 0.5157
Version'* Order 2 34.9076 0.56 0.5732
Error Sub (Version Order) 185 5784.0106
Within Suhiect Effects
Product 8 5156.6480 57.08 0.0001
Trans 2 537.0898 23.78 0.0001
Prod Version (Residual) 14 3044.2706 19.26 0.0001
Prod Order 8 325.4491 3.60 0.0004
Trans Order 2 32.1250 1.42 0.2500
Prod Version Order (Residual) 14 265.2929 1.67 0.1000
Error Sub (Prod) 1480 16712.9249


were discussed for aesthetic response. Individual t-tests for each of the stimulus sets for all three of the derived responses were conducted to ensure that the effect of unity was significant and in the proper direction across

stimulus sets (refer to Tables 111-12, 13, and 14).

The three hypotheses concerning the influence of aesthetic response on

derived responses all seem to be supported. The products that received higher beauty or aesthetic ratings (i. e., products exhibiting unity) also received higher ratings for liking, quality, and price than did products that received lower aesthetic ratings (i. e., products that did not exhibit unity). Although there seems to be a reduction in the magnitude of the effect of the unity design factor on the derived responses of perceived quality and expected price (as compared with attitudes) the findings do seem to provide evidence that design







76

TABLE III 17
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
NONLINEAR UNITY DIFFERENCE SCORE APPROACH

Source DF IType III SS F-Value- P > F
Between Subject Effects__________________Intercept 1 5141.7559 530.00 0.0001
Version 2 428.1894 22.07 0.0001
Order 1 12.1427 1.25 0.2647
Version Order 2 41.8216 2.16 0.1187
Error Sub (Version Order) 1188 11823.8659 _________Within Subject Effects ___Product 8 352.3060 9.50 0.0001
Trans 2 143.3854 15.47 0.0001
Prod Version (Residual) 14 506.0161 7.80 0.0001
Prod Order 8 146.6168 3.95 0.0001
Trans Order 2 52.9295 5.71 0.0100
Prod Version Order (Residual) 14 100.1989 1.54 0.1000
_Error Sub (Prod) 1504 ,6971.6544 _____ ___factors may indirectly influence "non-aesthetic" product evaluations by means

of aesthetic response.








77
TABLE III 18
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE RATING SCORE APPROACH

Source Type III SS I F-Value P > F
Between Subiect Effects
Version 2 9.127 0.39 0.6782
Order 1 1.608 0.14 0.7116
Version Order 2 32-987 1.41 0.2477
Error Sub (Version Order) 179 2098.955
Within Suhiect Effect-Product 8 460-3010 15.94 0.0001
Trans 2 8.4585 1.17 0.2500
Prod Version (Residual) 14 97.9144 1.94 0.0250
Prod Order 8 84.6168 2.93 0.0030
Trans Order 2 0.8922 0.12 0.2500
Prod Version Order (Residual) 14 102.4896 2.03 0.0250
Error Sub (Product) 1432 5170.5778
Feature A 1 19.6599 4.43 0.0368
A Version 2 111.1728 12.52 0.0001
A Order 1 7.9716 1.80 0.1820
A Version Order 2 0.4826 0.50 0.9471
Error Sub (A) 179 794.7515
Feature B 1 57.4549 15.22 0.0001
B Version 2 81.5176 10.80 0.0001
B Order 1 0.5777 0.15 0.6961
B Version Order 2 10.9047 1.44 0.2386
Error Sub (B) 179 675.6261
A*B 1 401.5016 83.54 0.0001
A B Version 2 22.2117 2.31 0.1021
A B Order 1 0.0029 0.00 0.9804
A B Version Order 2 2.0149 0.21 0.8111
Error Sub (A B) 179 860.2664
Prod A B 8 233.0967 15.44 0.0001
Prod A B Version 16 245.1410 8.12 0.0001
Prod A B Order 8 25.6402 1.70 0.0943
Prod A B Version Order 16 25.6986 0.85 0.6271
Error Sub (Prod A B) 1432 2702.5612








78

TABLE III 19
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE LINEAR UNITY SCORE

Source DF I Type III SS F-Value P > F
Between Suhiect Effects
Intercept 1 1606.0063 83.54 0.0001
Version 2 88.8466 2.31 0.1021
Order 1 0.0116 0.00 0.9804
Version Order 2 8.0595 0.21 0.8111
Error Sub (Version Order) 179 3441.0657
Within Subiect Effects
Product 8 932.3869 15.44 0.0001
Trans 2 172.8437 11.45 0.0001
Prod Version (Residual) 14 807.7205 7.64 0.0001
Prod Order 8 102.5606 1.70 0.0943
Trans Order 2 12.8424 0.85 0.2500
Prod Version Order (Residual) 14 89.9521 0.85 0.2500
_Error Sub (Prod) 1432 10810.2448


TABLE III 20
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE NONLINEAR UNITY SCORE

Source DF I Type III SS F-Value P>F
Between Suhiect Effects
Intercept 1 24.6230 3.79 0.0545
Version 2 243.5989 18.52 0.0001
Order 1 2.4103 0.37 0.5457
Version Order 2 3.1163 0.24 0.7893
Error Sub (Version Order) 188 1236.4512
Within Subject Effects
Product 8 195.0827 7.91 0.0001
Trans 2 180.0369 29.21 0.0001
Prod Version (Residual) 14 270.5720 6.27 0.0001
Prod Order 8 67.3174 2.73 0.0054
Trans Order 2 5.9871 0.97 0.2500
Prod Version Order (Residual) 14 65.0598 1.51 0.1000
Error Sub (Prod) 1504 4635.4572 ------------- L-








79

TABLE III 21
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE RATING SCORE APPROACH

Source DF I Type III SS F-Value P>F
Between Subiect Effects
Version 2 48.662 1.55 0.2149
Order 1 28.494 1.82 0.1795
Version Order 2 45-063 1.44 0.2405
Error Sub (Version Order) 179 2808-080
Within Subiect Effects
Prodifct 8 1089.9719 38.16 0.0001
Trans 2 58.7083 8.22 0.0010
Prod Version (Residual) 14 137.2604 2.75 0.0001
Prod Order 8 182.0655 6.37 0.0001
Trans Order 2 8.0139 1.12 0.2500
Prod Version Order (Residual) 14 76.0911 1.52 0.1000
Error Sub (Product) 1432 5112.3283
Feature A 1 250.9383 56.16 0.0001
A Version 2 83.3609 9.33 0.0001
A Order 1 1.9605 0.44 0.5086
A Version Order 2 4.8154 0.54 0.5844
Error Sub (A) 179 799.8424
Feature B 1 321.3547 86.02 0.0001
B Version 2 130.0014 17.40 0.0001
B Order 1 2.3247 0.62 0.4312
B Version Order 2 11.9936 1.61 0.2037
Error Sub (B) 179 668.6915
A*B 1 193.1558 57.67 0.0001
A B Version 2 26.2598 3.92 0.0216
A B Order 1 0.0024 0.00 0.9785
A B Version Order 2 4.2203 0.63 0.5337
Error Sub (A B) 179 599.4864
Prod A B 8 200.3073 15.86 0.0001
Prod A B Version 16 129.5848 5.13 0.0001
Prod A B Order 8 9.5617 0.76 0.6409
Prod A B Version Order 16 27.3044 1.08 0.3680
Error Sub (Prod A B) 1432 2260.7715








80

TABLE III1-22
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE LINEAR UNITY SCORE

Source DF IType III SS F-Value P > F
Between Subject Effects ___ _____ ____ ___Intercept 1 772.6232 57.67 0.0001
Version 2 105.0393 3.92 0.0216
Order 1 0.0098 0.00 0.9785
Version Order 2 16.8810 0.63 0.5337
Error Sub (Version Order) 179 2397.9457 ____Within Subject Effects __Product 8 801.2291 15.86 0.0001
Trans 2 24.3032 1.92 0.2500
Prod Version (Residual) 14 494.0359 5.59 0.0010
Prod Order 8 38.2466 0.76 0.6409
Trans Order 2 1.6577 0.13 0.2500
Prod Version Order (Residual) 14 107.5598 1.22 0.2500
Error Sub (Prod) 1432 9043.0861 _________TABLE III 23
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE NONLINEAR UNITY SCORE

Source DF Type III SS F-Value P > F
Between Subject Effects___________________Intercept 1 2557.21486 398.57 0.0001
Version 2 188.1044 14.66 0.0001
Order 1 7.6367 1.19 0.2767
Version Order 2 0.9127 0.07 0.9314
Error Sub (Version Order) 188 1206.1973 _____ ___Within Subject Effects__________Product 8 89.9727 4.40 0.0001
Trans 2 3.6682 0.71 0.2500
Prod Version (Residual) 14 86.8686 2.43 0.0100
Prod Order 8 39.4494 1.93 0.0519
Trans Order 2 8.2444 1.61 0.2500
Prod Version Order (Residual) 14 47.5262 1.33 0.2500
Error Sub (Prod) 1504 3842.0345 ------------








81

Summary

Overall, the results strongly support the hypothesis of a positive effect of unity on aesthetic responses (i. e., H1). The Factor A and Factor B interaction (i. e., unity main effect) was significant and this was shown to be the case across product categories and across three different types of product transformations. An effect of prototypicality (H2) on aesthetic response was indicated, but there was not support for the moderate schema incongruity hypotheses (H3 and 114).

The favorable aesthetic responses generated by consistency with the unity visual organization principle was shown to carry over into subjects' attitudes, perceptions of quality, and price expectations for products (i. e., H7, 118, and 119). The Feature A x Feature B interaction was significant for all three derived responses. In addition, unity was shown to influence subjects' feelings of familiarity (i. e., perceived familiarity). Although not predicted, a novelty effect on price expectations seemed to be indicated by the results (H5). These results are summarized in Figure 111-4. In Figure 111-4 it can be seen that the Unity/Atypical product (--) is always rated higher than the Nonunity/Atypical products [Note: MAX refers to the more highly rated of the two Non-unity/Atypical products and MIN refers to the Non-unity/Atypical product that received the lowest ratings].

These results clearly suggest that the unity design factor can significantly influence both aesthetic responses and non-aesthetic product evaluations and thus play an important role in determining consumer behavior. The replication across the nine product categories demonstrates that the effects observed are quite robust. Furthermore, the ability to produce








82

these effects using three very different types of (product) transformations underscores the fact that unity can (and does) have broad application as a design factor.








83











7 .5 --- - - - - - - -

7 --- - - - - - - -



6 -- --- --- ---- --- --- --- ---B eauty
-0-Familiarity
0 ~ Liking

0 Quality

4 .5 --- -- --- --- --- - -- -- P rice





3
(-H-) (MAX) (MIN) (Unity/Prototypic ality





FIGURE 111-4
MEAN AESTHETIC AND DERIVED RESPONSES













CHAPTER rV
EXPERIMENT 2


Overview

Experiment 1 investigated the effects of unity and prototypicality on aesthetic and derived responses by presenting subjects with all four versions of a stimulus set simultaneously and asking them to rate the versions on semantic differential scales that measured the constructs of interest. This experiment is a between-subjects replication of Experiment I in which subjects evaluate only one design from each stimulus set. In this study, product versions were evaluated without the context of other product versions from the same product category. The between-subjects design of this experiment also serves to reduce the possible demand artifacts that maybe involved in a withinsubjects experimental design that tests the influence of unity and prototypicality on aesthetic responses, attitudes, quality perceptions, and price expectations.



Experimental Design

A sub-set of the stimuli that were employed in Experiment 1 were used in this experiment. One transformation was chosen from each product category. The chosen transformation was the one with the strongest linear 84








85

unity score effect in Experiment 1. Although all nine product categories were included, the bathroom scales stimulus set was omitted from subsequent analyses for two reasons. First, this simplified the analyses of the results. Second, this stimulus set was the only reversal of the unity effect observed in Experiment 1. Each product category was represented and an effort was made to select an equal number of each of the three types of manipulations (i. e., shape, texture, and trim). The nine stimulus sets that were employed were the bathroom scale/texture11 T. V. Remote Control/trim, fiashlight'trim, lamp! texture, refrigerator/shape, telephone/shape, hair dryer/texture, dresser/trim, and clock/shape.

The overall design of the experiment was a 2(Feature A) x 2(Feature B) x 8(Products) mixed factorial design. Just as in the first experiment the manipulation of Features A and B resulted in the manipulation of prototypicality and unity. In this experiment, however, each cell of the crossed Feature A and B factors [i. e., product versions: Unity/Prototypical (++), NonUnity/Atypical (+-), Non-Unity/Atypical (-+), Unity/Atypical (--)] was contained in one of four questionnaire versions. That is, each questionnaire contained all eight product categories (and the extra product); each subject received one





This product had proven to be problematic in Experiment 1 but was included for the purpose of maintaining continuity between the product
presentations of the two experiments.








86

of four questionnaire versions and each questionnaire version contained one cell (i. e., product variation, e. g., Unity/Prototypical) of the 2(Feature A)x 2(Feature B) manipulation from each of the product categories (see Appendix D for a diagram of this design). The product versions (e. g., Unity/Prototypical) were distributed evenly throughout the four versions of the questionnaire. Although the three types of transformations were represented across the eight stimulus sets, transformation type was not treated as a factor in this design. The order in which the stimulus products were presented was reversed for half of the subjects.



Experimental Procedure

Two-hundred and forty volunteer subjects enrolled in the Introductory Marketing course at the University of Florida participated in this experiment. The procedure for this experiment was the same as the procedure for Experiment 1 except that in each of the four questionnaire versions subjects saw and rated only one version of each product for each of the eight stimulus sets. The dependent measures were the same measures that were used in Experiment 1.








87

Results and Discussion

The manipulation check for the effects of a change in either Feature A or Feature B on aesthetic responses indicated that, as expected, there was an overall reduction in the ratings of product appearance as compared with the prototypical product version. The average effect of a change in either of these features was weaker than they were in Experiment 1 with the average effect of a change in Feature A being -.425 and Feature B -.281. The effect of a change in Feature A or Feature B from the category prototype for all eight stimulus sets are shown in Table IV-1- In four of the eight cases the change



TABLEIV-1
MANIPULATION CHECK


Product Feature A Feature B

1. T.V. Remote Tr -0.001 0.343

2. Flashlight Tr 0.096 -0.059

3. Lamp Tx -0.862 -0.414

4. Refrigerator Sh -1.352 -0.827

5. Telephone Sh -0.687 -0.327

6. Hair dryer Tx -0.488 -0.798

7. Dresser Tr -0.171 0.019

8. Clock Sh 0.064 -0.182
Mean effect of change across -0.425 -0.281
products I








88

of either Feature A or Feature B resulted in a lower beauty rating. The effects of a change in the product features were mixed for the remaining four stimulus sets. Of these four mixed-effects stimulus sets only one (T. V. remote trim) is consistent with the manipulation check results of Experiment 1. The fact that in three instances feature manipulations that had resulted in a reduction of beauty ratings in Experiment 1 resulted in slightly increased ratings in Experiment 2 would seem to indicate that the strength of the manipulations is, not surprisingly, affected by the context of other product variations (i. e., whether or not the products being rated can be compared to other product variations or brands).



The Influence of Uniky and Prototypicality on Aesthetic and Derived Responses

As was the case in Experiment 1, the critical tests concerning the influence of unity and prototypicality on aesthetic and derived responses involves the Feature A x Feature B interaction (i. e., a main effect of unity). A significant Feature A x Feature B interaction across products would indicate that changes of product features that are made in a way that increases unity [i. e., Unity/Atypical (--)] lead to more positive aesthetic (and derived) responses. Such a result is inconsistent with the prototypicality explanation of aesthetic response which predicts that product versions that are more similar to the category prototype will receive more positive ratings.








89

Analysis

The data were analyzed using a mixed ANOVA with between-subjects factors of questionnaire Version and Order. Products was a within-subjects factor. The analysis was performed on product ratings since the betweensubjects design precluded the use of the unity score approaches at the individual level. Unity scores (i. e., linear and nonlinear scores) may be calculated using the mean product ratings at the aggregate level. However, a nonlinear unity score calculated in this way will tend to understate the effect of unity since the minimum case (i. e., the lowest rated Non-unity/Atypical product variation) may not be the same for each subject. Thus, the minimum Non-unity/Atypical mean will not necessarily be an aggregation of subjects' nonlinear unity scores. This means that the nonlinear unity scores for this between-subjects experiment represent a more conservative test of the influence of unity than the nonlinear unity scores in Experiment 1. The more conservative nonlinear unity scores reported in this experiment will be distinguished from those reported earlier by a prime (') mark.



Hypothesis Testing

The means for the beauty ratings for each of the eight stimulus sets are presented in Table IV-2. The analysis of variance for factors affecting beauty ratings is presented in Table IV-3. The analysis shows that overall the interaction between Features A and B was significant, F (1,1610) = 6.4238,








90
TABLE IV- 2
MEANS FOR BEAUTY RATINGS

Unity/ Non-Unity/ Non-Unity/ Unity/ Typical Atypical Atypical Atypical Product (++) (+-) (-+) () U UN,
1. T. V. Remote Tr 4.36 4.36 4.70 3.88 -0.41 -0.48

2. Flashlight Tr 4.85 4.95 4.79 5.07 0.09 0.28

3. Lamp Tx 5.00 4.14**a 4.59** 4.98 0.63** 0.84
b
4. Refrigerator Sh 4.16 2.80*** 3.33* 3.41 0.61
0.72**
5. Telephone Sh 3.97 3.28 3.64 3.62 0.34
0.35
6. Hair dryer Tx 5.23 4.74* 4.43 4.23 -0.20
0.15
7. Dresser Tr 4.03 3.86 4.05 4.12 0.26
0.12
8. Clock Sh 3.03 3.10 2.85 2.71 -0.14
-0.11
Total 4.33 3.90*** 4.05** 4.00** 0.19* 0.19

a Indicates significance from Prototypical product.
b Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.


p < .025. The linear unity scores formed from the mean product ratings

indicated that six of the eight stimulus sets were in the expected direction (i.

e., positive). The "mean" nonlinear unity scores indicated that five of the eight

stimulus sets were in the expected direction. Follow-up tests for each product

indicated that individually only two products were significant [Lamp F(1,233)

= 8.85, p < .0032; Refrigerator F(1,233) = 7.74, p < .0058]. However, the fact








91

TABLE IV 3
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE


Source DF Type III SS I F-Value P>F
Between Subject Effects
Version 3 10.8722 0.66 0.5788
Order 1 37.9800 6.89 0.0092
Version Order 3 17.0346 1.03 0.3798
Error Sub (Version Order) 230 1267.2469 Within Suhiect Effects
Product 7 814.4207 42.51 0.0001
Feature A 1 4.0309 1.47 0.2500
Feature B 1 26.6562 9.73 0.0100
A B 1 17.5812 6.42 0.0250
Prod Version 18 105.3644 2.13 0.0100
Prod Order 7 42.3589 2.24 0.0285
A Order 1 0.6343 0.23 0.2500
B Order 1 1.7413 0.64 0.2500
A B Order 1 2.7135 0.99 0.2500
ro, V rsion Order 18 51.2698 1.04 0.2500
Error Sub (Product) 1610 4406.5316 I 1


that this interaction was significant for the analysis of all the products indicates that consistency with the unity visual organization principle is

positively related to aesthetic response.

The ANOVA for the ratings of product beauty also indicates that there

were significant effects of Order and Product as well as a significant Products x Version interaction. These effects were addressed in the discussion of Experiment I and the explanation of each of them is basically the same for this experiment. The effect of Products reflects the differential strengths of the Feature A and Feature B manipulations as they were adapted and applied to








92

each product category. An examination of the manipulation check for the impact of changes in product features A and B on beauty ratings (refer to Table IV-1) reveals that there are differences in the strength of the manipulations across the eight products, This is also reflected in the interaction between Products and Version (i. e., cells of the crossed feature A and B "factors" of the design; e. g., Unity/Prototypical).

There were also significant order effects [Order F(1,230) = 6.89, p < .0092; Product Order F(7,1610) = 2.24, p < .0285]. An examination of the means for products for each of the presentation orders suggests that as in Experiment 1, these effects may be attributable to subjects increasing acceptance of the use of simple drawings to depict products.



Derived Responses

The influence of unity and prototypicality on attitude, quality perceptions, and price expectations was also examined. The means for the liking, quality, and price ratings are presented in Table IV-4, Table IV-5, and Table IVA respectively. The patterns exhibited for the liking and quality means are very similar to the pattern exhibited by the beauty means that were presented in the preceding section. Interestingly, the pattern for the overall means for price expectations were similar to those observed in Experiment 1 with the Unity/Atypical (--) product versions being the most highly rated and the Unity/Prototypical (++) product versions being the lowest rated product.









93
TABLE IV- 4
MEANS FOR LIKING RATINGS


Unity/ Non-Unity/ Non-Unity! Unity/ Typical Atypical Atypical Atypical
Product (++) (+-) (-+) (.-) U L Ux
1. T. V. Remote 5.38 5.56 5.81 4.88 -0.56 -0.68

2. Flashlight 5.63 5.76 5.60* 6.12 0.20 0.52

3. Lamp 5.87 4.97*a 5.16* 5.89 0.82** 0.92*
b
4. Refrigerator 5.07 3.65*** 3.90** 3.89** 0.24

0.71*
5. Telephone 4.50 3.91 4.47 4.12 0.21

0.12
6. Hair dryer 6.36 5.21*** 5.10 5.20 0.10

0.63**
7. Dresser 4.82 4.35 4.57 4.58 0.23

0.24
8. Clock 3.24 3.69 3.56 3.26 -0.30
-0.38
Total 5.11 4.63*** 4.77** 474** 0.22*** 0.16

a Indicates significance from Prototypical product.
b Indicates significance of the differences between Unity/Atypical and

Non-unity/Atypical conditions.
Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.


The ANOVAs for the influence of unity and prototypicality on liking, quality, and price are presented in Table IV-7, Table IV-8, and Table IV-9, respectively.

The influence of unity (i. e., the Feature A x Feature B interaction) was significant for liking F(1,1617) = 7.04, p < .001, approached significance for quality F(1,1603) = 3.17, p < .10, and was not significant for price F (1,1603)

= .24 p > .25.




Full Text
TABLE IV 3
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
91
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
10.8722
0.66
0.5788
Order
1
37.9800
6.89
0.0092
Version Order
3
17.0346
1.03
0.3798
Error Sub (Version Order)
230
1267.2469
Within Subiect Effects
Product
7
814.4207
42.51
0.0001
Feature A
1
4.0309
1.47
0.2500
Feature B
1
26.6562
9.73
0.0100
A B
1
17.5812
6.42
0.0250
Prod Version
18
105.3644
2.13
0.0100
Prod Order
7
42.3589
2.24
0.0285
A Order
1
0.6343
0.23
0.2500
B Order
1
1.7413
0.64
0.2500
A B Order
1
2.7135
0.99
0.2500
Prod Version Order
18
51.2698
1.04
0.2500
Error Sub (Product)
1610
4406.5316
that this interaction was significant for the analysis of all the products
indicates that consistency with the unity visual organization principle is
positively related to aesthetic response.
The ANOVA for the ratings of product beauty also indicates that there
were significant effects of Order and Product as well as a significant Products
x Version interaction. These effects were addressed in the discussion of
Experiment 1 and the explanation of each of them is basically the same for this
experiment. The effect of Products reflects the differential strengths of the
Feature A and Feature B manipulations as they were adapted and applied to


APPENDIX D
DIAGRAM OF EXPERIMENT 2


TABLE IV 8
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE
97
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
31.9286
2.16
0.0933
Order
1
52.2380
10.61
0.0013
Version Order
3
26.4376
1.79
0.1498
Error Sub (Version Order)
229
1127.2916
Within Subiect Effects
Product
7
783.7063
37.16
0.0001
Feature A
1
1.0528
0.35
0.2500
Feature B
1
0.2176
0.07
0.2500
A B
1
9.5530
3.17
0.1000
Prod Version
18
83.3248
1.54
0.1000
Prod Order
7
21.0434
1.00
0.4310
A Order
1
2.7504
0.91
0.2500
B Order
1
7.1185
2.36
0.2500
A B Order
1
17.3136
5.75
0.0250
Prod Version Order (Residual)
18
50.2271
0.93
0.2500
Error Sub (Product)
1603
4829.7360
of Products and Order were observed as well as the significant interaction of
Products and Version.
Although the effect of unity on quality and price was not significant in
this case, these findings concerning derived responses are consistent with the
findings reported in Experiment 1. In Experiment 1 there also seemed to be
a reduction in the magnitude of the effect of the unity design factor on
perceived quality and expected price. This reduction in the effect, coupled with
the reduced influence of unity due to the removal of the context provided by


CHAPTER I
INTRODUCTION
There is a growing recognition that product design, and particularly
aesthetic aspects of product design, is emerging as a key marketing element
(Kotler and Rath 1984). Throughout the 1980s companies updated their
manufacturing methods, quality control, distribution networks, customer
service, and labor/management relations in order to compete with other foreign
and domestic manufacturers. Today, as more and more manufacturers are
able to achieve similar levels of price, quality, reliability, and technology,
product appearance is increasingly being acknowledged as the major difference
around which consumers can exercise a choice (Oakley 1990, p. 4).
Product design, or "competitive aesthetics" as it has been called (Reed
1990), is gaining recognition as a strategic activity that companies can use to
gain a sustainable competitive advantage (Kotler and Rath 1984; Whitney
1988). It is being used for every sort of product from Apple computers to La
Croix sparkling water cans. As Bruce Nussbaum pointed out in Business
Week (June 17, 1991, p. 62): "Recently, business has grown increasingly aware
that design sells. U. S. Companies, in particular, are rediscovering that good
design translates into quality products, greater market share, and heftier
profits." Although the technical aspects of products remain vitally important,
1


5
product aesthetics on consumer behavior. The managerial implications of
these findings are also discussed.


126
difference between the two products raises questions about consumers level of
awareness concerning the influence of aesthetics on their product evaluations.
It is hoped that in addition to improving our understanding of the
influence of aesthetics on consumer behavior the research that has been
presented here will serve as a foundation for a theory of consumer aesthetics.
This work takes an important step toward increasing our knowledge of the
relationship between product aesthetics and consumer behavior. The goal of
this research has been to investigate product aesthetics in a manner that
yields the concrete principles governing peoples responses to product designs.
The theory and propositions concerning unity, prototypicality, and aesthetic
response that were presented here lay a foundation for theorizing about the
aesthetic aspects of product design. The results from the three studies indicate
that there are, indeed, factors that systematically and significantly influence
consumers aesthetic responses and product perceptions. These findings
provide empirical evidence that design is not simply a superficial, frivolous
concern, but rather that it is an important variable that can have a significant
impact on consumers responses to products. This should encourage
researchers and managers alike to pay more attention to design issues. It
should also help to more firmly establish product design as a legitimate
marketing interest.


66
the construct of interest (e. g., attitude, perceived quality). Thus, if the
products that received higher (more positive) aesthetic response ratings also
receive significantly higher ratings with respect to scales measuring derived
responses than this would provide at least some evidence of the broader
influence that product aesthetics may exert on consumer behavior.
The analyses and findings concerning attitude, perceived quality, and
price expectations will be discussed together since the analyses employed were
identical for all three and the results were for the most part parallel. The
means for the liking, quality, and price ratings are presented in Table III-12,
Table XII-13, and Table III-14, respectively. For the most part, the pattern of
the means that was observed for the liking and quality derived responses was
similar to the pattern observed in the case of aesthetic response. That is, both
the Unity/Prototypical and the Unity/Atypical conditions received higher
ratings than either of the two Non-unity/Atypical conditions. The pattern
observed for the price ratings was somewhat different. The means for the
price ratings across stimulus sets contained quite a number of instances where
the Unity/Typical (++) product version was the lowest rated version (this was
true for eleven of the twenty-seven stimulus sets). In twenty of the stimulus
sets the Unity/Atypical () was the highest rated product version. One
possible explanation for this is that subjects may attribute the unusual or
novel appearance of the atypical product versions to expensive "designer


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Dressers (chest of drawers')
Dislike 1 23456789 Like
Ugly 123456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 12345 6789 Low Quality
Low Pnce 123456789 High Price
Dislike 1 2 3 4 5 6 7 3 9 Like
t2 gly 1 23456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 123456789 Low Quality
Low Price 123456789 High Pnce
Dislike
2
3
4
5
6 "
8
9
Like
Dislike
>
3
4
5
6
7
8
9
Like
Ugly
2
3
4
5
6 7
8
9
Beautiful
Ugly
2
3
4
5
6
7
5
q
Beautiful
Familiar
-i
3
4
5
6 7
8
9
Unfamiliar
Familiar
"
3
4
5
6
7
$
9
Unfamiliar
High Quality
*>
3
4
5
6 7
8
9
Low Quality
High Quality
2
3
4
5
6
7
8
9
Low Quaiirv
Low Pnce
2
3
4
5
6 7
8
9
High Price
Low Price
2
3
4
5
6
7
8
9
High Price
153


99
prototypicality, seems unlikely since, if that were true, both of the non
unity/atypical (+-, -+) conditions should always be rated higher than the unity
atypical (--) condition and that was not the case.
Perceived Familiarity
The relationship between unity and perceived familiarity was also
examined in this experiment. The means for the familiarity ratings for each
of the eight stimulus sets are presented in Table IV-10. The analysis of
variance for the influence of unity on perceived prototypicality is presented in
Table IV-11. In addition to the significant Product and Version (Residual)
effects that have been discussed in the previous two sections, there is a
significant interaction of Feature A and Feature B which suggests that unity
can encourage feelings of perceived familiarity, F (1,1610) = 28.76, p < .001.
The unity scores for the familiarity ratings indicated that five of the linear
scores and six of the nonlinear unity scores were in the hypothesized direction.
Follow-up tests conducted to examine the effect of unity on perceived famil
iarity for individual products indicated that the Refrigerator F(l,233 = 33.17,
p < .001), Telephone F(l,232) = 5.14,p < .0242), and Clock F(l,232 = 3.29, p <
.0709) were significant. Perceived familiarity may be particularly influenced
by product shape since all three of these stimulus sets utilized the shape
transformation. These results build on those of Experiment 1 to suggest that


88
of either Feature A or Feature B resulted in a lower beauty rating. The effects
of a change in the product features were mixed for the remaining four stimulus
sets. Of these four mixed-effects stimulus sets only one (T. V. remote trim)
is consistent with the manipulation check results of Experiment 1. The fact
that in three instances feature manipulations that had resulted in a reduction
of beauty ratings in Experiment 1 resulted in slightly increased ratings in
Experiment 2 would seem to indicate that the strength of the manipulations
is, not surprisingly, affected by the context of other product variations (i. e.,
whether or not the products being rated can be compared to other product
variations or brands).
The Influence of Unity and Prototypicality on Aesthetic and Derived Responses
As was the case in Experiment 1, the critical tests concerning the
influence of unity and prototypicality on aesthetic and derived responses
involves the Feature A x Feature B interaction (i. e., a main effect of unity).
A significant Feature A x Feature B interaction across products would indicate
that changes of product features that are made in a way that increases unity
[i. e., Unity/Atypical (--)] lead to more positive aesthetic (and derived)
responses. Such a result is inconsistent with the prototypicality explanation
of aesthetic response which predicts that product versions that are more
similar to the category prototype will receive more positive ratings.


Copyright 1993
by
Robert W. Veryzer, Jr.


105
Stimulus Materials
The same sub-set of product drawings that were employed in
Experiment 2 were used in this experiment and thus the manipulations of
unity and prototypicality were the same as before. In this study, however,
each product version [i. e., "target" product, either Unity/Prototypical (++),
Unity/Atypical (--), or Non-unity/Atypical (+-)] was shown with a written
product description and was rated against a "control" product. The control
product was one of the Non-unity/Atypical product versions with a very
favorable product description (the Non-unity/Atypical condition for each
stimulus set that received higher beauty ratings was selected for use as the
control product). The product descriptions developed for the target products
were less favorable than the product descriptions for the control products.
"Strong" (i. e., favorable) product descriptions dominated "weak" (i. e., less
favorable) product descriptions on two of five or six attributes. The product
descriptions, which were developed from catalog descriptions of products, were
pilot tested (without the product drawings) in order to establish that the
"strong" descriptions were, in fact, perceived as being stronger than the "weak"
product descriptions. The strong and weak product descriptions are presented
in Appendix E.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Lamps
Dislike 1
3
4
5
6 7
8
Q
Like
Dislike
2 3
4 5
6
7 S
9
Ugly 1
*7
3
4
5
6 7
8
9
Beautiful
Ugly
2 3
4 5
6
7 8
9
Familiar 1
2
3
4
5
6 7
8
9
Unfamiliar
Familiar
2 3
4 5
6
7 8
9
High Qualirv 1
3
4
5
6 7
8
9
Low Quality
High Qualirv
2 3
a 5
t
7 8
9
Low Pnce 1
3
5
6 7
8
9
High Pnce
Low Price
4 5
6
7 8
9

Dislike
2
3 4
5
6 7
8
9
Like
Dislike
;
345
5
7 8
Ugly
2
3 4
5
6 7
8
9
Beautiful
Ugly
2
3 4 5
6
7 8
rammer
2
3 4
5
6 7
8
9
Unfamiliar
Familiar
;
3 4 5
6
7 S
High Quality
2
3 -
5
7
8
9
Low Quality
High Quality
2
3 4 5
6
7 8
Low Pnce
2
3 4
5
6 7
8
9
High Price
Low Price
;
3 4 5
6
7 8
Like
Beautiful
Unfamiliar
Low Qualify
High Price
Like
Beautiful
Unfamiliar
Low Qualirv
High Pnce
139


Experimental Procedure 108
Results and Discussion 108
Summary 114
VI GENERAL DISCUSSION 116
APPENDICES
A LIST OF PRODUCTS EXPLORED FOR USE AS
STIMULI 127
B STIMULUS SETS 129
C DIAGRAM OF EXPERIMENT 1 157
D DIAGRAM OF EXPERIMENT 2 159
E STRONG AND WEAK PRODUCT DESCRIPTIONS 161
REFERENCE LIST 164
BIOGRAPHICAL SKETCH 171
v


93
TABLE IV 4
MEANS FOR LIKING RATINGS
Product
Unity/
Typical
(++)
Non-Unity/
Atypical
(+-)
Non-Unity/
Atypical
(-+)
Unity/
Atypical
(-)
UL
UN
1. T. V. Remote
5.38
5.56
5.81
4.88
-0.56
-0.68
2. Flashlight
5.63
5.76
5.60*
6.12
0.20
0.52
3. Lamp
5.87
4.97*a
5.16*
5.89
0.82**
b
0.92*
4. Refrigerator
5.07
3.65***
3.90**
3.89**
0.71*
0.24
5. Telephone
4.50
3.91
4.47
4.12
0.12
0.21
6. Hair dryer
6.36
5.21***
5.10
5.20
0.63**
0.10
7. Dresser
4.82
4.35
4.57
4.58
0.24
0.23
8. Clock
3.24
3.69
3.56
3.26
-0.38
-0.30
Total
5.11
4.63***
4.77**
4.74**
0.22***
0.16
a Indicates significance from Prototypical product.
b Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.
The ANOVAs for the influence of unity and prototypicality on liking, quality,
and price are presented in Table IV-7, Table IV-8, and Table IV-9, respectively.
The influence of unity (i. e., the Feature A x Feature B interaction) was
significant for liking F(l,1617) = 7.04, p < .001, approached significance for
quality F(l,1603) = 3.17, p < .10, and was not significant for price F (1,1603)
= .24 p > .25.


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
THE INFLUENCE OF UNITY AND PROTOTYPICALITY ON
AESTHETIC RESPONSES TO NEW PRODUCT DESIGNS
By
Robert W. Veryzer, Jr.
August 1993
Chairman: Dr. J. Wesley Hutchinson
Major Department: Marketing
This dissertation investigates the influence of the aesthetic aspects of
product appearance on consumers product preferences and product
evaluations. Two factors, unity and prototypicality, are identified and
discussed as being the principal factors that influence consumers aesthetic
responses to product designs. The research provides a basis for a theory of
consumer aesthetics and has implications for new product development,
product quality, and marketing strategy.
The dissertation postulates that consumers responses to product designs
are influenced by the designs consistency with the visual organization
principle of unity (i. e., a congruity among the elements of a design) and its
level of prototypicality (i. e., familiarity; shared features with the category
vi


110
of the previous two experiments, that is, products that exhibit unity are more
highly rated than products that do not exhibit unity, then they would support
the conclusion that the effect of unity on product evaluations is quite robust.
The means for the liking ratings are presented in Table V 1 and the
ANOVA for the effect of unity on subjects attitudes toward the products is
presented in Table V 2. The analysis indicates that there is a significant
effect of Cell, F (2,960) = 4.52, p < .025, which refers to the three target
product versions (Unity/Prototypical, Non-unity/Atypical, or Unity/Atypical)
seen by the subjects in each condition (i. e., questionnaire version) saw. The
effect of unity is also present in the effect of Version, F(2,120) = 3.43, p <
.0357), since the product version (i. e., Unity/Prototypical, Non-unity/Atypical,
and Unity/Atypical) for each product category are contained in the three
questionnaire versions. Contrasts between the different cells (i. e., product
versions) of this design indicated that the difference between the mean for the
Unity/Atypical products (M = 6.06) and the mean for the Non-unity/Atypical
products (M = 5.67) was significant, F(l,120) = 4.79, p < .0306).^ Although
the nonlinear unity scores calculated from the mean product ratings indicated
that seven of the nine were in the expected (i. e., positive) direction, follow-up
tests on the individual products indicated that the contrast between the
15
The difference between the means for the Unity/Prototypical products
(M = 5.90) and the Unity/Atypical products (M = 6.06) was not
significant.


81
Summary
Overall, the results strongly support the hypothesis of a positive effect
of unity on aesthetic responses (i. e., HI). The Factor A and Factor B
interaction (i. e., unity main effect) was significant and this was shown to be
the case across product categories and across three different types of product
transformations. An effect of prototypicality (H2) on aesthetic response was
indicated, but there was not support for the moderate schema incongruity
hypotheses (H3 and H4).
The favorable aesthetic responses generated by consistency with the
unity visual organization principle was shown to carry over into subjects
attitudes, perceptions of quality, and price expectations for products (i. e., H7,
H8, and H9). The Feature A x Feature B interaction was significant for all
three derived responses. In addition, unity was shown to influence subjects
feelings of familiarity (i. e., perceived familiarity). Although not predicted, a
novelty effect on price expectations seemed to be indicated by the results (H5).
These results are summarized in Figure III-4. In Figure III-4 it can be seen
that the Unity/Atypical product () is always rated higher than the Non
unity/Atypical products [Note: MAX refers to the more highly rated of the two
Non-unity/Atypical products and MIN refers to the Non-unity/Atypical product
that received the lowest ratings].
These results clearly suggest that the unity design factor can
significantly influence both aesthetic responses and non-aesthetic product
evaluations and thus play an important role in determining consumer
behavior. The replication across the nine product categories demonstrates that
the effects observed are quite robust. Furthermore, the ability to produce


89
Analysis
The data were analyzed using a mixed ANOVA with between-subjects
factors of questionnaire Version and Order. Products was a within-subjects
factor. The analysis was performed on product ratings since the between-
subjects design precluded the use of the unity score approaches at the
individual level. Unity scores (i. e., linear and nonlinear scores) may be
calculated using the mean product ratings at the aggregate level. However, a
nonlinear unity score calculated in this way will tend to understate the effect
of unity since the minimum case (i. e., the lowest rated Non-unity/Atypical
product variation) may not be the same for each subject. Thus, the minimum
Non-unity/Atypical mean will not necessarily be an aggregation of subjects
nonlinear unity scores. This means that the nonlinear unity scores for this
between-subjects experiment represent a more conservative test of the
influence of unity than the nonlinear unity scores in Experiment 1. The more
conservative nonlinear unity scores reported in this experiment will be
distinguished from those reported earlier by a prime () mark.
Hypothesis Testing
The means for the beauty ratings for each of the eight stimulus sets are
presented in Table IV-2. The analysis of variance for factors affecting beauty
ratings is presented in Table IV-3. The analysis shows that overall the
interaction between Features A and B was significant, F (1,1610) = 6.4238,


TABLE III 13 continued
MEAN QUALITY RATINGS AND UNITY SCORES
70
Product
(++)
(+-)
(-+)
(-)
UL
%
7. Hair dryer
Sh
6.39
5.39**
4 57***
4.59***
0.50**
0.48*
Tr
4.40
5.08
5.22
5.63
-0.18
0.82**
Tx
5.82
4.00***
4.03***
5.81
]_ 77***
2.05***
8. Dresser
Sh
5.05
5.29
5.26
5.95
0.22*
1.32***
Tr
4.81
4.34*
4.32*
6.52
1.29***
2.38***
Tx
5.18
4.70*
4.75
5.64
0.63**
9. Clock
Sh
5.13
3.53***
2 7l_***
4.84
1.32***
1.57***
Tr
4.64
4.10**
4.20*
4.61
0.43*
0.69**
Tx
4.87
3.98***
4.19**
4.74
0.73***
0.92***
Totalc
5.16
4.86***
4.78***
5.46
0.48***
02***
Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
Indicates unity score is significantly different from zero.
Statistical significance based on ANOVA results reported in the text.
Indicates differences significant at p < .001 in one-tailed-tests.
Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
Indicates differences significant at p < .1 in one-tailed-tests.


77
TABLE III 18
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
9.127
0.39
0.6782
Order
1
1.608
0.14
0.7116
Version Order
2
32.987
1.41
0.2477
Error Sub (Version Order)
179
2098.955
Within Subiect Effects
Product
8
460.3010
15.94
0.0001
Trans
2
8.4585
1.17
0.2500
Prod Version (Residual)
14
97.9144
1.94
0.0250
Prod Order
8
84.6168
2.93
0.0030
Trans Order
2
0.8922
0.12
0.2500
Prod Version Order (Residual)
14
102.4896
2.03
0.0250
Error Sub (Product)
1432
5170.5778
Feature A
1
19.6599
4.43
0.0368
A Version
2
111.1728
12.52
0.0001
A Order
1
7.9716
1.80
0.1820
A Version Order
2
0.4826
0.50
0.9471
Error Sub (A)
179
794.7515
Feature B
1
57.4549
15.22
0.0001
B Version
2
81.5176
10.80
0.0001
B Order
1
0.5777
0.15
0.6961
B Version Order
2
10.9047
1.44
0.2386
Error Sub (B)
179
675.6261
A B
1
401.5016
83.54
0.0001
A B Version
2
22.2117
2.31
0.1021
A B Order
1
0.0029
0.00
0.9804
A B Version Order
2
2.0149
0.21
0.8111
Error Sub (A B)
179
860.2664
Prod A B
8
233.0967
15.44
0.0001
Prod A B Version
16
245.1410
8.12
0.0001
Prod A B Order
8
25.6402
1.70
0.0943
Prod A B Version Order
16
25.6986
0.85
0.6271
Error Sub (Prod A B)
1432
2702.5612


124
or influenced by cultural forces (Bornstein, Ferdinandsen, and Gross 1981;
Segall 1976). Although the acquired principles are undoubtably modified from
time to time, aesthetic response preference patterns are fairly stable within
individuals over time (Huber and Holbrook 1981). Similarities and differences
observed between individuals may be due to similarities or differences in the
physical, socio-economic, or cultural environment in which people live.
Although on the surface it would seem that aesthetic response operates
on a conscious level involving rational evaluation, it has been suggested that
the intervening cognitive response system may operate below the threshold of
consciousness at a subconscious or pre-conscious level (Holbrook and
Hirschman 1982). Aesthetic response seems to operate on at least two levels -
conscious awareness and nonconscious awareness (Holbrook and Hirschman
1982). The conscious level involves attending to the object and registering
feelings or appreciation of the object. This level is the conscious registering of
the unconscious input and is therefore primarily a function of the nonconscious
awareness level (Zajonc 1980). The nonconscious level of awareness involves
perceiving the object and determining its consistency with rules (e. g., design
principles) which have been acquired. Regardless of whether these rules are
innate or learned nonconsciously, they often seem to be applied without
conscious awareness. Thus, while differences in the appearance of products
may be readily perceived (i. e., conscious awareness), the underlying process


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Lamps
Dislike 123-i5s789 Like
Ugly 123456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 123456789 Low Quality
Low Price 123456789 High Price
Dislike ]
12 3 4 5
6 8
9
Like
Ugly 1
2 3 4 5
0 7 8
9
Beautiful
Familiar 1
2 3 4 5
s 7 S
9
Unfamiliar
High Qxiah tv 1
2 3 4 5
6 7 8
9
Low Quality
Low Pnce 1
2 3 4 5
6 7 8
9
High Pnce
Dislike
i 2
3
*
5
6 7
S
9
Like
Ugly 1
2
3
4
5
6 7
8
9
Beautiful
Familiar 1
2
3
4
5
6 7
8
9
Unfamiliar
High Quality 1
-
3
4
5
6 7
8
9
Low Quality
Low Price 1
2
3
4
5
6 7
s
9
High Price
Dislike !
3 4
5
5 ~
S 9
Like
Ugly 1
3 4

6 ~
5 9
Beautiful
Familiar 1
i
3 4
5
6
8 9
Unfamiliar
High Quality 1
-
3 4
5
6
? 9
Low Quality
Low Price 1
2
3 4
5
6
8 9
High Price
141


80
TABLE III 22
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE
LINEAR UNITY SCORE
Source
DF
Type III SS
F-Value
P > F
Between Subject Effects
Intercept
1
772.6232
57.67
0.0001
Version
2
105.0393
3.92
0.0216
Order
1
0.0098
0.00
0.9785
Version Order
2
16.8810
0.63
0.5337
Error Sub (Version Order)
179
2397.9457
Within Subiect Effects
Product
8
801.2291
15.86
0.0001
Trans
2
24.3032
1.92
0.2500
Prod Version (Residual)
14
494.0359
5.59
0.0010
Prod Order
8
38.2466
0.76
0.6409
Trans Order
2
1.6577
0.13
0.2500
Prod Version Order (Residual)
14
107.5598
1.22
0.2500
Error Sub (Prod)
1432
9043.0861
TABLE III 23
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE
NONLINEAR UNITY SCORE
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
2557.21486
398.57
0.0001
Version
2
188.1044
14.66
0.0001
Order
1
7.6367
1.19
0.2767
Version Order
2
0.9127
0.07
0.9314
Error Sub (Version Order)
188
1206.1973
Within Subiect Effects
Product
8
89.9727
4.40
0.0001
Trans
2
3.6682
0.71
0.2500
Prod Version (Residual)
14
86.8686
2.43
0.0100
Prod Order
8
39.4494
1.93
0.0519
Trans Order
2
8.2444
1.61
0.2500
Prod Version Order (Residual)
14
47.5262
1.33
0.2500
Error Sub (Prod)
1504
3842.0345


TABLE III 12 continued
MEAN LIKING RATINGS AND UNITY SCORES
68
Product
(++)
(+-)
(-+)
()
ul
uN
7. Hair dryer
Sh
7.40
5.45
3.11
3.50
1.09***
0.71**
Tr
5.22
5.37
5.14
5.74
0.24+
1 24***
Tx
6.97
3.06
2.80
5.61
3.31***
3.23***
8. Dresser
Sh
5.45
5.26
4.74
5.86
0.65**
1.85***
Tr
5.89
3.38
3.69
7.03
2.90***
3.85***
Tx
6.13
3.84
4.68
4.71
1.18**
1.56**
9. Clock
Sh
6.44
2.58
2.81
5.13
3.10***
2.75***
Tr
5.63
3.74
3.79
4.81
1.39***
1 45***
Tx
5.69
3.85
3.51
5.75
2.03***
2.38***
Total0
5.97
4.38
4.21
5.24
1.30***
1 73***
a
b
c
***
*
+
Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
Indicates unity score is significantly different from zero.
Statistical significance based on ANOVA results reported in the text.
Indicates differences significant at p <
Indicates differences significant at p <
.001 in one-tailed-tests.
.01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
Indicates differences significant at p < .1 in one-tailed-tests.


TABLE IV 7
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
96
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
34.3446
1.27
0.2844
Order
1
38.0624
4.23
0.0408
Version Order
3
95.6050
3.54
0.0154
Error Sub (Version Order)
231
2077.8006
Within Subiect Effects
Product
7
1162.2629
47.72
0.0001
Feature A
1
6.2858
1.81
0.2500
Feature B
1
30.7590
8.84
0.0100
A B
1
24.4887
7.04
0.0010
Prod Version
18
160.0874
2.56
0.2500
Prod Order
7
24.8918
1.02
0.4135
A Order
1
0.6051
0.17
0.2500
B Order
1
0.4279
0.12
0.2500
A B Order
1
14.6689
4.22
0.0500
Prod Version Order (Residual)
18
72.1990
1.15
0.2500
Error Sub (Product)
1617
5626.2035
product indicated that individually three products were significant for the
influence of unity on liking [Lamp F(l,234) = 9.26, p < .0026; Refrigerator
F(l,234) = 4.84 p< .0287; Hair dryer F(l,234) = 6.96, p < .0089, and T. V.
Remote was marginally significant F(l,234) = 3.57, p < .0602], two products
were significant for the influence of unity on quality [Telephone F( 1,231) =
5.40, p < .0210; Dresser F(l,232) = 4.36, p < .0378; and Lamp was marginally
significant F(l,232) = 3.20, p < .0747], and none of the products was significant
with respect to the influence of unity on price. Once again, significant effects


51
TABLE III 2 continued
MEAN BEAUTY RATINGS
Product
(++)
(+-)
(-+)
(-)
UL
uN
7. Hair dryer
Sh
6.57
4 gg***
2.85***
3.30***
0.90***
0.61**
Tr
4.64
4.98
4.98
5.42
0.03
0.92***
Tx
6.33
2.75***
2.56***
5.17**
3.06***
2.94***
8. Dresser
Sh
5.08
4.95
4.54*
5.66
0.62***
1.80***
Tr
5.44
3.56***
3.69***
6.61
2 4^***
3.26***
Tx
5.64
3.57***
4 44**
4.54**
0.99***
1.42**
9. Clock
Sh
5.72
2.39***
2.72***
4.75**
2.65***
2.58***
Tr
4.92
3.49***
3.48***
4.26**
1 07***
1.03***
Tx
5.17
3.57***
3.45***
5.00
1.56***
1 71***
Totalc
5.34
4.10***
4.01***
4 90***
1.05***
1.48***
a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
b Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.


TABLE V 1
LIKING
111
Product
Unity/
Prototypical
Non-Unity/
Atypical
Unity/
Atypical
UN
1. Bathroom scale
5.52
5.86
6.40
0.54
2. T. V. Remote
6.30
6.02
6.40*a
0.38
3. Flashlight
5.57
5.49
5.98
0.49
4. Lamp
6.64
6.05
6.19
0.14
5. Refrigerator
5.37
4 36***a
6.55
2
6. Telephone
5.88
4.74**
5.40**
0.66
7. Hair dryer
5.79
7.00
5.93
-1.07
8. Dresser
6.14
5.29
6.00
0.71
9. Clock
5.93
6.23
5.71
-0.52
Total
5.90
5.67**
6.06
0.39*
a Indicates significance from Prototypical product.
D Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.


122
by a company to create an entire "line" of products that target different
consumer segments at different price points, levels of quality, etc., while
utilizing the same internal components. While this would be an extreme use
(i. e., exploitation) of product design, the more likely and desirable use is to
employ design to better communicate the nature (i. e., positioning and proper
use) of a product to consumers.
An important issue to be investigated further is the relationship between
unity and prototypicality. The findings of Experiments 1 and 2 indicated that
consistency with the unity visual organization principle is positively related to
perceived familiarity. One possible explanation for this is that there may be
a natural tendency in people toward organization principles such as unity and
symmetry. Whether the tendency of the viewer to look for organization and
prefer it to chaos is innate or learned is still the subject of debate. The debate
between biological and cultural determinism in aesthetics is but a small part
of a larger ongoing debate on the relative roles of "nature" and "nurture" in
human behavior. In all likelihood, the tendency to perceive visual elements in
certain ways or as integrated wholes involves both biological and cultural
influences. The natural tendency toward organization with respect to
principles such as unity and symmetry may stem (at least in part) from our
earliest encounters with ourselves and our environments (Johnson 1987). Our
bodies (i. e., forms) and those of the animals and plants that we encounter
exhibit the "regularities" that we have come to know as unity, symmetry,


37
(e. g., inconsistent perspective) and unintended degradation (e. g., blurry edges,
cloudy surfaces) into a stimulus.
The need for a method that afforded the construction of previously
"uncreated" (i. e., nonexistent) products that were markedly different from
existing products and the need to control for extraneous aesthetic influences
led to the use of line drawings. The use of drawings does entail a trade-off of
realism for "producibility" and greater experimental control. While this
reduction in realism is unfortunate, it is not uncommon for research that is
conducted in a laboratory setting. The use of drawings was also more practical
for the purpose of reproducing the stimuli for inclusion in the booklets used in
the experiment. This method of constructing stimuli necessitated the selection
of products that could be clearly communicated through simple line drawings.
A number of products in a wide variety of product categories were
examined in catalogs and in stores in order to determine their suitability for
use as stimuli in this experiment. An initial series of product, "studies" (i. e.,
sketches) was done to explore ways of manipulating products along the unity
and prototypicality dimensions. A number of products that seemed to meet the
four requirements were then selected for further development in product
2
The possibility of modifying computer scanned images was explored.
This approach proved to be unsatisfactory for the purposes of creating
the stimuli for this particular experiment because the radical changes
that had to be made to the computer scanned images introduced
aesthetic confounds into the stimuli and often resulted in severe
degradation of the image.


78
TABLE III 19
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE
LINEAR UNITY SCORE
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
1606.0063
83.54
0.0001
Version
2
88.8466
2.31
0.1021
Order
1
0.0116
0.00
0.9804
Version Order
2
8.0595
0.21
0.8111
Error Sub (Version Order)
179
3441.0657
Within Subiect Effects
Product
8
932.3869
15.44
0.0001
Trans
2
172.8437
11.45
0.0001
Prod Version (Residual)
14
807.7205
7.64
0.0001
Prod Order
8
102.5606
1.70
0.0943
Trans Order
2
12.8424
0.85
0.2500
Prod Version Order (Residual)
14
89.9521
0.85
0.2500
Error Sub (Prod)
1432
10810.2448
TABLE III 20
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE
NONLINEAR UNITY SCORE
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
24.6230
3.79
0.0545
Version
2
243.5989
18.52
0.0001
Order
1
2.4103
0.37
0.5457
Version Order
2
3.1163
0.24
0.7893
Error Sub (Version Order)
188
1236.4512
Within Subiect Effects
Product
8
195.0827
7.91
0.0001
Trans
2
180.0369
29.21
0.0001
Prod Version (Residual)
14
270.5720
6.27
0.0001
Prod Order
8
67.3174
2.73
0.0054
Trans Order
2
5.9871
0.97
0.2500
Prod Version Order (Residual)
14
65.0598
1.51
0.1000
Error Sub (Prod)
1504
4635.4572


Please
products
the
proposed
Bathroom scalp.:
131


14
harmony, repetition, etc.) can significantly influence perceptions. The
preceding discussion suggests the following hypothesis:
HI: Aesthetic responses are more positive for objects (products)
exhibiting high consistency with the visual organization principle
of unity than they are for objects (products) that are not
consistent with this visual organization principle.
Prototypicality
Another factor that seems to exert an influence on aesthetic response is
that of experience. Familiarity has been shown to lead to positive affect
(Kunst-Wilson and Zajonc 1980). This would seem to suggest that more typical
or familiar items should be better liked (Loken and Ward 1990). Typicality,
or prototypicality, is concerned with the degree to which an object is
representative of a category. A prototype can be defined as the central
representation of a category, as possessing the average or mean value of the
attributes of that category and as representing the averaged members of the
class (Langlois and Roggman 1990; Rosch 1978). According to the prototypi
cality view, people respond most favorably to objects that are highly protypical
and less favorably to objects that are less prototypical (Glass and Holyoak
1986, p. 170; Langlois and Roggman 1990). The prototypicality explanation of
preference maintains that more prototypical examples tend to be better liked
(Loken and Ward 1990; Nedungadi and Hutchinson 1985). Although prototypi-


Control Stimulus
107
Target Stimulus
"Weak" Product Description
"Strong" Product Description
Unity/Prototypical
Non-Unity/Atypical
(++)
(-+)
(Control)
Unity/Atypical
Non-Unity/Atypical
(-)
(-+)
(Control)
Non-Uni ty/Atypical
Non-Unity/Atypical
(+-)
(-+)
(Control)
Design: KTarget product) x 3(Unity/Prototypical, Unity/Atypical,
Non-unity/Atypical) x 9(Products) Mixed Design.
FIGURE V 1
EXPERIMENT 3


REFERENCE LIST
Ajzen, leek (1989), "Attitude Structure and Behavior," in Attitude Structure
and Function, eds. Anthony R. Pratkanis, Steven J. Breckler, and
Anthony G. Greenwald, Hillsdale, NJ: Lawrence Erlbaum Associates,
Publishers, pp. 241-274.
Auld, Frank (1981), "A Theory Deriving Preference from Conflict," in Advances
in Intrinsic Motivation and Aesthetics, ed. Hy I. Day, New York: Plenum
Press, pp. 203-224.
Bell, Stephen S., Morris B. Holbrook, and Michael R. Solomon (1991),
"Combining Esthetic and Social Value to Explain Preferences for
Product Styles with the Incorporation of Personality and Ensemble
Effects," The Journal of Social Behavior and Personality. Vol. 6, No. 6,
pp. 243-273.
Berkowitz, Marvin (1987), "The Influence of Shape on Product Preferences,"
Advances in Consumer Research. Vol. 14, p. 559.
Berlyne, D. E. (1970), "Novelty, Complexity, and Hedonic Value," Perception
and Psychophysics. Vol. 8, pp. 279-286.
Berlyne, D. E. (1971), Aesthetics and Psychology. New York: Appleton-
Century-Crofts.
Berlyne, D.E. (1974), Studies in the New Experimental Aesthetics. New York:
John Wiley and Sons.
Bornstein, Marc H., Kay Ferdinandsen, and Charles G. Gross (1981),
"Perception of Symmetry in Infancy," Developmental Psychology.
Vol. 17, No. 1, pp. 82-86.
Carruthers, Margaret (1970), "Color-Form Dominance and Memory for Color,"
M. A. thesis, Psychology Dept., University of Glasgow.
Ching, Francis D. K. (1979), Architecture: Form, Space, and Order. New York:
Van Nostrand Reinhold Company.
164


115
presented here indicate that while this certainly does seem to be the case for
attitudes toward products, the additional information seems to moderate the
influence of product aesthetics (i. e., unity) on perceptions of product quality.
As was pointed out in Chapter III, attitudes seem to be more closely linked (or
more susceptible) to aesthetic responses than other derived responses such as
quality. It may be that the more general nature of attitudes results in their
being more inclusive and thus allows (consciously or unconsciously) aesthetic
information to be weighted more heavily than it is in the case of derived
responses that are more focused on specific product attributes.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Refrigerators
Dislike
2
3
4
;

7 8
9
Like
Dislike
2
3
4
5
6
7 -3
9
Like
Ugly
2
3
-
5
6
7 8
9
Beautiful
Ugly
-)
3
4
5
6
7 8
9
Beautiful
Famiiiar
*)
3
4
5
6
7 8
9
Unfamiliar
Famiiiar
2
3
4
5
6
7 8
9
Unfamiliar
High Quality
9
3
4
5
6
7 8
9
Low Quality
High Quality
2
3
4
5
6
7 8
9
Low Quality
Low Pnce
2
3
4
5
6
7 8
9
High Price
Low Price
2
3
4
5
6
7 8
9
High Price
Disiike
3 4
5
6
7
8
9
Like
Disiike
2
3
A
5
6
8
9
Like
Ugly
*
3 4
5
6
7
8
9
Beautiful
Ugly
3
4
5
6
7 8
9
Beautiful
Famiiiar 1
-
3 4
5
6
7
8
9
Unfamiliar
Familiar
2
3
4
5
6
7 S
o
Unfamiliar
High Quality
:
3 4
5
6
7
8
9
Low Quality
High Quality
-i
3
4
5
6
7 8
0
Low Quality
Low Price
2
3 4
5
6
7
8
9
High Price
Low Price
2
3
4
5
6
7 8
9
High Pnce
144


35
Stimulus Development
The construction of stimuli that are to be used in research that examines
aesthetic influences requires a great deal of care. Many aspects of an objects
appearance have the potential to affect aesthetic responses. The influence of
aesthetic aspects of a stimulus object (e. g., a product or a picture of a product)
such as color, perspective, shading, etc. that could affect an aesthetic response
must be eliminated or controlled. In addition to the problems and limitations
inherent in controlling for extraneous aesthetic influences, the construction of
visual (i. e., pictorial) stimuli for aesthetic research is further complicated by
the fact that it is usually difficult to precisely determine the strength of
variables of interest (Nunnally 1981).
Although there are quite a number of studies that have manipulated
prototypicality (e. g., Hutchinson and Alba 1991), there is little in the way of
precedent for constructing stimuli that simultaneously exhibit different levels
of prototypicality and unity. The construction of sets of stimuli for this
experiment necessitated an exploration of the ways that unity and
prototypicality could be manipulated across a range of products. There were
four principal requirements that guided the development of the stimulus sets.
The first requirement, which pertained to the selection of particular products,
was that the product class had to have a strong category prototype. In order
to investigate the hypotheses concerning prototypicality it was necessary for
there to be a strong prototypical product design (i. e., form, configuration) for


125
by which the differences are transformed into an aesthetic response often
seems to occur nonconsciously.
The relationship between design principles and aesthetic response seems
to develop, at least in part, through learning. While this learning process may
be a formal one as in the case of educational programs for art and architecture,
it often seems to be the result of an informal and nonconscious process. Thus,
the phenomenon of aesthetic response may involve the nonconscious
development of design principle internal processing algorithms (design
principle IPAs) as well as the nonconscious application of these design
principle algorithms. Objects (products) that are consistent with a persons
design principle IPAs would be expected to produce more positive affect than
objects that violate a persons relevant design principle IPAs (Veryzer 1993).
Another area that merits further investigation is the tendency of people
to underweight the influence of product aesthetics on their evaluations of
products. Berkowitz (1987) found evidence of the misattribution of product
preferences by the perceivers of products. In a study that examined peoples
preferences for two food products, ratings on product attributes revealed that
subjects did not attribute the reasons for their choices to visual appeal or more
pleasing shape even though shape was the only actual difference between the
two types of products. The fact that consumers attributed their preferences to
other aspects such as "freshness" and "taste" even though shape was the only


118
The "comparative context" (i. e., different product versions rated
together) of Experiment 1 was similar to shopping situations where competitive
product offerings in the same category are displayed together (e. g., electronics
stores). This is also the case for products displayed in catalogs that feature
goods made by more than one manufacturer. The majority of products seem
to be displayed along with competitive product offerings and therefore are
evaluated in comparative contexts. In this sense Experiment 1 would seem to
be ecologically valid. However, most advertising and some products are
evaluated under conditions similar to the "non-comparative" context employed
in Experiment 2. The evaluative context in Experiment 2 was "non
comparative" in that product versions were evaluated singularly, without the
benefit of other similar product versions as reference points. This is the case
for most advertising. It is also the case for products such as automobiles,
boats, furniture, etc. that often involve exclusive distribution networks. The
evaluation contexts for these products is usually one where the consumer is not
able to directly (i. e., side by side) compare competitive product offerings. With
respect to these purchase situations Experiment 2 would seem to be
ecologically valid.
The third experiment, which focused on derived responses, found that
the unity aspect of product appearance influenced subjects attitudes toward
products despite the availability of information contained in product
descriptions. This study involved a context that was similar to many everyday


109
appearance conditions: Unity/Prototypical, Non-unity/Atypical, Unity/Atypical)
that were coupled with the target products (i. e., the weak product
descriptions) would influence the ratings of the target products.^
Analysis
The design was analyzed using a mixed ANOVA to assess the effects of
Unity, Product, Order/Position, and their possible interactions. Questionnaire
Version and Order/Position were between-subjects factors. Subjects ratings
of the control products were averaged and used as a covariate in the analysis
in order to reduce the error due to individual rating differences across subjects.
Hypothesis Testing
The critical test for examining whether or not the unity aspect of
product appearance influences subjects attitudes toward products and
perceptions of product quality despite the presence of additional information
involves the differences between the three target conditions. These conditions
differ only with respect to the appearance of the product version
(Unity/Prototypical, Non-unity/Atypical, Unity/Atypical) that is coupled with
the weak product description. If the results are consistent with the findings
14 Coupling the appearance conditions with the weak product
descriptions avoids ceiling effects which might otherwise occur if the
appearance conditions had been coupled with the strong product
descriptions.


16
H2: Objects (products) that are more prototypical (i. e., have
more shared features with the category schema) will receive more
favorable aesthetic responses than objects that are less
prototypical (i. e., fewer shared features with the category
schema).
Although prototypicality/familiarity seems to provide a satisfactory
explanation of aesthetic response in some cases (e. g., Kunst-Wilson and Zajonc
1980; Loken and Ward 1990), such an explanation does not seem adequate in
others (e. g., Meyers-Levy and Tybout 1989). In fact, in some cases it is the
converse of prototypicality (i. e., novelty, distinctiveness) that seems to account
for positive aesthetic response (Woll and Graesser 1982). Some research that
examined novelty and complexity may provide insight into this apparent
inconsistency. In a study that examined novelty ratings of simple and complex
shapes, Eisenman (1968) found that more complex polygons (i. e., those with
more sides) were rated as being more novel. In a related study, Berlyne (1970)
examined the effect of repeated presentation on hedonic value for simple and
complex (i. e., novel) patterns. Berlyne had subjects rate two simple and two
complex patterns six times on a 7-point pleasingness scale. Between
consecutive tests the subjects saw each of the patterns eight times without
having to record a judgement. The results confirmed that ratings of complex
(i. e., novel) patterns rose and then fell after reaching a maximum at the third
test. Ratings of the simple (i. e., less novel) patterns, which were initially


85
unity score effect in Experiment 1. Although all nine product categories were
included, the bathroom scales stimulus set was omitted from subsequent
analyses for two reasons. First, this simplified the analyses of the results.
Second, this stimulus set was the only reversal of the unity effect observed in
Experiment 1. Each product category was represented and an effort was made
to select an equal number of each of the three types of manipulations (i. e.,
shape, texture, and trim). The nine stimulus sets that were employed were the
bathroom scale/textureT. V. Remote Control/trim, flashlight/trim, lamp/
texture, reffigerator/shape, telephone/shape, hair dryer/texture, dresser/trim,
and clock/shape.
The overall design of the experiment was a 2(Feature A) x 2(Feature B)
x 8(Products) mixed factorial design. Just as in the first experiment the
manipulation of Features A and B resulted in the manipulation of
prototypicality and unity. In this experiment, however, each cell of the crossed
Feature A and B factors [i. e., product versions: Unity/Prototypical (++), Non-
Unity/Atypical (+-), Non-Unity/Atypical (-+), Unity/Atypical (--)] was contained
in one of four questionnaire versions. That is, each questionnaire contained
all eight product categories (and the extra product); each subject received one
11
This product had proven to be problematic in Experiment 1 but was
included for the purpose of maintaining continuity between the product
presentations of the two experiments.


Please indicate vour reaction to the appearance of the proposed
products by marking (circling) the scales below each.
TV Remote Controls
Dislike
Ugly
Familiar
High Qualiry
Low Pnce
1 3 : 6 8 9 Like
1 3 d 6 8 9 Beautiful
1 2 j- -i r 6 8 9 U ni amiliar
1 - 6 7 8 9 Low Qualirv
Dislike
Ugly
Familiar
High Quality
12 3 4
6 7 8 9
High Pnce
Low Pnce
i 2 3 4 5 6 7 8 9 Like
1 3 4 5 6 7 8 9 Bcaumui
1 2 456789 Unfamiliar
*2 3 4 5 6 7 8 9 Low Quality
i 27456789 High Pnce
crac.?
B3&
>333
Disiixe
2 3 4 f6
~ s
0
Like
Dislike
2
3 4 5
r 7 8
9
Lute
U gly
2 3 4 5 o
7 8
9
Beautiful
Ugly
2
3 4 5
i 8
9
Beautiful
Familiar !
2 3 4 5 6
- 8
9
Unfamiliar
Familiar
2
3 4 5
ft 7 8
0
Unfamiliar
High Quality
2 3 4 5 6
8
9
Low Quality
High Quality
2
3 4 5
6 7 8
9
Low Quabr.
Low Price
2 3 4 5 6
7 8
9
High Price
Low Price
2
3 4 5
6 7 8
9
High Pnce
134


49
The means for the product beauty ratings for each of the twenty-seven
stimulus sets are presented in Table III-2. Across the twenty-seven stimulus
sets the products that did not exhibit unity (M = 4.06) were rated lower than
the products that did exhibit unity (M = 5.12). The analysis of variance
presented in Table III-3 indicates that this difference which is captured in the
Feature A x Feature B interaction is significant F (1,184) = 338.16, p < .0001.
This supports the unity hypothesis which predicted that aesthetic responses
are more positive for objects (products) that exhibit high consistency with the
visual organization principle of unity than they are for objects (products) that
are not consistent with the unity visual organization principle.
The significance of the Factor A x Factor B interaction (i. e., main effect
of unity) was also tested in parallel analyses that utilized the linear unity
scores (U--_) formed from subjects ratings of all four versions of each stimulus
set [i. e., (Unity/Prototypical + Unity/Atypical) (Non-unity/Atypical + Non-
Unity/Atypical)] and the nonlinear unity scores (U^) [i. e., (Unity/Atypical) -
minimum (Non-unity/Atypical)]. The analysis of variance tables for each of
these approaches are shown in Table III-4 and Table III-5. The unity
hypothesis (i. e., Feature A x Feature B interaction) is tested by determining
whether or not the intercept is zero. That is, if the intercept were actually
zero, what would the probability be of obtaining, by chance alone, a value as
large or larger than the one actually obtained? The results reported in Table


34
Unity
Feature B
Feature A
Feature B
(a)
Schema Incongruity
Feature A
+
(b)
Feature B
Feature A
+
0
0
-
Feature B
( c )
Novelty
Feature A
-
0
0
-h
( d)
+ indicates the highest (i. e., most positive) rating(s) for a product version
contained in the stimulus set.
- indicates the lowest rating(s) for a product version contained in the set.
0 indicates a rating between the highest and lowest ratings.
FIGURE III-2
POSSIBLE PATTERNS OF RESULTS


BIOGRAPHICAL SKETCH
Robert Whitman Veryzer, Jr. is the third of four sons born to Robert and
Marion Veryzer. He attended Birmingham Seaholm High School in
Birmingham, Michigan. From 1979 to 1983 Robert attended Olivet College
where he was the recipient of the Olivet College Presidential Scholarship and
was graduated with honors. After working for a year for a small
manufacturing company, he left to pursue a Master of Business Administration
degree in marketing at Michigan State University. In addition to his studies,
Robert worked for the State of Michigan as a marketing assistant and later for
General Motors Corporation-Oldsmobile Product Planning Department as a
product research analyst. After completing his M. B. A. degree, he spent one
semester conducting an independent research project that involved examining
the role of product design in business. Robert then went to work for General
Foods Corporation as an assistant product manager. In the Fall of 1989, he
entered the Ph.D. program in marketing at the University of Florida. Robert
has accepted an offer to join the faculty of Rensselaer Polytechnic Institute.
171


102
Even though rating product versions without the context of other versions
reduced the magnitude of the effects, the influence of unity on aesthetic
responses, attitudes, and perceived familiarity was still significant. The
finding that the influence of unity on quality and price is not as strong as it
is on the other responses is consistent with the findings of Experiment 1.
Clearly, context (i. e., availability to compare product versions with other
products in the same product category during the evaluation task) plays a role
in moderating the influence of factors such as unity, but the findings presented
here demonstrate that the unity effects are robust and that the influence of
unity on product evaluations may be pervasive.


39
entire stimuli set contained one hundred and eight drawings of products.
Examples of these stimuli sets are presented in Appendix B.
Experimental Design
The overall design of the experiment is a 2(Feature A) x 2(Feature B) x
9(Products) x 3(Version) x 2(Order) mixed factorial design. As was pointed out
in the preceding discussion the manipulation of features A and B result in the
manipulation of prototypicality and unity. The twenty-seven replicate sets of
stimuli were organized into three questionnaire versions. Each of the three
questionnaire versions contained all nine of the product categories (each
product category was made up of a set of four variations of the product i. e.,
Feature A x Feature B) but for only one of the three transformation types
(shape, trim, or texture). So for example, questionnaire version one contained
the telephone stimulus set (set of four product variations) that were altered
using the trim transformation; questionnaire version two contained the
telephone stimulus set altered using the texture transformation; and
questionnaire version three contained the telephone stimulus set altered using
the shape transformation. While each questionnaire contained each of the nine
product categories, in three of the nine cases the trim transformation was
utilized, in three of the nine cases the texture transformation was utilized, and
in three of the nine cases the shape transformation was utilized. A diagram
of this design is presented in Appendix C. Thus, Feature A, Feature B,


Please indicate vour
products bv marking
C
reaction to the
(circling) the scales
appearance of the
below each.
proposed
fiathroorn_s£ales
132


92
each product category. An examination of the manipulation check for the
impact of changes in product features A and B on beauty ratings (refer to
Table IV-1) reveals that there are differences in the strength of the
manipulations across the eight products. This is also reflected in the inter
action between Products and Version (i. e., cells of the crossed feature A and
B "factors" of the design; e. g., Unity/Prototypical).
There were also significant order effects [Order F(l,230) = 6.89, p <
.0092; Product Order F(7,1610) = 2.24, p < .0285]. An examination of the
means for products for each of the presentation orders suggests that as in
Experiment 1, these effects may be attributable to subjects increasing
acceptance of the use of simple drawings to depict products.
Derived Responses
The influence of unity and prototypicality on attitude, quality
perceptions, and price expectations was also examined. The means for the
liking, quality, and price ratings are presented in Table IV-4, Table IV-5, and
Table IV-6, respectively. The patterns exhibited for the liking and quality
means are very similar to the pattern exhibited by the beauty means that were
presented in the preceding section. Interestingly, the pattern for the overall
means for price expectations were similar to those observed in Experiment 1
with the Unity/Atypical () product versions being the most highly rated and
the Unity/Prototypical (++) product versions being the lowest rated product.


33
category prototype. Thus, this product is quite atypical (-) with respect to
prototypicality. This version does, however, exhibit unity since there is a
visual connection or consistency among the two product features. As can be
seen in the figure, this manipulation of the two product features results in
three levels of prototypicality and two levels of unity being produced.
Four possible patterns of results are shown in Figure III-2. These
patterns have been labeled to indicate the explanation of aesthetic response
that each supports. The unity explanation of aesthetic response would be
supported by results showing that both unity conditions were more highly
rated than both of the non-unity conditions (Figure III-2a). Such results would
suggest that a product variations consistency with the unity visual
organization principle positively influenced subjects aesthetic responses. The
predictions of the schema incongruity hypotheses are the opposite of those for
the unity explanation of aesthetic response. The schema incongruity
hypotheses predict that moderately atypical stimuli will be preferred to stimuli
that are prototypical or extremely atypical (Figure III-2b). The prototypicality
explanation of aesthetic response would be supported by the results depicted
in Figure III-2c. Here, the more prototypical a product variation is the higher
its ratings. A pattern of results exactly opposite those for the prototypicality
explanation would suggest an effect of novelty (Figure III-2d). In this way, the
basic 2(Feature A) x 2(Feature B) design allows the hypotheses presented
earlier to be examined.


TABLE OF CONTENTS
ACKNOWLEDGEMENTS iii
ABSTRACT vi
CHAPTERS
I INTRODUCTION 1
II CONCEPTUAL BACKGROUND ON THE
FACTORS THAT INFLUENCE AESTHETIC
RESPONSES 6
Introduction 6
Factors Affecting Aesthetic Responses 9
Visual Organization Principles 9
Prototypicality 14
Derived Responses 21
III EXPERIMENT 1 28
Overview 28
Stimuli 28
Experimental Design 39
Experimental Procedure 40
Results and Discussion 41
Summary 81
IV EXPERIMENT 2 84
Overview 84
Experimental Design 84
Experimental Procedure 86
Results and Discussion 87
Summary 101
V EXPERIMENT 3 104
Overview
Stimulus Materials ,
Experimental Design
104
105
106


54
TABLE III 5
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
NONLINEAR UNITY SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subject Effects
Intercept
1
3873.5708
473.48
0.0001
Version
2
340.8709
20.83
0.0001
Order
1
3.2588
0.40
0.5287
Version Order
2
1.3466
0.08
0.9210
Error Sub (Version Order)
190
1554.4164
Within Subject Effects
Product
8
298.1844
9.67
0.0001
Trans
2
63.3658
8.21
0.0001
Prod Version
14
484.9146
8.98
0.0001
Prod Order
8
123.7531
4.01
0.0100
Trans Order
2
41.8300
5.42
0.0100
Prod Version (Residual)
14
75.8172
1.40
0.2500
Error Sub (Prod)
1520
5861.4233
Unity/Atypical () products (M = 4.90) suggests that prototypicality does have
some effect on aesthetic responses.
The analysis of variance presented in Table III-3 as well as those shown
in Table III-4 and Table III-5 also show that there were a number of factors
and interactions among factors that resulted in significant effects in this
design. Most of these significant effects can be attributed to differences in the
strength of the manipulations used to create each stimulus set and the
differences across product categories. The effect of Product was significant,
F (8,1472) = 40.78, p < .0001, as was the effect of Transformation type
F (2,1472) = 24.91, p < .0001. Since stimulus sets (i. e., Transformation types)
are contained in each of the versions of the questionnaire the effects of these


158
TABLE C-l
DIAGRAM OF EXPERIMENT 1
Bathroom
Scale
TV
Remote
Flash-
Light
Lamp
Refrig
erator
Tele
phone
Hair
dryer
Dresser
Clock
Version 1
Sh
Tx
Sh
Tr
Tx
Tr
Sh
Tx
Tr
Version 2
Tr
Sh
Tr
Tx
Sh
Tx
Tr
Sh
Tx
Version 3
Tx
Tr
Tx
Sh
Tr
Sh
Tx
Tr
Sh
Sh = Shape Transformation
Tr = Trim Transformation
Tx = Texture Transformation
Note: The order in which the stimulus sets were presented was reversed for half of the subjects.


160
TABLE D-l
DIAGRAM OF EXPERIMENT 2
TV
Remote
Flash
light
Lamp
Refrig
erator
Tele
phone
Hair
dryer
Dresser
Clock
Questionnaire
Version 1
(++)
(+-)
(-+)
(-)
(++)
(+-)
(-+)
(--)
Questionnaire
Version 2
(-+)
(++)
()
(+-)
(-+)
(++)
(-)
(+-)
Questionnaire
Version 3
(+-)
(-)
(++)
(-+)
(-+)
()
(++)
(-+)
Questionnaire
Version 4
(--)
(-+)
(+-)
(++)
(--)
(-+)
(+-)
(++)


APPENDIX C
DIAGRAM OF EXPERIMENT 1


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
TV Remote
Controls
//aaa/A
nr-ir-t r, h\
/iMssl/l
jbF
Sill
f~~ y
Dislike 123456'89 Like
Ugly 1 23456789 Beautiful
Familiar 12 3 4567S9 Unfamiliar
High Quality 123456789 Low Quaiitv
Low Pnce 123456789 High Pnce
Dislike 1 2 3 4567S9 Like
Ugly 1 2 3 4 5 6 7 8 9 Beautiful
Familiar 1 23456789 Unfamiliar
High Quality 123456789 Low Quality
Low Pnce 1 2 3 4 5 6 7 8 9 High Pnce
7 1
!§§§!)
0s§//¡
¡38g¡)
iSasaUI
F j/
y
Dislike 12 3- 56789 Like
Ugly 12 3 -56789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 12 3 456-89 Low Quality
Low Pnce 12 3 456789 High Pnce
Dislike 1 2345~7sa Like
Ugly ; 23456789 Beautiful
Familiar 1 23456789 Unfamiliar
High Quality 1 2 3 4 5 6 7 S 9 Low Quailt>
Low Price 1 23456789 High Pnce
135


73
products" and thus relate novelty (atypicality) in general to the higher prices
associated with such products.
The analysis of variance results for liking, quality, and price are
reported in Table III-15 to Table IIT23. For all three of the derived responses
there was a significant effect of unity. In the case of attitude (i. e., liking
ratings) the effect of unity (i. e., Factor A x Factor B interaction) was highly
significant across all three variations of the dependent measure: rating scores
F (1,185) = 378.04, p < 0.0001; Linear unity scores F (1,185) = 378.04, p <
.0001; and pure-unity score F (1,188) = 530.00, p < .0001. That aesthetic
responses would have a strong influence on attitudes is not too surprising since
many conceptualizations of attitude maintain that attitudes contain an
affective component or that affective responses are antecedents of attitudes
(Ajzen 1989; Edwards 1990). The effect of unity on (perceived) quality was not
as strong as it was for liking but it was still significant except in the analysis
that utilized nonlinear unity scores where it was marginally significant: rating
scores F (1,179) = 83.54, p < .0001; linear unity scores F (1,179) = 83.54, p <
.0001; nonlinear unity scores F (1,188) = 3.74, p < .0545. The effect of unity
on ratings of price was significant across all the dependent measure variations:
rating scores F (1,179) = 57.67; linear unity scores F (1,179) = 57.67, p < .0001;
nonlinear unity scores F (1,188) = 398.57, p < .0001.
Once again, the patterns of interactions observed among Products,
Transformation type, and Order for these derived responses were the same as


APPENDIX E
PRODUCT DESCRIPTIONS USED IN EXPERIMENT 3


60
rating task. This was controlled for to some degree by reversing the
presentation order of stimuli in the second order.
Perceived Familiarity
Hypothesis 6 predicted that consistency with the visual organization
principle of unity was positively related to perceived familiarity. The means
for the familiarity ratings for each of the twenty-seven stimulus sets are
presented in Table III-8. The analysis of variance tables for the rating score,
linear unity score, and nonlinear unity score approaches are presented in
Tables IIX-9, HI-10, and III-ll, respectively.
The analysis of variance for the subject ratings of familiarity (Table III-
9) indicates a significant unity effect for perceived familiarity F (1,181) =
254.47, p < .0001. The results for the analysis of variance using linear unity
scores were very similar to those observed using subject ratings with the unity
effect achieving the same level of significance F (1,181) = 254.47, p < .0001.
The analysis of variance for the nonlinear score approach, which is probably
the most appropriate analysis in this case since it removes the influence of the
Unity/Prototypical cell of the design, also indicates a significant effect of unity
on perceived familiarity F (1,188) = 10.65, p < .0013.^ These results
suggest that unity can encourage feelings of perceived familiarity.
The pattern of interactions observed among Product, Transformation
type, and Order were the same as were discussed for the influence of
unity on aesthetic responses.


TABLE III 8
MEAN FAMILIARITY RATINGS AND UNITY SCORES
61
Product
(++)
(+-)
(-+)
(-)
UL
%
1. Bathroom scale
Sh
7.62
7 14**a
5.40***
5.56***
0 3i**b
0.45*
Tr
7.06
6.97
7.13
6.98
0.025
0.27
Tx
7.05
6.77
6.23*
6.10**
0.03
0.42
2. TV Remote Control
Sh
7.59
5.08**
5 44***
4 94***
0.97***
0.65
Tr
6.62
6.13*
6.51
6.89
0.53**
1 45***
Tx
7.35
5.17***
5.32***
5.86***
1 31***
1.00***
3. Flashlight
Sh
7.93
4 80***
4.16***
4.05***
1 54***
0.27
Tr
7.49
5.52***
5.79***
5.49***
0.84***
0.35
Tx
7.49
4.52***
4 27***
6.35***
2.43***
2.21***
4. Lamp
Sh
8.22
5.95***
3.98***
3.57***
0.89***
-0.18
Tr
7.21
6.10***
6.81*
6.60*
0.45**
0.67**
Tx
7.43
6.38**
6.18**
7.46
4 49***
2.02***
5. Refrigerator
Sh
7.97
3.16***
3.25***
3.84***
2.68***
0.92***
Tr
7.75
5.52***
5.28***
6.61***
1.88***
^ 74***
Tx
8.22
4.21**
4 29***
4.52***
2.16***
0.67**
6. Telephone
Sh
7.26
3.84***
4 77***
3 41***
1.10***
-0.31
Tr
7.16
6.19***
5.46***
5.52***
0.50***
0.34**
Tx
6.89
6.24**
5.73***
6.41*
0.65***
0.82**


8
what they know" (Pye 1978, p. 11). In fact, the well worn maxim that "form
follows function" has even been challenged (Lewalski 1988; Pye 1978).
Interestingly enough, designers are beginning to look to the disciplines of
experimental psychology and consumer behavior in order to understand the
cognitive condition of design (Zaff 1987).
The lack of understanding concerning what makes an object
aesthetically pleasing is particularly evident in the area of new product
development. The issue of aesthetic response and the factors that influence it
are frequently overlooked in many discussions of the new product development
process. Much of the work in this area seems to sidestep these considerations
by either ignoring the role of industrial design (i. e., the process of shaping or
giving form to goods that are to be mass produced) in the new product
development process or subsuming the industrial design function under the
engineering function (e. g., Gruenwald 1985; Urban and Hauser 1993). In
either case, the result is that the role of product appearance in the success of
a new product is not explicitly acknowledged or addressed. Given the
increasing importance of product aesthetics this omission is a rather serious
deficiency. Fortunately, the distinct role of industrial design and its
relationship to the engineering and marketing functions is beginning to receive
the attention that it merits (e. g., Lorenz 1986). However, most of the work
that addresses the role of industrial design in the new product development


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Telephones
Dislike 1 234567*9 Like
L si)' 1 3^56789 Beautiful
Familiar 133456789 Unfamiliar
High Quality 123456789 Low Quality
Low Pnce 123456789 High Pnce
Disilke 1 2 3 -i 5 6 7 S y Like
uS!y 12 3 456789 Bcauui'Ji
Familiar 1 2 3 4 5 6 7 8 9 Un/amiiiar
High Quality 12 3 4 5 6 7 8 9 Low Quality
Low Pnce 12 3 456789 High Pnce
Dislike
3
a 5
6 7
8 9
Like
Ugly 1
*>
3
4 5
6 7
8 9
Beautiful
Familiar 1
-
3
4 5
5 7
5 9
Unfamiliar
High Quaiitv l
3
4 5

5 9
Low Quality
Low Pnce 1
2
3
4 5
6 7
8 9
High Price
Dislike
2 3 4 5
6 7
8
4 Luce
Ugly
2 3 4 5
6 7
8
9 Beauuiu.
Familiar
2 5 4 5
6 7
s
0 Lnfarmuar
High Quality 1
2 3 4 5
6 7
§
9 Low Quail?.
Low Pnce 1
2 3 4 5
6 7
S
0 High Pnce
145


TABLE III 14
MEAN PRICE RATINGS AND UNITY SCORES
71
Product
(++)
(+-)
(-+)
(-)
uT,
%
1. Bathroom scale
Sh
4.42
4.57
4.86
4.93
-0.04
1.02***
Tr
3.53
5.09
3.79
5.38
0.01
1.65***
Tx
3.66
5.74
4.38
5.36
-0.55
1 H***
2. TV Remote Control
Sh
4.38
5.22
5.08
5.56
-0.18
0.91***
Tr
3.80
4.89
4.84
6.56
-0.31
2.08***
Tx
4.62
4.92
4.85
5.55
0.21
1.03***
3. Flashlight
Sh
4.49
4.89
5.21
5.46
-0.10
0.91***
Tr
4.62
4.33+a
4.68
4.57
0.06
0.47*
Tx
4.34
4.14
4.17
5.09
0.55**b
4. Lamp
Sh
4.45
4.95
5.10
5.27
-0.13
0.90***
Tr
3.53
4.52
4.95
5.28
-0.32
1.05***
Tx
3.80
4.82
4.49
5.74
0.07
1 54***
5. Refrigerator
Sh
5.58
4.81*
4.66**
5.89
0.99**
1.48***
Tr
5.07
4.84
4.93
6.25
0.72***
1.69***
Tx
5.50
4.58**
4.87*
5.10
0.57**
0.77**
6. Telephone
Sh
5.26
5.11
5.03
5.84
0.44**
1 lg***
Tr
4.85
4.97
4.25**
4.33
0.03
0.28*
5.63
5.34
5.01**
6.14
0.69**
1.30***


I certify that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of I^optor^f I^hilo^qghy.
J/Wesley Hutchinson, Chairman
associate professor of Marketing
I certify that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of Doctor of Philosophy
Richard J. Lutz
Professor of Marketing
I certify that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of Doctor of Philosophy.
Associate Professor of Marketing
I certify that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of Doctor of Philosophy.
David G. Mick
Assistant Professor of Marketing
I certifjr that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of Doctor of Philosophy.
^Jonathan Hamilton
Associate Professor of Economics
This dissertation was submitted to the Graduate Faculty of the
Department of Marketing in the College of Business Administration and to the
Graduate School and was accepted as partial fulfillment of the requirements
for the degree of Doctor of Philosophy.
August, 1993
Dean, Graduate School


42
aesthetic response was a -1.34. This pattern was consistent across most of the
sets of stimuli for each product transformation (i. e., stimulus sets).5 The
effects of a change in Feature A or Feature B from the category prototype are
shown in Table III-l. In twenty-one of the twenty-seven stimulus sets the
change of either Feature A or Feature B resulted in a lower beauty rating for
the product variation as compared with the rating for the category prototype.
In two stimulus sets the effects of a change in the product features were
mixed. That is, a change in one feature (either A or B) had a negative effect
on aesthetic response but a change in the other feature (B or A) had a positive
effect (see the lamp/trim and lamp/texture stimulus sets in Table III-l). In
only four of the twenty-seven stimulus sets were the effects of a change in
Feature A and Feature B both (separately) in the opposite (i. e., positive)
direction (see the bathroom scale/trim, bathroom scale/texture, TV remote
control/trim, and the hair dryer/trim stimulus sets in Table III-l). Overall, the
manipulation check demonstrates that for the majority of the stimulus sets a
reduction in favorable aesthetic response does occur when the products are
altered in such a way as to be both less prototypical and less unified than the
category prototype. It should be noted that there is variation in the reduction
in favorable aesthetic response across the twenty-seven stimulus sets. This is
to be expected given the difficulties of precisely estimating the strength of each
5 "Stimulus set" will be used throughout this discussion to refer to one of
the transformation types (shape, trim, or texture) of a product category
(e. g., telephone/shape transformation set).


58
TABLE III 7
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES -
NONLINEAR UNITY SCORES
Product
Ujt Score for
Beauty
S. E.
t-test
1. Bathroom scale
Sh
0.8788
0.1985
4.4272***
Tr
1.3788
0.1988
6.9359***
Tx
0.4308
0.2816
1.5296+
2. TV Remote Control
Sh
1.4091
0.2653
5.3113***
Tr
2.2308
0.2287
9.7551***
Tx
1.3636
0.2981
4.5741***
3. Flashlight
Sh
0.5539
0.2352
2.3551**
Tr
0.6061
0.2678
2.2628*
Tx
3.0000
0.3260
9.2017***
4. Lamp
Sh
0.3939
0.1753
2.2479*
Tr
1.4242
0.2275
6.2592***
Tx
2.0769
0.2768
7.5047***
5. Refrigerator
Sh
2.1969
0.2800
7.8451***
Tr
2.0308
0.2685
7.5629***
Tx
1.3182
0.2663
4.9493***
6. Telephone
Sh
0.9394
0.2678
3.5073**
Tr
0.2121
0.1441
1.4725+
Tx
1.4091
0.2443
5.7688***


APPENDIX A
PRODUCTS EXPLORED FOR USE AS STIMULI


19
vs. soft drink) that was specified in the first sentence of the product description
(e. g., subjects were told that they were going to be evaluating either a new
"beverage" or a new "soft drink" and later they were told that the product had
either "high preservatives" or had "all natural ingredients"). Across their three
experiments Meyers-Levy and Tybout found support for Mandlers view that
the process of responding to levels of schema congruity influences evaluations,
and that moderate schema incongruity enhances evaluations. Moderate
schema incongruity led to more favorable evaluations than either schema
congruity or extreme schema incongruity. This discussion suggests the
following hypotheses:
H3: Moderate schema incongruity (i. e., distinctiveness, novelty)
leads to a more favorable aesthetic response than does complete
congruity between a product and its product class (i. e.,
prototypicality).
H4: Moderate schema incongruity leads to more favorable
aesthetic responses than does extreme schema incongruity.
It is also possible that people may prefer more novel products due to
variety seeking (Holbrook and Hirschman 1982) or perhaps because of a
products salience relative to other products (Loken and Ward 1990; Woll and
Graesser 1982).


63
TABLE III 9
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
73.7530
0.55
0.5789
Order
1
233.9050
3.48
0.0638
Version Order
2
8.6500
0.06
0.9377
Error Sub (Version Order)
181
12174.5330
Within Subiect Effects
Product
8
946.3903
20.96
0.0001
Trans
2
724.2835
64.16
0.0001
Prod Version (Residual)
14
677.6986
8.58
0.0001
Prod. Order
8
38.3970
0.85
0.5582
Trans Order
2
46.4860
4.12
0.0250
Prod Version Order (Residual)
14
189.0356
2.39
0.0100
Error Sub (Product)
1448
8172.4271
Feature A
1
1459.8545
259.23
0.0001
A Version
2
300.8827
27.71
0.0001
A Order
1
7.6979
1.37
0.2439
A Version Order
2
17.3149
1.54
0.2177
Error Sub (A)
181
1019.2983
Feature B
1
919.4761
185.88
0.0001
B Version
2
14.4650
1.46
0.2345
B Order
1
1.4952
0.30
0.5831
B Version Order
2
4.1657
0.42
0.6570
Error Sub (B)
181
895.3306
A B
1
2362.5960
254.47
0.0001
A B Version
2
69.0487
3.72
0.0261
A B Order
1
21.4075
2.31
0.1306
A B Version Order
2
22.0981
1.19
0.3066
Error Sub (A B)
181
1680.4642
Prod A B
8
549.3698
28.33
0.0001
Prod A B Version
16
434.3241
11.20
0.0001
Prod A B Order
8
109.7385
5.66
0.0001
Prod A B Version Order
16
52.1372
1.34
0.1618
Error Sub (Prod A B)
1448
3510.0931


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Hair-drvers
Dislike
Ugly
Familiar
High Quality
Low Pnce
3
4
5
6
7
8
9
Like
Dislike
i
3
4
5
6
7
8
9
Like
*>
3
4
5
6
7
8
9
Beautiful
Ugly
3
4
5
6
7
8
9
Beautiful
n
3
4
5
6
7
8
9
Unfamiliar
Familiar 1
2
3
4
5
6
7
8
9
Unfamiliar
"i
3
4
5
6
7
8
9
Low Quality
High Quality
3
4
5
6
7
8
9
Low Quality
3
4
5
6
7
8
9
High Pnce
Low Price
-
3
4
S
6
7
8
9
High Pnce
Dislike
2
3 4
5
6
7
8
9
Like
Dislike
3
4
5
6
7
8
9
Like
Ugly
2
3 4
5
6
7
8
9
Beautiful
Ugly
-
3
4
5
6
7
8
9
Beautiful
Familiar
-
3 4
5
6
7
8
9
Unfamiliar
Familiar 1
-
3
5
6
7
8
9
Unfamiliar
High Quality
2
3 4
5
6
7
8
9
Low Quality
High Quality
3
4
5
6
7
8
9
Low Quality
Low Pnce
2
3 4
5
6
7
8
9
High Price
Low Price
3
4
5
6
7
8
9
High Price
149


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Lamps
Dislike
2 3-1
5 6
7 8
9
Like
Dislike
Ugly
: 4
5 6
7 8
9
Beautiful
Ugly
Familiar !
: 3 4
5 6
7 8
9
Unfamiliar
Familiar
High Quality
5 6
~ s
9
Low Quality
High Quality
Low Price
2 3 4
5 6
7 8
9
High Pnce
Low Pnce
2 3 4 5 6 7 8 9 Like
23456789 Beautiful
23456789 Unfamiliar
2 3 -156789 Low Quality
2 3 456789 High Pnce
Dislike
: 3 4
5 6 7
8
9
Like
Dislike
2
4 5
6
7
8
9
Like
Ugly
: 3 4
5 6 7
8
9
Beautiful
Ugly 1
-
3
4 5
6
*7
8
9
Beautiful
Familiar
: 3 4
5 6 7
8
9
Unfamiliar
Familiar
2
3
4 5
6
8
Q
Unfamiliar
High Quality
2 3 4
5 6 7
8
9
Low Quality
High Quality 1
2
3
4 5
6
7
8
9
Low Quality
Low Pnce
2 3 4
5 6 7
8
9
High Price
Low Price 1
2
3
4 5
6
7
8
9
High Price
140


3
surprisingly little in the way of design/aesthetic theory has been offered that
aids in our understanding of how aesthetic responses are formed (Berlyne
1974, p. 5; Gorski 1987, p. 68; Holbrook and Zirlin 1985; Pye 1978, p. 88).
Even though disciplines concerned with basic research such as experimental
aesthetics remain dormant or in a "protracted infancy" (Berlyne 1971), other
disciplines such as industrial (product) design look to them to address the lack
of design theory (Gorski 1987, p. 68). The end result is that managers
continue to exhibit much "unease" when it comes to making decisions about
design and managing design projects (Oakley 1990, p. 7). This is particularly
problematic for managers charged with transferring new technology out of the
laboratory and into the market. It is at this point that the appearance of a
product plays a crucial role in communicating the products identity and use to
consumers.
Although consumer research seems to be ideally suited for the study of
aesthetic response due to its unique combination of scientific research methods
and its tangible research context (i. e., consumer product focus), consumer
researchers have taken only tentative steps toward exploring aesthetics and
its relationship to consumers behavior. Much of this work has been concerned
with debating the definition and scope of consumer aesthetics rather than
examining the nature and influence of aesthetics and aesthetic responses
(Holbrook and Zirlin 1985). Thus, aesthetic research has proceeded without
the benefit of a conceptual foundation (Holbrook and Zirlin 1985; Olson 1981).


48
scores, and nonlinear unity scores) in order to insure that the results reported
were not simply an artifact of a particular analysis scheme.
Hypothesis Testing
The critical test for examining the hypotheses concerning the influence
of unity and prototypicality on aesthetic and derived responses involves the
Feature A x Feature B interaction. In this design, the Feature A x Feature B
interaction is the main effect of unity. This interaction "tests" the competing
predictions of the prototypicality and unity hypotheses. If the
Unity/Prototypical (++) products of this design receive higher ratings than the
Non-Unity/Atypical (+-) and Non-unity/Atypical (-+) products and these in turn
are more highly rated than the Unity/Atypical (--) products then the
prototypicality explanation of aesthetic (and derived) response would seem to
be supported (Hypothesis 2). If, however, the Unity/Prototypical (++) products
and the Unity/Atypical (--) products were both rated significantly higher than
the Non-unity/Atypical (+-, -+) products then support for the unity hypothesis
would be indicated (Hypothesis 1). The schema incongruity explanation of
aesthetic response would be supported if the Non-unity/Atypical (+-, -+)
products were rated significantly higher than the prototypical products (++)
and/or the extremely atypical products (). If aesthetic response was found to
be positively related to atypicality than this would suggest a novelty effect.


52
TABLE III 3
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
42.9350
1.63
0.1991
Order
1
13.7790
1.05
0.3080
Version Order
2
107.6630
4.08
0.0184
Error Sub (Version Order)
184
2425.7720
Within Subiect Effects
Product
8
1081.1417
40.78
0.0001
Trans
2
165.0833
24.91
0.0001
Prod Version (Residual)
14
566.1260
12.02
0.0001
Prod Order
8
199.3217
7.52
0.0001
Trans Order
2
5.4222
0.82
0.2500
Prod Version Order (Residual)
14
43.1955
0.93
0.2500
Error Sub (Product)
1472
4878.0299
Feature A
1
124.2720
23.97
0.0001
A Version
2
252.2238
24.33
0.0001
A Order
1
10.6083
2.05
0.1542
A Version Order
2
6.1115
0.59
0.5556
Error Sub (A)
184
953.7624
Feature B
1
52.8499
11.86
0.0007
B Version
2
114.6225
12.86
0.0001
B Order
1
21.7458
4.88
0.0284
B Version Order
2
17.8531
2.00
0.1378
Error Sub (B)
184
819.8276
Feature A Feature B
1
1941.8813
338.16
0.0001
A B Version
2
67.4647
5.87
0.0034
A B Order
1
3.1291
0.54
0.4614
A B Version Order
2
5.6170
0.49
0.6140
Error (A B)
184
1056.6273
Prod A B
8
933.9437
53.26
0.0001
Prod A B Version
16
787.4382
22.45
0.0001
Prod A B Order
8
94.0783
5.37
0.0001
Prod A B Version Order
16
35.9705
1.03
0.4255
Error Sub (Prod A B)
1472
3226.3444


64
TABLE III 10
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
LINEAR UNITY SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
71.9460
10.65
0.0013
Version
2
426.8649
31.61
0.0001
Order
1
54.3667
8.05
0.0050
Version Order
2
11.3613
0.84
0.4328
Error Sub (Version Order)
188
1269.5378
Within Subiect Effects
Product
8
727.7061
20.07
0.0001
Trans
2
661.0854
72.93
0.0001
Prod Version (Residual)
14
557.1115
8.78
0.0001
Prod Order
8
60.4733
1.67
0.1016
Trans Order
2
1.2788
0.14
0.2500
Prod Version Order (Residual)
14
84.6541
1.33
0.2500
Error Sub (Prod)
1504
6816.5729


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Flashlights
Dislike 2 3 -1 5 o 7 8 9 Like
Ugly 1 2 .' 456789 Beautiful
Familiar 1 23-5 6789 Unfamiliar
High Quality 1 23456789 Low Quality
Low Pnce 1 23456789 High Pnce
Dislike 1 2 5 4 5 6 7 5 Like
Ugly 1 2 3 4 5 6 7 8 3 Beautiful
Familiar 12 3 45678 Unfamiliar
High Quality 1 2 3 J 5 6 7 8 a Low Quality
Low Pnce 1 2 3 4 5 6 7 8 3 High Pnce
Dislike
2 3 4 ?67
s
9
Like
Dislike
6
7
y
~ Like
Ugly
2 3 4 5 6 7
8
9
Beautiful
Ugly
: ? -t 5
6
7
J
0 Beautiful
Familial
2 3 4 5 6 -
8
9
Unfamiliar
Familiar
2 3 4 5
6
7
S
0 Unfamiliar
High Quality
2 3 4 5 6 7
8
9
Low Quality
High Quality
2 3 4 5
6
7
8
0 Low Quaiir\
Low Pnce
2 3 4 5 6 7
8
9
High Price
Low Price
2 3 4 5
6
7
8
Q High Pnce
137


9
process stops short of identifying and investigating the factors that influence
aesthetic responses to products.
Factors Affecting Aesthetic Response
Aesthetic response is a complex phenomenon that is not yet well
understood. As with any complex phenomenon it is likely to involve a number
of factors; however, prior research seems to suggest that there are two
principal factors: the visual organization principle of unity and prototypicality,
that may significantly influence aesthetic responses. Some of the significant
prior research that suggests that these factors may play an important role in
influencing aesthetic responses will now be reviewed. The influence of
aesthetics/aesthetic response on non-aesthetic perceptions (i. e., derived
responses) will also be discussed.
Visual Organization Principles
The greatest void in aesthetic theory and research (and the greatest
opportunity) concerns the identification of specific factors that systematically
influence aesthetic responses. Although very general "rules" or "principles"
have sometimes been offered (e. g., "unity in variety" Hutcheson 1725), the
rules are usually vague and unspecified. Even though aestheticians seem to
rely on principles such as "unity" and "balance" and aesthetics research has
frequently employed scales that attempt to measure dimensions such as


62
TABLE III 8 continued
MEAN FAMILIARITY RATINGS AND UNITY SCORES
Product
(++)
(+-)
(-+)
(--)
UL
uN
7. Hair dryer
Sh
8.05
6.48***
4 46***
4.40***
0.83***
0.21
Tr
7.05
6.38**
6.54*
6.44**
0.33*
0.52**
Tx
7.67
3.98***
4.30***
5.69***
2.68***
2.02***
8. Dresser
Sh
7.76
6.62***
5.98***
6.29***
0.73**
0.56*
Tr
7.34
4.87***
4 97***
7.51
2.52***
2.95***
Tx
7.43
5.68***
5.63***
6.37*
1 18***
1.15***
9. Clock
Sh
7.28
4 51***
4 75***
6.43**
2.19***
2.09***
Tr
7.33
5.68***
5.78***
6.08***
0.97***
0.77**
Tx
7.05
5.44***
5.76***
6.86
1.32***
1.50***
Total0
7.45
5.53***
5.34***
5.79***
0 94***
Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
k Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.


76
TABLE III 17
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
NONLINEAR UNITY DIFFERENCE SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subject Effects
Intercept
1
5141.7559
530.00
0.0001
Version
2
428.1894
22.07
0.0001
Order
1
12.1427
1.25
0.2647
Version Order
2
41.8216
2.16
0.1187
Error Sub (Version Order)
188
1823.8659
Within Subiect Effects
Product
8
352.3060
9.50
0.0001
Trans
2
143.3854
15.47
0.0001
Prod Version (Residual)
14
506.0161
7.80
0.0001
Prod Order
8
146.6168
3.95
0.0001
Trans Order
2
52.9295
5.71
0.0100
Prod Version Order (Residual)
14
100.1989
1.54
0.1000
Error Sub (Prod)
1504
6971.6544
factors may indirectly influence "non-aesthetic" product evaluations by means
of aesthetic response.


20
H5: Objects (products) that are more novel (i. e., atypical) will
receive more favorable aesthetic response ratings than objects
that are less novel (i. e., more prototypical).
These hypotheses would predict that a product that exhibits a singular
change from the category prototype should receive more positive aesthetic
ratings than either the category prototype or a product that exhibits multiple
changes (i. e., extreme schema incongruity). While this result might be
expected if familiarity is entirely a function of memory, it is possible that the
unity design factor may influence feelings of "perceived familiarity." In an
instance where two attributes of a stimulus were altered such that the
stimulus was no longer prototypical but it did exhibit unity, the relational
similarity (i. e., unity) to the category prototype may generate a sense of
"perceived" familiarity (Goldstone, Medin, and Gentner 1991). This discussion
suggests the following hypothesis:
H6: Consistency with the visual organization principle of unity
is positively related to perceived familiarity.
In addition to the hypotheses that have already been presented
concerning the influence of specific factors on aesthetic response, there are
several other important hypotheses concerning the influence of the aesthetic
responses that are fostered by these factors. These hypotheses are developed
and presented in the section that follows.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Dressers (chest of drawers)
Dislike
2
3
-1
5
6
7 8
9
Like
Dislike
:
3
4
5
6
7
8
9
Like
U glv
i
3
4
3
6
7 8
q
Beautiful
Ugly
3
4
5
6
7
8
9
Beautiful
Familiar
2
3
4
5
6
7 8
9
Unfamiliar
Familiar
3
4
5
6
7
8
9
Unfamiliar
High Quality
p
3
4
5
6
7 8
9
Low Quality
High Quality
3
4
5
6
7
8
9
Low Quality
Low Price
2
3
4
5
6
7 8
9
High Price
Low Price
2
3
4
5
6
7
8
9
High Pnce
Dislike
-i
3
4
5
5 7
8
9
Like
Dislike
2
3
4
5
6
7
8
9
Like
Ugly
p
3
5
6 7
8
9
Beautiful
Ugly
2
3
4
5
6
7
8
9
Beautiful
Familiar
2
3
4
5
6 7
8
9
Unfamiliar
Familiar
-
3
4
5
6
7
8
9
Unfamiliar
High Quality
P
3
4
5
6 7
8
9
Low Quality
High Quality
p
3
:
6
p
8
9
Low Qualm
Low Price
3
4
f
6 7
8
9
High Price
Low Price
2
3
4
5
6
7
8
9
High Price
152


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Hair-drvers
Dislike
2
3 4
5
6
7
8
9
Like
Dislike
Ugly
3 4
5
6
7
8
9
Beautiful
Ugly
Familiar 1
-
3 4
5
6
7
8
9
Unfamiliar
Familiar
High Quality
3 4
5
6
7
8
9
Low Quality
High Quality
Low Price
o
3 4
5
6
7
8
9
High Pnce
Low Pnce
1 2 3
6 7 8 9 Like
6 7 8 9 Beautiful
2 3
6 7 8 9 Unfamiliar
1
6 7 8 9 Low Quality
3-156
9 High Pnce
Disiike
6 7
s
Q
Like
Dislike
3 3-5 6"
8
-
LiKe
Ugly
2 3 4 5
6 7
8
9
Beautiful
Ugly
-3-56"
8
0
Beautiful
Familiar
2 3 4 5
6 7
8
9
Unfamiliar
Familiar
3 3 5 6
8
9
Unfamiliar
High Quality
2 3 4 5
6 7
8
9
Low Quality
High Quality
2 3 56-
8
Q
Low Quality
Low Pnce
2 3 4 5
6 7
8
9
High Price
Low Pnce
2 3 5 6 7
8
9
High Pnce
150


169
Segall, Marshall H. (1976), "Visual Art: Some Perspectives from Cross-Cultural
Psychology," in Beyond Aesthetics: Investigations into the Nature of
Visual Art, ed. Don Brothwell, Thames and Hudson: London, 1976.
Sproles, George B. (1981), "Analyzing Fashion Life Cycles Principles and
Perspectives," Journal of Marketing. Vol. 45 (Fall), pp. 116-124.
Urban, Glen L. and John R. Hauser (1993), Design and Marketing of New
Products. Englewood Cliffs, NJ: Prentice-Hall, Inc.
Veryzer, Robert W. (1993), "Aesthetic Response and the Influence of Design
Principles on Product Preferences," in Advances in Consumer Research.
eds., Leigh McAlister and Michael L. Rothschild, Provo, UT: Association
of Consumer Research, Vol. 20, pp. 224-228.
Wallendorf, Melanie (1980), "The Formation of Aesthetic Criteria Through
Social Structures and Social Institutions," in Advances in Consumer
Research, ed., Jerry C. Olson, Ann Arbor, MI: Association for Consumer
Research, Vol. 7, pp. 3-6.
Whitney, Danial E. (1988), "Manufacturing by Design," Harvard Business
Review, (July-August), pp.83-90.
Wilkie, William L. (1990), Consumer Behavior, 2nd ed., New York: John Wiley
and Sons.
Wohlwill, Joachim F. (1981), A Conceptual Analysis of Exploratory Behavior:
The "Specific-Diverse" Distinction Revisited, in Advances in Intrinsic
Motivation and Aesthetics, ed. Hy I. Day, New York: Plenum Press, pp.
341-364.
Winer, B. J., Donald R. Brown, and Kenneth M. Michels (1991), Statistical
Principles in Experimental Design, New York, NY: McGraw-Hill, Inc.
Woll, Stanley and Arthur Graesser (1982), "Memory Discrimination for
Information typical and Atypical of Person Schemata," Social Cognition,
Vol. 1, No. 42, pp. 287-310.
Zaff, Brian S. (1987), The Cognitive Condition of Design, ed. Brian S. Zaff,
Symposium Proceedings (March), 1987.
Zajonc, R. B. (1980), "Feeling and Thinking: Preferences Need No Inferences,"
American Psychologist, 35, pp. 151-175.


APPENDIX B
STIMULUS SETS


TABLE IV 10
MEANS FOR FAMILIARITY RATINGS
100
Non-
Non-
Product
Unity/
Unity/
Unity/
Unity/
Typical
Atypical
Atypical
Atypical
(++)
(+-)
(-+)
(-)
UT,
Un
1. T. V. Remote
7.93
7.48
7.75
7.97
0.34
0.49
2. Flashlight
7.56
7.69
7.85
7.72
-0.13
0.03
3. Lamp
7.84
7.88
7.78
7.13
-0.35
-0.65
4. Refrigerator
8.45***
a
4 51***
5.77***
5.93***
2.05***
1.42**
5. Telephone
7.74
5.31
6.15
5.38***
0.83**
0.07**
6. Hair dryer
7.85
7.43
7.46
7.62
0.29
0.19
7. Dresser
7.54
7.21
7.27
6.82
-0.06
-0.39
8. Clock
8.09
7.57
7.70
8.07
0.45+
0.50
Total
7.87
6.88***
7.22***
7 08***
0.43***
0.21
a Indicates significance from Prototypical product.
b Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Flashlights
Dislike 1
6 7
8
Q
Like
Ugly 1
2 3 -i 5
6 7
8
9
Beautiful
Familiar i
2 3 4 5
6 7
8
9
Unfamiliar
High Quality 1
2 3 4 5
6 7
8
9
Low Quality
Low Pnce i
2 3 4 f
6 7
8
9
High Pnce
Dislike !23456'8 Like
Ugiy 1 2 3 4 5 6 8 0 Beautiful
Familiar 12345678a Unfamiliar
High Quality 12 3 456-89 Low Quaiir.
Low Pnce 12 3 4 5 6 7 8 9 High Pnce
Disiike 1
2 3-56
7 8
0
Like
Dislike
2 3-5
ft
0
Like
Ugly
2 5 5 6
7 8
9
Beautiful
Ugly
2 3 4 5
6
- Jj
9
Beautiful
Famihar
2 3-56
7 8
9
Unfamiliar
Familiar
2 3 4 5
6
7 8
0
Unfamiii
High Quality
2 3-56
7
9
Low Quality
High Quality
2 3 4 5
6
6
Q
Low Quaj
Low Pnce 1
2 3-56
~ s
9
High Pnce
Low Price
2 3-5
6
7 8
9
High Pnc
138


ACKNOWLEDGEMENTS
A number of people deserve credit for their role in helping me to
accomplish the goal of finishing this dissertation. First and foremost is
Dr. J. Wesley Hutchinson, the chairman of my committee, to whom I am
extremely grateful. The time and effort that he has contributed to this
endeavor are greatly appreciated. This work would not have been possible
without his benevolent guidance.
I am very grateful to Dr. Richard Lutz, Dr. Chris Janiszewski, and
Dr. David Mick for their- guidance in refining my ideas and for the
encouragement that they have given me. I would also like to thank
Dr. John Lynch, Jr. for his guidance on the statistical analyses and
Dr. Jonathan Hamilton for contributing an economists point of view.
I am indebted to my parents, Robert and Marion Veryzer, for their love
and support, and for always being there when I need them.
Finally, I would like to thank my brother David who has been a
tremendous source of inspiration to me.
in


17
higher than those for the more complex (novel) patterns, fell throughout the
tests until they finally flattened out (see Figure II 1). This would seem to
explain why (and roughly when) prototypicality will be liked better in some
cases and novelty will be better liked in others.
Mandler (1982) has theorized that the level of congruity between a
product and a more general product category schema may influence the nature
of information processing and thus product evaluations. Products that are
moderately incongruent with their associated category schemas are said to
stimulate processing that leads to a more favorable evaluation relative to
products that are either congruent or extremely incongruent with the category
schema. Mandler suggests that moderate incongruities are those that can be
successfully and readily resolved by the processor.
Meyers-Levy and Tybout (1989) conducted a series of experiments to test
Mandlers schema (in)congruity hypothesis. The general method, which was
modified slightly over the course of the three studies that were conducted,
consisted of presenting subjects with descriptions and samples of beverage/soft
drink products and having them evaluate the products along dimensions such
as appeal, taste, quality, interest in trial, etc. In their design, schema
congruity and schema incongruity were manipulated by varying a single
attribute in the product description (high preservatives vs. all natural
ingredients) so as to alter the structural and descriptive congruence of the
product description with the schema activated by the category label (beverage


7
uncommon in interdisciplinary research, the different orientations of the
disciplines that are or have been concerned with aesthetic phenomena and the
vagueness of the theories that have been offered to explain aesthetic responses
have severely inhibited progress toward gaining an understanding of aesthetic
responses.
Some aesthetic response theories, which seem to view aesthetic response
as idiosyncratic, maintain that their are no laws or principles of aesthetics
(Mothersill 1989), while others suggest that inner tendencies of the visual
system result in laws that govern perception and thereby influence aesthetic
response (Katz 1950; Koffka 1935). There are aesthetic theories that point to
fashion trends or the influence of culture as the determinant of systematic
aesthetic responses for all classes of products (Pleydell-Pearce 1970; Sproles
1981). There are views of aesthetic response that would suggest that the
consumers preference for a specific product was determined by a desire for
unity in variety (Auld 1981; Berlyne 1971; Lauer 1979). Other views maintain
that preference is related to prototypicality (Loken and Ward 1990; Nedungadi
and Hutchinson 1985). There are also those that periodically suggest that the
term "aesthetic response" applies only to works of art (Holbrook 1981), only to
be contradicted by those who maintain that all objects have an aesthetic
component (Berlyne 1974; Wallendorf 1980). Even the artists and product
designers who determine the forms of the objects seem to have trouble
agreeing on the basics of design (Lauer 1979) and as yet "have not formulated


57
TABLE III 6 continued
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES -
LINEAR UNITY SCORES
Product
Uj Score for
U3eauty
S. E.
t-test
7. Hair dryer
Sh
0.8940
0.4167
4.2902***
Tr
0.0303
0.3029
0.2000
Tx
3.0539
0.5295
11.5353***
8. Dresser
Sh
0.6137
0.3105
3.9526***
Tr
2.4077
0.4661
10.3311***
Tx
0.9925
0.5844
3.3966***
9. Clock
Sh
2.6462
0.4438
11.9243***
Tr
1.0682
0.4164
5.1308***
Tx
1.5531
0.3905
7.9536***
*** Indicates differences significant at p < .001 in one-tailed-tests
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.
manipulations is examined by looking at the difference between the
Unity/Atypical product and the lowest rated Non-Unity/Atypical product for
each stimulus set, all of the scores are positive and the scores for twenty-five
of the stimulus sets are significant at the p < .05 level (twenty-two of these are
significant at the p < .01 level or higher). These analyses suggest that even
though there are differences in the strengths of the manipulations due to
transformation type and product class, the effect of unity is relatively
consistent (i. e., significantly positive) across both products and
transformations.


170
Zeithaml, Valarie A. (1988), "Consumer Perceptions of Price, Quality, and
Value: A Means-End Model and Synthesis of Evidence," Journal of
Marketing, Vol. 52, pp. 2-22.


4
Consumer researchers have not developed one, and they have not found one
in any of the other disciplines to adopt as a starting point for building
aesthetic theory.
This dissertation attempts to address this void in aesthetic theory by
examining the influence of the aesthetic aspects of product appearance on
consumers product preferences and perceptions. This research represents
important steps toward understanding the implications of aesthetics and
product design for consumer research and the formulation of a theory of
consumer aesthetics.
In Chapter II the two principal factors, unity and prototypicality, that
influence consumers aesthetic responses to a products design are discussed
and the supporting research is reviewed. Several hypotheses are developed
that involve the relationship of the factors to aesthetic and derived responses.
Chapters III and IV each present the method and results of an
experiment designed to investigate the influence of the unity and
prototypicality factors on aesthetic and derived responses. Chapter III also
includes a detailed discussion of how the stimuli employed in all of the
experiments were developed.
Chapter V presents the method and results of an experiment that
focuses on derived responses.
Chapter VI provides a general discussion of the findings and
implications of these results for research that examines the influence of


APPENDIX E
PRODUCT DESCRIPTIONS FOR EXPERIMENT 3
"Stroner"
"Weak"
Bathroom Scales:
Bathroom Scales:
*5 Year Warranty
*3 Year Warranty
^Maximum Weight 330 lbs.
*Maximum Weight 320 lbs.
*Easy to Read Dial
*Easy to Read Dial
*5 lbs.
*5 lbs.
TV Remote Control
TV Remote Control:
*Controls 8 devices (TV, VCR,
*Controls 6 devices (TV, VCR,
Cable, CD, Satellite, and 3
Cable, CD, and 2 auxiliaries)
auxiliaries)
*Performs all functions of the
*Performs all functions of the
original remote control
original remote control
*On screen Programming Keys
*On screen Programming Keys
*Uses 4 AAA Batteries
*Uses 4 AAA Batteries
Flashlight:
Flashlight:
* Adjustable Beam (from spot to
*Dual Reflector System
flood)
*3 Year Warranty
*Lifetime Warranty
*Uses 2-D Batteries
*Uses 2-D Batteries
*Waterproof
* Waterproof
Lamp:
Lamp:
*2 way switch (high/low)
*Rotary on/off switch
*Uses up to 100 watt bulb
*Uses up to 100 watt bulb
*Flexible Gooseneck
*Flexible Gooseneck
*Weighted (desktop base)
*Weighted (desktop base)
Refrigerator:
Refrigerator:
*21.6 cu. ft.
*20 cu. ft.
*Freezer has Removable shelf
*Frostless
*Frostless
*Energy Efficient
*Energy Efficient
*6 Storage compartments
*6 Storage compartments
*2 Door Shelves
*2 Door Shelves
162


108
page (left or right side) was reversed for half of the subjects and was balanced
across questionnaire versions. The order in which the stimulus products (i. e.,
sets of target and control versions) were presented was also reversed for half
of the subjects and this was confounded with the product position
manipulation. Thus, the experiment was a l(target product with weak product
description) x 3(Product appearance: Unity/Prototypical, Non-unity/Atypical,
Unity/Atypical) x 9(Product) x 2(position/order) design that was conducted for
two dependent measures (quality or liking).
Experimental Procedure
Two-hundred and fifty seven subjects participated in this experiment.
The procedure that was followed was very similar to the procedures employed
in Experiments 1 and 2 except that in this study subjects were presented with
two product versions (control and target) on a page and each version was
accompanied with a product description. All conditions of this design were run
concurrently.
Results and Discussion
It was predicted that the unity aspect of product appearance would
influence subjects ratings of products in the same way that it had in earlier
experiments despite the addition of written product information. If this were
the case, one would expect that the particular product drawings (i. e.,


TABLE III 14 continued
MEAN PRICE RATINGS AND UNITY SCORES
72
Product
(++)
(+-)
(-+)
()
UL
uN
7. Hair dryer
Sh
5.17
4.58**
3.92**
4.28**
0.45**
0.75**
Tr
3.91
4.67
4.72
5.58
0.05
1.15***
Tx
5.03
4.44*
4.21**
5.87
1.08***
1.82***
8. Dresser
Sh
4.30
4.67
5.39
5.98
0.11
1.62***
Tr
4.39
4.10
4.36
6.33
1.08***
2.25***
Tx
4.65
4.42
4.80
5.32
0.31*
1 14***
9. Clock
Sh
4.61
3.09***
3.31***
4.39
1 24***
1.51***
Tr
3.90
3.33**
3.52+
3.92
0.46**
0.71**
Tx
4.09
3.20**
3.44**
3.94
0.68**
0.88***
Totalc
4.50
4.60c
4.55
5.33
0.29***
1 19***
a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.


CHAPTER VI
GENERAL DISCUSSION
The results of these three studies provide strong evidence of the
influence of unity on aesthetic responses. Unity was also shown to
significantly influence attitudes toward products, perceived familiarity, and in
some situations perceptions of product quality. The context of the evaluation
situation (i. e., whether or not other product variations or written product
descriptions were presented for comparison along with the product being
evaluated) was shown to moderate the influence of unity (i. e., aesthetic
response) on perceived quality. In addition to these insights concerning the
influence of unity on aesthetic and derived responses, these studies help to
clarify the role of prototypicality in influencing aesthetic and derived
responses. Although prototypicality did seem to have a positive effect on
aesthetic responses and attitudes toward products, it did not seem to have a
positive effect on perceived quality or price expectations. In fact,
prototypicality seemed to have a negative effect on price expectations.
In Experiment 1 a products consistency with the unity visual
organization principle was shown to significantly influence subjects aesthetic
responses toward products. The favorable aesthetic response generated by a
116


41
four of the product versions that were shown on each page. Subjects were then
allowed to proceed through the task of rating the products at their own pace.
The subjects rated each product design on 9-point semantic differential scales
that measured aesthetic response (beautiful/ugly), familiarity
(familiar/unfamiliar), attitude (like/dislike), quality (high quality/low quality),
and price (high price/low price).^ Following the rating tasks, subjects were
asked to write about how they determined the ratings that they gave to the
proposed products. The entire procedure took approximately one-half hour to
complete.
Results and Discussion
The manipulation check of the effect of a change in either Feature A or
Feature B on aesthetic responses found that there was, in fact, a reduction in
ratings of the appearance of the products (compared to those of the prototypical
product version) as measured on the semantic differential scale anchored by
Beautiful/Ugly. This is to say that in the case of the product variations that
were altered so that one feature (either Feature A or Feature B) was no longer
prototypical or unified the effect of the change was a reduction in positive
aesthetic response. The average effect of a change in Feature A on aesthetic
responses was -1.24. Likewise, the average effect of a change in Feature B on
4
The direction of some of these scales was reversed in order to reduce
possible response bias on the part of subjects.


24
very minor motivations. Ratings on ten attributes showed the basis of
consumer preference in a more systematic way. Overall preference ratios were
as follows: more like fresh 3.2; more natural 3.1; taste/flavor 2.3;
quality -- 2.3; size -- 1.7; texture -- 1.5; shape 1.5; and appearance 1.1.
Panelists did not misperceive criteria like ease of preparation (1.0) and ease
of holding (.8) for which shape had objectively little impact.
Berkowitz suggests that the data seem to indicate a chain of interrelated
inferences which stem from the shape of the product rather than a single
direct linkage. He notes that the findings indicate that an attribute
communicated and presumably noticed may not be considered by consumers
to be discriminating, but the attribute(s) that it triggers may be considered to
be discriminating. Thus, the squared-off shape of the one test item may have
fostered an association (or cued categorization) with processing or processed
products while the more natural looking product may have been associated
with freshness or "fun experiences such as summer family barbecues when
fresh corn was served" (Berkowitz 1987, p. 559). Berkowitz notion of
"interrelated inferences" is similar to the concepts of perceptual categorization
and perceptual inferences (Wilkie 1990). Wilkie (1990) has pointed out how
these processes, which translate sensory inputs into a mental "identification"
of a particular stimulus and develop (i. e., construct) beliefs concerning the
stimulus based on other information such as stimulus properties, lead to


163
APPENDIX E continued
"Strone"
"Weak"
Telephone:
Telephone:
*10 Number Memory
*6 Number Memory
*Receiver and Ringer volume
*Receiver and Ringer volume
control
control
*Mute and Redial Buttons
*Mute and Redial
*3 lbs.
*3 lbs.
Hair dryer:
Hair dryer:
*1,600 watts
*1,500 watts
*2 speeds and 3 heat settings
*2 speeds and 2 heat settings
*Dual Voltage
*Dual Voltage
*Super-quiet operation
*Super-quiet operation
*2 lbs.
*2 lbs.
Dresser:
Dresser:
*2 Spacious Drawers
*2 Spacious Drawers
*Drawers have steel slides with
*Metal on wood Drawer guide
ball bearings for smooth
*Durable chip resistant finish
operation
*Hardwood Back
*Durable chip-resistant finish
*Hardwood Back
Clock:
Clock:
*Illuminated Dial (2 brightness
*Illuminated Dial
settings)
*Snooze Button
*Snooze Button
*Adjustable soft/loud Alarm
* Adjustable soft/loud Alarm
*9 v Battery Backup
*9 v Battery Backup


90
TABLE IV 2
MEANS FOR BEAUTY RATINGS
Product
Unity/
Typical
(++)
Non-Unity/
Atypical
(+-)
Non-Unity/
Atypical
(-+)
Unity/
Atypical
(-)
uT,
Un
1. T. V. Remote Tr
4.36
4.36
4.70
3.88
-0.41
-0.48
2. Flashlight Tr
4.85
4.95
4.79
5.07
0.09
0.28
3. Lamp Tx
5.00
4 14**a
4.59**
4.98
0.63**
b
0.84
4. Refrigerator Sh
4.16
2.80***
3.33*
3.41
0.72**
0.61
5. Telephone Sh
3.97
3.28
3.64
3.62
0.35
0.34
6. Hair dryer Tx
5.23
4.74*
4.43
4.23
0.15
-0.20
7. Dresser Tr
4.03
3.86
4.05
4.12
0.12
0.26
8. Clock Sh
3.03
3.10
2.85
2.71
-0.11
-0.14
Total
4.33
3.90***
4.05**
4.00**
0.19*
0.19
~~ Indicates significance from Prototypical product.
Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.
p < .025. The linear unity scores formed from the mean product ratings
indicated that six of the eight stimulus sets were in the expected direction (i.
e., positive). The "mean" nonlinear unity scores indicated that five of the eight
stimulus sets were in the expected direction. Follow-up tests for each product
indicated that individually only two products were significant [Lamp F(l,233)
= 8.85, p < .0032; Refrigerator F(l,233) = 7.74, p < .0058]. However, the fact


123
balance, proportion, etc. Experience with these regularities would most
certainly influence subsequent perception and over time become internalized.
In this way people may develop a level of familiarity with and preference for
certain relations (i. e, regularities; e. g., unity, symmetry). Goldstone, Medin,
and Gentner (1991) have investigated how relational similarities (i. e.,
descriptions of connections between two or more objects or attributes; e. g.,
same color) affect similarity judgments. Their work shows that relations and
attributes are psychologically distinct and that relations can significantly affect
similarity judgments. Thus, relations (i. e., visual organization principles)
which are learned (consciously or nonconsciously) would seem to have the
potential to influence judgments of a products prototypicality.
There is also a need for further investigation into the systematic nature
of aesthetic response. This work as well as others (e. g., Veryzer 1993)
indicates that the systematic nature of aesthetic response in the visual domain
stems from the perceptual tendency toward organization. This tendency has
been studied by the Gestalt psychologists and aestheticians and is the basis for
general rules of perception such as the Gestalt laws (e. g., proximity,
similarity) and design principles (e. g., unity, contrast, proportion). While the
source of the visual organization principles (i. e., design principles, Gestalt
laws) that operate to organize perception is still open to debate, there is
evidence that the perception principles are present very early in life and that
preferences related to these principles develop over time and may be modified


TABLE IV 5
MEANS FOR QUALITY RATINGS
94
Product
Unity/
Typical
C++)
Non-Unity/
Atypical
(+-)
Non-Unity/
Atypical
(-+)
Unity/
Atypical
(--)
UL
UN
1. T. V. Remote
5.26
5.28
5.50
5.00
-0.26
-0.28
2. Flashlight
5.13
5.29
5.24
5.13
-0.14
-0.11
3. Lamp
4.72
4.33*a
4.67
5.05
0.39+b
0.72
4. Refrigerator
4.76
4.45
4.09+
4.50
0.36
0.41
5. Telephone
4.72
4.31*
4.38*
5.09
0.56*
0.78
6. Hair dryer
5.85
5.47
5.38
5.15
0.08
-0.23
7. Dresser
4.30
4.03**
4.33
5.10
0.52*
1.07*
8. Clock
2.93
3.58
3.34
3.22
-0.39
-0.12
Total
4.71
4.59
4.62
4.78
0.14+
0.28
^ Indicates significance from Prototypical product.
Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.
The linear and nonlinear unity scores for liking both indicated that the
mean product ratings for six of the eight stimulus sets were in the expected
direction (i. e., positive). Five of the eight stimulus sets displayed positive
linear unity scores for quality ratings. Four of the nonlinear unity scores
formed from the mean product ratings were positive. Although the unity
scores for the price ratings indicated that five of the linear unity scores and six


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
TV Remote Controls
/ taacHi if
C2QQ ¡I
/ aaa //
ocjca f
m
fags)
/ sso j
/ aa a //
/ aa a if
C3E3C3 f
iaipcy /
/ /
V
/
Dislike 5 6 8 9 Like
L civ 1 2 5 4*6789 Beautiful
Famiiiar 12 3 456789 Unfamiliar
High Quality 12 3 456789 Low Qualitv
Low Pnce 12 3 456789 High Pnce
Dislike 1 2 3 4 5 6 7 8 4 Like
Ugly 1 234 r 6789 Beautiful
Familiar 12 3 4 5 6 7 8 9 Unfamiliar
High Quality 1 2 3 4 5 6 7 8 9 Low Quality
Low Pnce 1 2 3 4 5 6 7 8 9 High Pnce
SBSi|
/ C2C2Q i
/ aut a //
/ C3QQ //
/ mr-vtf S
/ /ao' //
/ L3LJC3 h
/ Lsaa j
/ C3CJL3 //
/ OOZ3 //
j jqS 1
Dislike 12 5 4J6789 Like
Ugly 1 2 3 4 5 6 7 8 9 Beautiful
Fanuhar 1 2 3 4 5 6 8 9 Unfamiliar
High Quality 12 3 456_S9 Low Qualm
Low Pnce 123456789 High Price
Dislike 1 234567s Like
Ugly 1 2 3 -1 5 6 7 8 Beautiful
Farruhar 2 3 4 ; 6 7 8 9 Unfamiliar
High Quality : 234 5 *780 Low Quality
Low Price 1 2 3 4 5 6 7 8 9 High Pnce
133


TABLE III 12
MEAN LIKING RATINGS AND UNITY SCORES
67
Product
(++)
(+-)
(-+)
(-)
UL
UN
1. Bathroom scale
Sh
6.31
5.92**a
5.06**
4.92**
0.12
0.83***
Tr
4.89
5.91
5.22
6.49
0.13+b
1 71***
Tx
4.47
6.44
5.25
5.28
-0.97
0.56+
2. TV Remote Control
Sh
5.37
4.58*
5.09
5.20
0.40*
1 42***
Tr
4.33
4.73
5.44
7.09
0.62**
2 74***
Tx
5.97
4.61**
5.00**
5.74
1.10**
1.68***
3. Flashlight
Sh
7.08
4.09***
3
3 4y***
1.53***
0.78**
Tr
6.16
3.95***
4.53***
4 26***
0.96***
0.77**
Tx
6.72
2.92***
2 40***
5.48**
3.36***
3.23***
4. Lamp
Sh
6.56
5.95*
4.34***
4.53***
0.37**
0.66**
Tr
5.08
4.80
6.02
6.25
0.26*
1.61**
Tx
5.36
5.30
4.98
6.73
0.91**
2.58***
5. Refrigerator
Sh
6.97
2.06***
2.46***
3.98***
3.25***
2 14***
Tr
6.13
4.14***
4 20***
6.05
1.96***
2.55***
Tx
6.89
2.31***
2.37***
3.40***
2.72***
1.36***
6. Telephone
Sh
5.98
3.03***
4 20***
3.63***
1 19***
0.92**
Tr
6.02
5.65+
4 29***
4.40***
0.26*
0.44**
Tx
6.05
5.26**
4.97**
6.42
1 io***
1 76***


TABLE IV 9
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE
98
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
11.7739
0.77
0.5127
Order
1
11.2379
2.20
0.1393
Version Order
3
17.6058
1.15
0.3301
Error Sub (Version Order)
229
1169.4883
Within Subiect Effects
Product
7
1224.6204
73.91
0.0001
Feature A
1
1.6873
0.71
0.2500
Feature B
1
18.6191
7.87
0.0100
A B
1
0.5761
0.24
0.2500
Prod Version
18
43.3833
1.02
0.2500
Prod Order
7
64.1632
3.87
0.0003
A Order
1
0.0001
0.00
0.2500
B Order
1
0.1359
0.06
0.2500
A B Order
1
0.3289
0.14
0.2500
Prod Version Order (Residual)
18
49.1493
1.15
0.2500
Error Sub (Product)
1603
3794.1377
having all versions appear on the same page, could explain the lack of a
12
significant unity effect for quality and price. The alternative explanation,
that the absence of a comparative context increases the importance of
12
Partial omega squares (Keppel 1991, p. 223) were calculated in order to
estimate the magnitude of the unity effect. The partial omega squares
for the effect of unity on beauty ratings for Experiment 1 were .31 for
the rating score and U^ approaches and .37 for the score approach.
The partial omega square for the effect of unity on ratings of product
beauty for Experiment 2 was .006. Although these estimates indicate
that the effect of unity is quite large in Experiment 1 and very small in
Experiment 2, care should be taken in interpreting the importance of
these effects. See Keppel (1991, pp. 66-68 and 224) for discussions
concerning the importance of small effects and the difficulty of
comparing the sizes of omega squared estimates.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Clocks
Dislike
- -
5
6 '
6
9
Like
Dislike
;
3 4
<
6
7
8
9
Like
Ugly
5
6 7
S
9
Beauuful
Ugly
*i
3 4
5
6
7
8
9
Beauuful
Familiar
2
? 4
5
6
8
9
Unfamiliar
Familiar
I
3 4
5
6
7
8
9
Uni amiiiar
High Quality
3 4
5
6 7
6
9
Low Quaiiry
High Qualirv
:
3 4
5
6
7
8
9
Low Quality
Low Pnce
2
3 4
5
6 7
8
9
High Pnce
Low Pnce
2
3 4
5
6
7
8
9
High Pnce
Dislike
: ?
4 5

s
9
Like
Dislike
z
3 4
j
6 7
8
Q
Like
Ugly
-
d 5
6 7
8
9
Beautiful
Ugly
:
3 4
5
6 7
8
O
Beautiful
Familiar
: 3
4 5
6 "
2
0
Uni amiiiar
F amiiiar 1
2
3 4
5
6 7
8
0
Unfamiliar
High Quaiiry
- 3
4 5
6 7
S
9
Low Quaiirv
High Quality
2
3 4
5
6 7
8
9
Low Quaiiry
Low Pnce
2 3
4 5
6 7
8
9
High Price
Low Pnce
2
3 4
;
6 7
8
0
High Pnce
154


2
more and more manufacturers of all kinds of products are having to come to
terms with the reality that product appearance is often the major factor
influencing (consciously or unconsciously) what people buy and how much they
are willing to pay (Oakley 1990, p. 5).
The influence of aesthetic factors on product preferences and perceptions
is an important but often neglected area of study in consumer research. The
design of products inherently involves aesthetics. The aesthetic aspects of a
product give rise to the registering of affect or pleasure due to the conscious
or unconscious influences of the characteristics of the product (Holbrook and
Zirlin 1985). The response arising from the interaction between the aesthetic
aspects of an object and the perceiver of the object has been termed the
"aesthetic response" (Olson 1981). There is a growing recognition among
marketing researchers that aesthetic responses can significantly affect
consumer behavior and thus aesthetics is gaining recognition as an important
marketing variable (Kotler and Rath 1984; Wallendorf 1980). The influence
of aesthetics is increasingly being acknowledged as an important part of new
product development (Whitney 1988), marketing strategy (Kotler and Rath
1984), product quality (Garvin 1984; Zeithaml 1988), product differentiation
(Dickson and Ginter 1987), and competitive advantage (Holt 1985; Kotler and
Rath 1984).
Despite the growing awareness of the significant influence that the
aesthetic aspects of products can have on product preferences and perceptions,


CHAPTER II
CONCEPTUAL BACKGROUND ON THE FACTORS THAT INFLUENCE
AESTHETIC RESPONSES
Introduction
Even though there is a growing awareness of the important role that
aesthetics play in influencing product preferences, our understanding of
aesthetic responses is extremely limited. Our primitive understanding of
aesthetic response phenomena may be due to the highly fragmented approach
that has characterized its study. Aesthetics have been studied in a number of
fields including philosophy, art history, psychology, experimental aesthetics,
industrial design, and more recently consumer behavior. Each field has
contributed to our comprehension of aesthetics and yet, little progress has been
made in understanding the specifics of aesthetic response (Berkowitz 1987;
Wohlwill 1981). Within the many conceptualizations of aesthetic response that
have emerged from these different fields of study, there is little agreement and
limited insight regarding why a particular object is perceived as pleasurable
or beautiful while another is viewed as unattractive. The question concerning
what makes an object aesthetically pleasing, regardless of whether it is a
painting or a portable stereo, has received diverse and typically vague answers
from the disciplines that have studied it. While a lack of convergence is not
6


114
TABLE V 4
ANALYSIS OF VARIANCE QUALITY
Source
DF
Type III SS
F-Value
P > F
Between Subjects Effects
Intercept
1
2.5064
0.68
0.4101
Covariate
1
474.7499
129.42
0.0001
Version
2
21.0677
2.87
0.0604
Order and Position
1
4.7474
1.29
0.2575
Version Order
2
1.1033
0.15
0.8605
Error
123
451.1998
Cell
2
0.6545
0.22
0.2500
Prod Version
14
200.7009
9.85
0.0010
Error (Product)
984
1431.8646
Cell Order
2
0.3782
0.13
0.2500
Prod Version Order
14
47.7123
2.34
0.1000
Error (Product)
984
1431.8646
(M = 5.98) and the mean for Non-unity/Atypical products (M = 5.79) significant
F (1,123) = .24, p > .25). Seven of the nine nonlinear unity scores were in the
expected direction. Follow-up tests on the individual products indicated that
the contrast between the Unity/Atypical and Non-unity/Atypical product
versions was only significant for the refrigerator product F (1,123 = 6.84, p <
.01).
Summary
It was predicted that the unity aspect of product appearance would
influence subjects attitudes towards products and their perceptions of product
quality despite the availability of written attribute information. The results


166
Holbrook, Morris B. and Robert B. Zirlin (1985), "Artistic Creation, Artworks,
and Aesthetic Appreciation: Some Philosophical Contributions to Non
profit Marketing," Advances in Nonprofit Marketing, Vol. 1, pp. 1-54.
Holt, S. (1985), "Design, the Ninth Principle of Excellence: The Product Half
of the Business Equation," Innovation, (Fall), pp. 2-4.
Huber, Joel, and Morris B. Holbrook (1981), "The Use of Real Versus Artificial
Stimuli in Research on Visual Esthetic Judgements," in Symbolic
Consumer Behavior, eds. Elizabeth C. Hirschman and Morris B.
Holbrook, Ann Arbor, MI: Association for Consumer Research, pp. 60-68.
Hutcheson, F. (1725), An Inquiry into the Original of Our Ideas of Beauty and
Virtue, London: Darby.
Hutchinson, J. Wesley and Joseph W. Alba (1991), "Ignoring Irrelevant
Information: Situational Determinants of Consumer Learning," Journal
of Consumer Research, Vol. 18, (December), pp. 325-345.
Johnson, Mark (1987), The Body in the Mind, Chicago: The University of
Chicago Press.
Katz, David (1950), Gestalt Psychology, New York: The Ronald Press
Company.
Keppel, Geoggrey (1991), Design and Analysis: A Researchers Handbook, 3rd
ed., Englewood Cliffs, NJ: Prentice-Hall, Inc.
Koffka, K. (1935), Principles of Gestalt Psychology, New York: Harcourt, Brace
and Company.
Kohler, Wolfgang (1929), Gestalt Psychology: An Introduction to New Concepts
in Modern Psychology, New York: Liveright Publishing Corporation
(1947).
Kotler, Philip and G. Alexander Rath (1984), "Design: A Powerful But
Neglected Strategic Tool," The Journal of Business Strategy, Vol. 5
(Fall), pp. 16-21.
Kunst-Wilson, W. R. and R. B. Zajonc (1980), "Affective Discrimination of
Stimuli That Cannot Be Recognized," Science, Vol. 207, pp. 557-558.


CHAPTER V
EXPERIMENT 3
Overview
Experiment 2 examined the influence of unity and prototypicality on
aesthetic and derived responses when products were evaluated without the
context of other product versions from the same product category. The
experiment described in this chapter is a further investigation of the influence
of product aesthetics on derived responses. In this experiment, subjects are
provided with written product descriptions as well as drawings of the product
versions that they are asked to rate. The experiment examines whether or not
there is a moderating effect of the additional information (i. e., written product
descriptions) on the influence of unity. It is expected that as in the experiment
discussed earlier, unity will significantly influence subjects ratings of products
even though the evaluation task (i.e., rating product versions on scales
measuring either liking or quality) does not explicitly require consideration of
product appearance. In addition to demonstrating the influence of unity on
derived responses, such a finding would also provide evidence of the
importance of product aesthetics in general.
104


Please indicate vour reaction to the appearance of the
products by marking (circling) the scales below each.
proposed
Bathroom scales
Dislike
12 3 4
5 ei s
0
Like
Ugly 1
2 3 4
5 6 ~ i
9
Beautiful
Familiar 1
2 3 4
5 6'8
9
Unfamiliar
High Quality 1
2 3 4
5 6^8
9
Low Quality
Low Price 1
2 3 4
5 6 S
9
High Pnce
Dislike
Ugly
Familiar
High Quality
Low Pnce
3 4 5 6 7 8 0 Like
3 4 5 6 7 8 9 Beautiful
3 4 5 6 7 8 0 Unfamiliar
3 4 5 6 7 8 9 Low Quality
3 4 5 6 7 8 9 High Pnce
Dislike 1 2 3 4 5
Ugly 12 3 4 5
Familiar 12 3 4 5
High Qualiry 1 2 3 4 5
Low pnce 12 3-5
6 £ 9 Like
7 8 0 Beautiful
7 8 9 Unfamiliar
- 89 Low Quality
^ 0 High Pnce
Dislike
Ugh
Familiar
High Quajjr\
Low Pnce
6 1 s q Like
8 9 Beautiful
6 8 9 Unfamiliar
8 0 Low Quailr\
* > 9 High Pnce
130


21
Derived Responses
Attitude
Aesthetic responses seem to influence derived responses (i. e., non-
aesthetic evaluations) although their influence is often attributed to other
factors (e. g., Berkowitz 1987). One area that seems to be influenced by
aesthetic responses is that of attitude. Bell et al. (1991) have suggested that
aesthetic responses are key determinants of general liking. In their study,
which was described earlier, they found evidence that general liking was
influenced by aesthetic response (recall that aesthetic response in their study
was influenced by perceived unity which was manipulated by means of the
ensemble that was depicted).
In general, "attitudes" may be conceptualized as evaluative judgements.
The term is usually used to refer to an individuals disposition to respond
favorably or unfavorably to an object, person, institution, or event, or to any
other discriminable aspect of an individuals world (Ajzen 1989, p. 241). Ajzen
(1989) has noted that affective reactions may feed into the overall evaluative
response to an attitude object and thus may be at least partly responsible for
the evaluative direction and intensity of a persons beliefs.
The influence of affect due to aesthetic factors (i. e., happy or angry
looking faces, taste, smell) on attitude formation and change was examined in
a study by Edwards (1990). In two experiments, Edwards examined the
hypotheses that the sequence of affect and cognition in an attitudes formation


TABLE IV 6
MEANS FOR PRICE RATINGS
95
Product
Unity/
Typical
(++)
Non-Unity/
Atypical
(+-)
Non-Unity/
Atypical
(-+)
Unity/
Atypical
(-)
uT.
Un
1. T. V. Remote
4.31
4.71
4.95
4.33
-0.51
-0.38
2. Flashlight
3.80
3.93
3.83
4.03
0.04
0.20
3. Lamp
3.51
3.59
3.72
4.03
0.12
0.44
4. Refrigerator
5.05
5.30
4.92
5.02
-0.08
0.10
5. Telephone
4.88
5.31
4.75
5.55
0.19
0.80
6. Hair dryer
4.57
4.67
4.29
4.64
0.13
0.35
7. Dresser
4.18
4.17
4.19
4.50
0.16
0.33
8. Clock
2.19
2.68
2.61
2.45
-0.33
-0.16
Total
4.06
4.29
4.16
4.32
-0.04
0.21
a Indicates significance from Prototypical product.
Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
Indicates that the atypical condition was rated more favorably than the
prototypical condition at the p < .05 level.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.
of the nonlinear unity scores were positive, these scores reflect the influence
of the high ratings received by the Unity/Atypical product variations. Across
the eight stimulus sets the Unity/Prototypical product variations tended to
receive price ratings that were lower than those received by the Non-
Unity/Atypical product variations. This pattern (i. e., novelty effect) is
consistent with the results reported for Experiment 1. Follow-up tests for each


TABLE V 2
ANALYSIS OF VARIANCE LIKING
112
Source
DF
Type III SS
F-Value
P > F
Between Subjects Effects
Intercept
1
181.0534
24.72
0.0001
Covariate
1
129.7023
17.71
0.0001
Version
2
50.1938
3.43
0.0357
Order and Position
1
5.0247
0.69
0.4092
Version Order
2
18.5099
1.26
0.2864
Error
120
878.9758
Cell
2
18.6838
4.52
0.0250
Prod Version
14
156.2446
5.39
0.0100
Error (Product)
960
1984.9216
Cell Order
2
0.1300
0.03
0.2500
Prod Version Order
14
27.1226
0.94
0.2500
Error (Product)
960
1984.9216
Unity/Atypical and Non-unity/Atypical product versions was significant only
for the refrigerator product F (1,121 = 28.41, p < .0001). These results indicate
that product appearance, and particularly unity, does influence subjects
attitudes toward products even when written attribute information (i. e.,
technical information) is available concerning the products.
The means for the quality ratings are presented in Table V-3 and the
ANOVA for the effect of unity on subjects perceptions of quality is presented
in Table V-4. Although the pattern exhibited by the means is consistent with
the prediction that the unity aspect of product appearance influences perceived
quality, the effect of unity was not significant across products F (2,984 = .22,
p > .25), nor was the contrast between the mean for Unity/Atypical products


59
TABLE III 7 Continued
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES -
NONLINEAR UNITY SCORES
Product
Ujyr Score for
Beauty
S. E.
t-test
7. Hair dryer
Sh
0.6061
0.2470
2.4536**
Tr
0.9242
0.2226
4.1516***
Tx
2.9385
0.3375
8.7074***
8. Dresser
Sh
1.8030
0.2416
7.4624***
Tr
3.2615
0.2375
13.7356***
Tx
1.4242
0.4108
3.4666**
9. Clock
Sh
2.5846
0.2556
10.1118***
Tr
1.0303
0.2471
4.1696***
Tx
1.7121
0.2068
8.2777***
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.
While an interaction of order with the effect of unity (i. e., the Feature
A x Feature B interaction) was not indicated, order was found to interact with
factors such as Product. In part, the effect of order seems to be attributable
to subjects increasing acceptance of the use of the simplified (i. e., line
drawings without color) drawings to depict products. Such an explanation is
suggested by the fact that, in general, when the stimuli for a particular
product category were rated toward the end of the rating task they received
slightly higher ratings than when they were rated at the beginning of the


82
these effects using three very different types of (product) transformations
underscores the fact that unity can (and does) have broad application as a
design factor.


Pleass indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
-Telephones
Dislike
3
4
5
6
7
8
Q
Like
Dislike
i
4 5
6
7
8
9
Like
Lgiv
2
3
4
5
6
7
S
9
Beautiful
Ugly
) ^
4 5
6
7
8
9
Beautiful
Familiar
3
4
5
6
7
8
9
Unfamiliar
Familiar
2 3
4 f
6
7
8
9
Unfamiliar
High Quality
3
4
5
6
7
8
9
Low Quality
High Quality
2 3
4 5
6
7
8
9
Low Quality
Low Pnce
2
3
4
5
6
7
8
9
High Price
Low Pnce
; ?
4 5
6
7
8
9
High Pnce
Dislike
;
3 4
5
6 7
8
9
Like
Disiike
I 3 4 5
7
8
9
Like
Ugiy
2
3 4
5
6 7
8
9
Beautiful
Ugly 1
7
8
9
Beautiful
Famiuar
2
3 4
5
6 7
s
9
Unfamiliar
Familiar
3 3 4 5 6
7
8
9
Unfamiliar
High Quality
2
2 2
5
6 7
8
9
Low Quality
High Quality 1
2 3-56
7
8
9
Low Quality
Low Pnce
3 4
5
6 7
S
9
High Price
Low Price
2 3,56
7
8
9
High Price
146


22
is an important determinant of its subsequent resistance to affective and
cognitive means of persuasion. Affect-based and cognition-based attitudes were
induced and subsequently challenged by either affective or cognitive means of
persuasion. The procedure used to create the two types of attitudes and the
means of persuasion involved varying the sequence of affect and cognition
while holding the content of communications constant. Edwards found that
affect-based attitudes exhibited more change under affective means of
persuasion than under cognitive means of persuasion. Cognition-based
attitudes, on the other hand, exhibited equal change under both forms of
persuasion. In addition, it was found that affect-based attitudes were
expressed with greater confidence than their cognition-based counterparts.
These findings demonstrate that aesthetic responses can play an important
role in determining attitudes.
H7: Attitudes (e. g., general liking) towards products are more
favorable for products receiving more positive aesthetic ratings
than for products receiving less positive aesthetic ratings.
Perceived Quality
Another important area where aesthetic responses seem to play an
important role in influencing non-aesthetic evaluations is that of product
quality. Garvin (1984) has identified aesthetics as one of the eight dimensions
of quality. Zeithaml (1988) has suggested that "intrinsic cues," which involve


165
Cohen, J (1977), Statistical Power Analysis for the Behavioral Sciences. (Rev.
ed.), New York: Academic Press, 1977.
Dickson, Peter R. and James L. Ginter (1987), "Market Segmentation, Product
Differentiation, and Marketing Strategy," Journal of Marketing. (April),
pp. 1-10.
Edwards, Kari (1990), "The Interplay of Affect and Cognition in Attitude
Formation and Change," Journal of Personality and Social Psychology,
Vol. 59, pp. 202-216.
Eisenman, R. (1968), "Novelty Ratings of Simple and Complex Shapes,"
Journal of General Psychology. 78, pp. 275-278.
Garvin, David A. (1984), "What Does "Product Quality" Really Mean?" Sloan
Management Review, (Fall), pp. 25-43.
Glass, Lewis A. and Keith J. Holyoak (1986), Cognition, Second ed., New York,
N.Y.: Random House.
Goldstone, Robert L., Douglas L. Medin, and Dedre Gentner (1991), "Relational
Similarity and the Nonindependence of Features in Similarity
Judgements," Cognitive Psychology, Vol. 23, pp. 222-262.
Gorski, D. (1987), The Cognitive Condition of Design, ed. Brian Zaff,
Symposium Proceedings (March) 1987.
Gruenwald, George (1985), New Product Development: What Really Works,
Chicago, IL: Crain Books.
Hartmann, George W. (1935), Gestalt Psychology: A Survey of Facts and
Principles, New York: The Ronald Press Company.
Holbrook, Morris B. (1981), "Introduction: The Esthetic Imperative in
Consumer Research," in Symbolic Consumer Behavior, eds. Elizabeth C.
Hirschman and Morris B. Holbrook, Ann Arbor, MI: Association for
Consumer Research, pp. 36-37.
Holbrook, Morris and Elizabeth Hirschman (1982), "The Experiential Aspects
of Consumption: Consumer Fantasies, Feelings, and Fun," Journal of
Consumer Research, Vol. 9 (2), pp. 132-140.


117
products consistency with the unity visual organization principle carried over
into subjects attitudes toward products, perceptions of product quality, and
product price expectations. In addition, unity was shown to influence subjects
perceptions of product familiarity. The effect of unity was quite robust as
these results were replicated across nine product categories using three
different types of product transformation. The results of Experiment 1 provide
a clear indication that the unity visual organization principle is an important
design factor that can significantly influence aesthetic responses and non-
aesthetic product evaluations.
Experiment 2 found that unity significantly influenced aesthetic
responses even when products were evaluated without the context of other
product versions from the same product category. The favorable aesthetic
response generated by consistency with the unity visual organization principle
was also shown to carry over to attitudes toward products and perceived
familiarity. Although the overall pattern of the means for the quality ratings
of the products was positive for the majority of the stimulus sets tested, it was
not significant. This experiment, which provided a between-subjects test of the
influence of the unity and prototypicality factors, indicates that just as in
Experiment 1 unity provides a better explanation of aesthetic response than
does prototypicality. This between-subjects replication further demonstrates
that the effects of unity are quite robust.


55
factors are also indicated in the Product x Version (Residual) interactions.8
The differential strength of the manipulations across stimulus sets was further
examined in order to ensure that the effects were in the right direction across
the twenty-seven stimulus sets. This was done by conducting t-tests using the
"unity" scores for each of the stimulus sets in order to assess the degree to
which the effect of unity was significant and in the proper direction across the
twenty-seven stimulus sets despite the differences in the strength of the
manipulations. The analysis for the linear unity scores is shown in Table III-6
(the unity scores were also presented in Table III-2). In twenty-one out of the
twenty-seven stimulus sets the linear unity scores are significantly different
from zero (p < .05; p < .01 for nineteen of the stimulus sets). In six cases the
interaction scores were not significant and in only one case was a score
negative and significant (i. e., in the "wrong" direction).^ The findings are
similar when the analysis is conducted using nonlinear unity scores. This
analysis is presented in Table III-7. When the differential strength of the
Q
In this design the main effect of Transformation type is examined using
a Latin square type orthogonal fraction of the complete five factor
design. This is consistent with the treatment suggested by Winer,
Brown, and Michels (1991, pp. 706-711) for Latin squares and related
designs.
Q
This negative case occurred for the texture manipulation of the
bathroom scale stimulus set. All three of the bathroom scale
product/cases were found to be somewhat problematic because this
product was not easily divided into two separate parts (e. g., telephone
stimulus handset, base). This was particularly true for the texture
manipulation of the bathroom scale (see bathroom scale/texture stimuli
in Appendix B).


43
TABLE III 1
MANIPULATION CHECK FOR THE EFFECT OF A CHANGE
IN FEATURE A OR B ON BEAUTY RATINGS
Product
Feature A
Feature B
1. Bathroom scale
Sh
-0.14
-0.17
Tr
1.00
0.14
Tx
1.77
0.86
2. TV Control
Sh
-0.73
-0.22
Tr
0.65
0.97
Tx
-0.98
-0.84
3. Flashlight
Sh
-1.83
-1.16
Tr
-3.04
-3.55
Tx
-2.36
-3.31
4. Lamp
Sh
-0.16
-0.14
Tr
0.02
-1.28
Tx
-0.07
1.06
5. Refrigerator
Sh
-4.00
-3.60
Tr
-1.09
-1.12
Tx
-3.80
-3.64
6. Telephone
Sh
-2.60
-1.01
Tr
-0.28
-1.25
Tx
-0.56
-0.99


18
Tests
Mean pleasingness ratings for complex (C) and simple (S) patterns in
successive test. Consecutive tests were separated by eight presentations
of each pattern. (Source: Berlyne 1970)
FIGURE II 1
BERLYNES EXPERIMENTAL RESULTS


36
the product category. A second requirement that directed the selection of
products for the stimuli sets concerned the (non)existence of a product that was
atypical but unified. Product categories that contained instances of products
that were atypical (i. e., novel) and unified were considered problematic
because in such cases people may be more favorably disposed (i. e., receptive)
toward atypical (but unified) product designs due to prior experience with or
knowledge of the product. If this were the case it could lead to results that
overstated the influence of the unity design factor. This is especially important
since Experiment 1 effectively provides a theory test between a prototypicality
explanation of aesthetic response and a unity explanation. A third
requirement involved product conduciveness to the manipulation of unity. In
order for a product to exhibit unity (i. e., a visual connection among elements,
repetition of form or pattern) or disunity the product had to have parts that
could be perceived and manipulated separately. The fourth requirement that
guided the development of the stimulus sets concerned the medium used to
create the stimuli and stimulus communicability. A medium was needed that
would allow the creation and presentation of stimulus products that were
drastically different from the products that currently existed. This made it
difficult to construct stimuli by altering (either photographically or using
computer scanned images) existing products or pictures of existing products
because such modification often results in introducing "aesthetic confounds"


TABLE III 13
MEAN QUALITY RATINGS AND UNITY SCORES
69
Product
(++)
(+-)
(-+)
(-)
uT,
UN
1. Bathroom scale
Sh
5.28
5.03
5.25
5.39
0.16*b
0.92***
Tr
4.53
5.90
4.65
6.11
0.05
1.50***
Tx
4.35
5.81
5.02
5.65
-0.44
0.95***
2. TV Remote Control
Sh
5.24
5.26
5.34
5.37
-0.00
0.68**
Tr
4.03
5.18
5.48
6.66
-0.03
^ rjrj***
Tx
5.41
5.18
5.23
5.66
0.31+
0.77**
3. Flashlight
Sh
5.61
4.95**a
4.90*
5.02*
0.42*
0.70**
Tr
5.07
4.70+
4.69*
5.00
0.34**
0.69**
Tx
5.42
4.26**
4.16**
5.68
1 24***
1.83***
4. Lamp
Sh
5.59
5.43
5.03*
5.08+
0.08
0.53**
Tr
4.39
4.84
5.53
5.82
-0.09
1.26***
Tx
4.23
5.39
4.87
6.45
0.19
1 91***
5. Refrigerator
Sh
6.02
3.97***
4 45***
5.03**
1.33***
1 24***
Tr
5.24
4.73*
4.58*
5.66
0.79***
445***
Tx
6.36
4.74***
4.85***
5.25**
1.00***
0.85***
6. Telephone
Sh
5.55
4.71**
4.85**
4.77**
0.36*
0.37*
Tr
5.23
5.34
4.79*
4.56**
-0.17
0.06
Tx
5.55
5.33
5.13*
5.89
0.53**
1


10
disorderly/orderly, weak/powerful, somber/bright, etc. as they pertain to
aesthetic stimuli (e. g., paintings), the specific contributions of these
dimensions to aesthetic response are not yet well understood. Although
principles such as "unity" and "balance" may be understood in terms of visual
organization principles such as the Gestalt laws of perception (Lewalski 1988;
Pickford 1972, p. 31), no reported research has been found that relates Gestalt
laws to general rules regarding aesthetic responses (see Veryzer 1993 for an
exception).
Unity
Design Principles such as "unity" describe perceived spatial relations
between the parts of a visual display. The design principle of unity refers to
a congruity among the elements of a design such that they look as though they
belong together or as though there is some visual connection beyond mere
chance that has caused them to come together (Lauer 1979). The tendency to
perceive groupings of constitutive elements in certain ways or as integrated
entities is an important aspect of perception. These tendencies are described
by Gestalt laws (Koffka 1935) and design principles (Ching 1979; Lauer 1979).
Although design principles are more general than the Gestalt laws of
perception, the two sets of rules of perception are related and in some
instances a set of Gestalt laws may be used to describe a design principle. For
example, the Gestalt laws of proximity (i. e., elements that are closest to each


75
TABLE III 16
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
LINEAR UNITY SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subject Effects
Intercept
1
11819.3718
378.04
0.0001
Version
2
380.7288
6.09
0.0027
Order
1
13.2590
0.42
0.5157
Version Order
2
34.9076
0.56
0.5732
Error Sub (Version Order)
185
5784.0106
Within Subject Effects
Product
8
5156.6480
57.08
0.0001
Trans
2
537.0898
23.78
0.0001
Prod Version (Residual)
14
3044.2706
19.26
0.0001
Prod Order
8
325.4491
3.60
0.0004
Trans Order
2
32.1250
1.42
0.2500
Prod Version Order (Residual)
14
265.2929
1.67
0.1000
Error Sub (Prod)
1480
16712.9249
were discussed for aesthetic response. Individual t-tests for each of the
stimulus sets for all three of the derived responses were conducted to ensure
that the effect of unity was significant and in the proper direction across
stimulus sets (refer to Tables III-12, 13, and 14).
The three hypotheses concerning the influence of aesthetic response on
derived responses all seem to be supported. The products that received higher
beauty or aesthetic ratings (i. e., products exhibiting unity) also received higher
ratings for liking, quality, and price than did products that received lower
aesthetic ratings (i. e., products that did not exhibit unity). Although there
seems to be a reduction in the magnitude of the effect of the unity design
factor on the derived responses of perceived quality and expected price (as
compared with attitudes) the findings do seem to provide evidence that design


120
third experiment in which subjects rated products on only one of the dependent
measures would seem to suggest that the results are not merely an artifact of
the procedure employed.
The findings concerning the effect of evaluation context on the influence
of product aesthetics has some interesting managerial implications which are
especially relevant to the design and marketing of new products. Experiment
3 seems to indicate that the context in which product evaluations occur (e. g.,
presence of competitive products, descriptive product information) can
influence how heavily consumers weight aesthetic information (i. e., product
appearance) in their product evaluations. This suggests that there may be a
trade-off between the influence of descriptive information and product
appearance in the promotion or merchandising of a product. If, for example,
the addition of certain descriptive product information causes consumers to
focus less on a very well designed products appearance, then the net effect of
adding the information may be a reduction in the total "impact" of the
promotional effort. Depending on the goals (i. e., desired impact; e. g.,
increased sales, awareness, quality perceptions, etc.) of the promotion this
could seriously affect the results of the campaign.
Experiments 1 and 2 suggest that context (i. e., the opportunity to
compare a product version with other products in the same product category
during the evaluation phase of the buying process) plays a role in mediating
the influence of the aesthetic aspects of products. This suggests that managers


THE INFLUENCE OF UNITY AND PROTOTYPICALITY ON
AESTHETIC RESPONSES TO NEW PRODUCT DESIGNS
By
ROBERT W. VERYZER, JR.
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1993


30
The transformation of the two product features was done in a way that
also manipulated each product variations consistency with the unity visual
organization principle. In the case of the prototypical variation the two
selected features displayed a visual connection with each other (e. g., a
repetition of the same shape). When one of the two selected features was
altered it was done in a way that decreased unity (e. g., did not display a
visual connection with other parts of the product) and therefore decreased the
unity exhibited by the product variation. This was the case for each of the
variations in which only one feature had been altered (i. e., the moderately
atypical variations). In the case of the product variation where both features
were altered unity was again achieved since the altered features displayed a
visual connection to each other even though the features were very different
from the features of the prototypical variation. Thus, changes in the two
selected product features resulted in two levels of unity (unified and un
unified), and three levels of prototypicality (prototypical, moderately atypical,
extremely atypical).
An example of how unity and prototypicality were manipulated by
transforming two of a products features is shown in Figure III-l. In this
example two features of a telephone (handset and base) are altered using the
shape transformation in order to make the product either more or less
prototypical and more or less unified. The upper left-hand cell of the figure
contains the most prototypical (++) form of the product. The prototypicality of


87
Results and Discussion
The manipulation check for the effects of a change in either Feature A
or Feature B on aesthetic responses indicated that, as expected, there was an
overall reduction in the ratings of product appearance as compared with the
prototypical product version. The average effect of a change in either of these
features was weaker than they were in Experiment 1 with the average effect
of a change in Feature A being -.425 and Feature B -.281. The effect of a
change in Feature A or Feature B from the category prototype for all eight
stimulus sets are shown in Table IV-1. In four of the eight cases the change
TABLE IV 1
MANIPULATION CHECK
Product
Feature A
Feature B
1. T.V. Remote Tr
- 0.001
0.343
2. Flashlight Tr
0.096
- 0.059
3. Lamp Tx
- 0.862
- 0.414
4. Refrigerator Sh
- 1.352
- 0.827
5. Telephone Sh
- 0.687
- 0.327
6. Hair dryer Tx
- 0.488
- 0.798
7. Dresser Tr
- 0.171
0.019
8. Clock Sh
0.064
- 0.182
Mean effect of change across
products
-0.425
- 0.281


29
manipulations were accomplished. The manipulations will then be related to
the basic design employed in the experiment. This will be followed by a
general discussion of how the stimulus products employed in this experiment
were selected and developed.
Stimulus Manipulations
The stimulus sets (i. e., design sets) were constructed by first selecting
a prototypical product (i. e., product form) from a product category. Two
prominent parts of the prototypical product were then selected for
manipulation. Three variations of the prototypical product were then produced
by altering either one or the other of the two selected features or both features.
This produced a stimulus set consisting of four product variations (i. e., the
original prototypical product variation, two variations that contained one
altered feature, and one variation in which both features had been altered).
The two product variations that shared one of the selected features with the
prototypical variation (i. e., only one feature had been altered) were moderately
atypical products. The product variation that did not share either of the two
selected features with the prototype (i. e., both features had been altered) was
the most atypical product of the set. Thus, within the stimulus set of four
products three levels of prototypicality were represented (i. e., prototypical,
moderately atypical, and extremely atypical).


CHAPTER IV
EXPERIMENT 2
Overview
Experiment 1 investigated the effects of unity and prototypicality on
aesthetic and derived responses by presenting subjects with all four versions
of a stimulus set simultaneously and asking them to rate the versions on
semantic differential scales that measured the constructs of interest. This
experiment is a between-subjects replication of Experiment 1 in which subjects
evaluate only one design from each stimulus set. In this study, product
versions were evaluated without the context of other product versions from the
same product category. The between-subjects design of this experiment also
serves to reduce the possible demand artifacts that maybe involved in a within-
subjects experimental design that tests the influence of unity and
prototypicality on aesthetic responses, attitudes, quality perceptions, and price
expectations.
Experimental Design
A sub-set of the stimuli that were employed in Experiment 1 were used
in this experiment. One transformation was chosen from each product
category. The chosen transformation was the one with the strongest linear
84


Response
103
B
Beauty

Familiarity

Liking
O
Quality
*
Price
FIGURE IV-1
MEAN AESTHETIC AND DERIVED RESPONSES


25
consumer inferences and thus play a major role in directing consumer behavior
(Wilkie 1990, p. 267).
Evidence for the influence of aesthetic response on quality and
ability/performance evaluations can also be found in the social perception
literature. Landy and Sigall (1974) found significant main effects for writer
attractiveness on evaluations of a writer and her work. Similarly, Lennon
(1990) found a significant effect of clothing attractiveness (clothing and
accessories that matched vs. clothing and accessories that did not match in
color, style or pattern) on perceived competence. The findings of these studies
suggest that the aesthetic aspect of products (i. e., objects) and the aesthetic
responses that they give rise to may exert an influence on non-aesthetic
aspects of products such as quality.
H8: Quality ratings are higher for products that receive more
positive aesthetic ratings than for products that receive less
positive aesthetic ratings.
Price
The possibility that aesthetic response may influence perceived product
quality raises some interesting questions with regard to the price-quality
relationship. This relationship has been examined primarily in terms of price
as cue to quality (Monroe 1973; Zeithaml 1988). Even though a positive price-
perceived quality relationship does appear to exist, results of studies that have


74
TABLE III 15
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
46.0270
1.66
0.1937
Order
1
27.3070
1.96
0.1627
Version Order
2
124.4730
4.48
0.0126
Error Sub (Version Order)
185
2571.1570
Within Subiect Effects
Product
8
1223.6860
40.79
0.0001
Trans
2
166.7490
22.23
0.0001
Prod Version (Residual)
14
547.9980
10.44
0.0001
Prod Order
8
141.8833
4.73
0.0001
Trans Order
2
7.1283
0.95
0.2500
Prod Version Order
14
42.2361
0.80
0.2500
(Residual)
1480
5550.5546
Error Sub (Product)
Feature A
1
350.7149
48.87
0.0001
A Version
2
439.3120
30.61
0.001
A Order
1
21.8458
3.04
0.0827
A Version Order
2
14.4999
1.01
0.3662
Error Sub (A)
185
1327.7566
Feature B
1
136.4246
24.32
0.0001
B Version
2
154.4816
13.77
0.0001
B Order
1
36.0534
6.43
0.0121
B Version Order
2
25.8407
2.30
0.1028
Error Sub (B)
185
1037.9498
A B
1
2954.8430
378.04
0.0001
A B Version
2
95.1822
6.09
0.0027
A B Order
1
3.3147
0.42
0.5157
A B Version Order
2
8.7269
0.56
0.5732
Error Sub (A B)
185
1446.0027
Prod A B
8
1289.1620
57.08
0.0001
Prod A B Version
16
895.3401
19.82
0.0001
Prod A B Order
8
81.3623
3.60
0.0004
Prod A B Version *
16
74.3545
1.65
0.0509
Order
1480
4178.2312
Error Sub (Prod A B)


CHAPTER III
EXPERIMENT 1
Overview
The hypotheses developed in Chapter II concern the role of design
factors in influencing aesthetic and derived responses. It was hypothesized
that aesthetic and derived responses would be more positive for products
exhibiting high unity than they would be for products that were not consistent
with the unity visual organization principle. It was also suggested that
although there is some question about the level of prototypicality that
maximizes positive aesthetic response, prototypicality, in general, would be
expected to lead to more favorable aesthetic and derived responses. This
chapter first discusses the stimuli and methodology that were used in
Experiment 1 to examine these hypotheses. The analyses and results of
Experiment 1 are then presented and discussed.
Stimuli
In order to examine the hypotheses concerning the influence of unity and
prototypicality stimulus sets were created that allowed these factors to be
manipulated independently. This discussion will first describe how these
28


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Dressers (chest of drawers)
Dislike
? 4
5
6
7
8
9
Like
Dislike
2
3
4
5

7
8
9
Like
Ugly
3 4
5
6
7
s
9
Beautiful
Ugly
n
3
4
5
6
7
8
9
Beautiful
Familiar
-
3 4
5
6
7
8
9
Unfamiliar
Familiar 1
-
3
4
5
6
7
8
9
Unfamiliar
High Quail rv
3 4
5
6
7
8
9
Low Quality
High Quality
3
4
5
6
7
8
9
Low Quality
Low Price
:
3 -i
<
6
7
8
9
High Price
Low Price
n
3
4
5
6
7
8
9
High Price
Dislike
9 5
6
7 8
9
Like
Dislike
2
3
4
5
7
J
9
Like
Ugly
3 9 5
6
7 8
9
Beautiful
Ugly
2
3
4
5
6
7
8
9
Beautiful
Familiar
-
3 9 5

7 8
9
Unfamiliar
Familiar
3
4
5
6
7
S
9
Unfamiliar
High Quality
-
3 4 5
6
7
8
9
Low Quality
High Quality
2
3
4
5
6
7
8
9
Low Quality
Low Price
3 4 5
6
7
8
9
High Price
Low Pnce
-i
3
4
5
6
7
8
9
High Price
151


113
TABLE V 3
QUALITY
Product
Unity/
typical
Non-Unity/
Atypical
Unity/
Atypical
Un
1. Bathroom scale
5.24
5.58
6.68
1.10
2. T. V. Remote
6.44
6.00
6.21
0.21
3. Flashlight
5.30
5.46
5.98
0.52
4. Lamp
6.91
5.19
5.83*a
0.64
5. Refrigerator
5.83
5 70***a
5.97
0.27**b
6. Telephone
6.05
5.44*
5.65***
0.21
7. Hair dryer
6.11
6.77*
5.44**
-1.33
8. Dresser
6.29
5.67
6.09
0.42
9. Clock
5.49
6.37
6.00
-0.37
Total
5.96
5.79
5.98
0.19
a Indicates significance from Prototypical product.
b Indicates significance of the differences between Unity/Atypical
and Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Refrigerators
Dislike
:
3 4
;
6 "
8
9
Like
Dislike
2
3
5

S
0
Like
Ugly
"i
3 4
5
6 ?
s
Q
Beautiful
Ugly
2
3
4
5
6 "
8
9
Beautiful
Familiar
*
3 4
5
6 "
8
9
Unfamiliar
Familiar 1
2
3
4
5
b 7
8
Q
Unfamiliar
High Quality
? -
5
6 7
8
9
Low Quality
High Quality
2
3
4
5
b 7
8
9
Low Quality
Low Price
;
3 4
7
S
9
High Price
Low Pnce
-
3
4
5
o 7
8
9
High Price
Dislike
2 3 4 3
6 "
s
9
Like
Dislike
p
4 5
D
7 8
Like
Ugly
2 3 4 5
6 7
8
9
Beautiful
Ugly
2 p
4 5
9
7 8
Beautiful
Familiar
2 3 4 >

8
9
Unfamiliar
Familiar 1
2 3
4 5
9
j
9
Unfamiliar
High Quality
2 3 4 5

S
9
Low Quality
High Quality
p
4 5

7 8
9
Low Qualirv
Low Pnce
2 3 4 5
6 7
8
9
High Price
Low Pnce
2 3
4 5
7 8
0
High Price
142


38
studies. A list of these products is presented in Appendix A. These product
studies led to the identification of three basic ways to transform the features
of the products so as to simultaneously accomplish the prototypicality and
unity manipulations. These transformations involved altering product features
by changing the shape of parts of the product, adding texture to parts of the
product, or adding trim to part of the product. Additional product studies were
then undertaken to determine which products could be manipulated using each
of the three transformations (shape, trim, and texture). These studies involved
producing four versions of each product (for each of the three types of
transformations) by systematically transforming two product features in order
to accomplish the unity manipulation (i. e., a visual connection among product
features) at a particular level of prototypicality (i. e., shared features with the
category prototype).
In order to increase the generalizability of the results and demonstrate
the robustness of the effects of interest, nine product categories were selected
for use as stimuli. These nine product categories were: alarm clocks, bathroom
scales, dressers, flashlights, hair dryers, lamps, refrigerators, telephones, and
television remote controls. This resulted in a total of twenty-seven replications
(nine product categories x three types of transformation). Each of these
products had to be produced in four versions (Unity/Prototypical ++, Non
unity/Atypical +-, Non-unity/Atypical -+, and Unity/Atypical ). Thus, the


Response
83

Beauty

Familiarity

Liking
o
Quality
*
Price
Unity/Prototypicality
FIGURE III-4
MEAN AESTHETIC AND DERIVED RESPONSES


12
a framed picture (each photograph contained all five types of living-room
furniture but differed with respect to the mix of traditional and contemporary
styles). Subjects rated the furniture in the randomly assigned photo in terms
of perceived unity (i. e., a visual connection among elements that suggests that
something beyond mere chance has caused them to come together), aesthetic
response, social impression, general liking, and intention to own. Subjects also
provided ratings on a number of items which measured personality and
motivation variables. Bell et al. (1991), found that aesthetic response did, in
o
fact, depend on perceived unity (R = .05, p < .001) and that perceived unity
depended on the product styles (R = .19, p < .001 with significant
contributions of chair, table, and ensemble). These findings suggest that the
principle of unity, which was operationalized in this study as the mix of the
two styles shown in each photograph, does seem to influence aesthetic
response.
The influence of unity has also been examined in the context of social
information processing studies. For example, Lennon (1990) investigated the
effects of clothing attractiveness on perceptions. In order to determine the
degree to which people perceive others differentially as a function of the
attractiveness of their clothing, slides of six different models in business attire,
three wearing attractive clothing and three wearing unattractive clothing, were
prepared. In the unattractive clothing condition, models wore garments and
accessories that did not match either in color, style, or pattern. In the


53
TABLE III 4
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
LINEAR SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
7767.5254
338.16
0.0001
Version
2
269.8589
5.87
0.0034
Order
1
12.5163
0.54
0.4614
Version Order
2
22.4679
0.49
0.6140
Error Sub (Version Order)
184
4226.5090
Within Subiect Effects
Product
8
3735.7748
53.26
0.0001
Trans
2
549.8955
31.36
0.0001
Prod Version (Residual)
14
2599.8573
21.18
0.0001
Prod Order
8
376.3133
5.37
0.0001
Trans Order
2
40.2259
2.29
0.2500
Prod Version Order (Residual)
14
103.6562
0.84
0.2500
Error Sub (Prod)
1472
12905.3777
III-4 and Table III-5 clearly indicate support for the unity hypothesis. The
intercept using the linear unity score (Ut) as the dependent measure was
significantly different from zero F (1,184) = 338.16, p < 0.0001. The
intercept as tested using the nonlinear unity score (U-^) approach was also
significantly different from zero F(l,190) = 473.48, p < 0.001.^ The fact that
there is a small, but significant t (190 = 6.04, one-tailed p < .001) difference
between the means of the Unity/Prototypical (++) products (M = 5.34) and the
/2
The results for the linear unity scores are identical to the results for the
ratings score approach.
7
The difference in the denominator degrees of freedom are due to missing
data that caused several subjects to be eliminated from the "unity score"
analysis.


65
TABLE III 11
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
NONLINEAR UNITY SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
9450.3839
254.47
0.0001
Version
2
276.1947
3.72
0.0261
Order
1
85.6300
2.31
0.1306
Version Order
2
88.3924
1.19
0.3066
Error Sub (Version Order)
181
6721.8569
Within Subiect Effects
Product
8
2197.4792
28.33
0.0001
Trans
2
395.5843
20.40
0.0001
Prod Version
14
1341.7119
9.88
0.0001
Prod Order
8
438.9539
5.66
0.0001
Trans Order
2
9.8527
0.51
0.2500
Prod Version Order (Residual)
14
198.696
1.46
0.2500
Error Sub (Prod)
1448
14040.3723
Derived Responses
The remaining three hypotheses concerned the effect of aesthetic
response on derived responses (i. e., attitude, quality, and price). It was
hypothesized that aesthetic responses would influence derived responses and
thus products that received more positive aesthetic ratings (i. e., products that
exhibit unity) would also be rated higher with regard to liking (i. e., attitude),
quality (i. e., perceived quality), and price (i. e., expected price) than would
products that received less positive aesthetic ratings. The critical tests for
examining these hypotheses are the same as the critical test for examining the
influence of unity on aesthetic response except the dependent measures reflect


13
attractive clothing condition, models wore clothing that was well matched and
wore accessories to complement their clothing (two pilot studies were
conducted to get a consensus regarding clothing attractiveness). Fifty-eight
female subjects listened to a pre-recorded audio tape consisting of thirty
suggestions relative to marketing a perfume. As a comment was heard, a slide
of the woman purported to have made the comment was projected. Subjects
rated the women on competence, work comfort (a measure of the extent to
which the respondent would feel comfortable working with the woman shown),
and sociability. A repeated measures analysis of variance revealed that there
was a main effect for clothing attractiveness on perceived competence (F(l,58)
= 52.14, p = .000). Models dressed in attractive clothing (M = 55.15) were
perceived to be more competent than models dressed in unattractive clothing
(M = 46.49). There was also a main effect for clothing attractiveness on
perceived work comfort (F(l,58) = 24.18, p = .000). Respondents indicated that
they would feel more comfortable working with attractively dressed models (M
= 22.02) than with unattractively dressed models (M = 19.10). Finally, there
was a main effect for clothing attractiveness on perceived sociability (F(l,58)
= 5.28, p = .025). Attractively dressed models (M = 11.41) were perceived to
be more sociable than those models who were unattractively dressed (mean =
10.86). These results provide further support for the view that configurations
of aesthetic elements (i. e., visual organization rules such as unity, color


TABLE III 2
MEAN BEAUTY RATINGS AND UNITY SCORES
50
Product
(++)
(+-)
(-+)
(-)
uT,
Un
1. Bathroom Scale
Sh
5.30
5.16
5.13
5.18
0.85
0.88***
Tr
4.42
5.42
4.56
5.81
0.12*b
1.38***
Tx
4.03
5.80
4.89
4.81
-0.89
0.43+
2. TV Control
Sh
4.82
4.09*a
4.60
4.95
0.51**
1 41***
Tr
3.98
4.63
4.95
6.42
0.44*
2.23***
Tx
5.36
4.38**
4.52**
5.28
0.84**
1.36***
3. Flashlight
Sh
6.31
3.95***
3.00***
3.05***
1.20***
0.55**
Tr
5.57
3.74***
4r 41***
3 97***
0 70***
0.61*
Tx
5.89
2.85***
2.34***
5.15*
2 94***
3.00***
4. Lamp
Sh
5.64
5.66
4.36***
4.36**
0.00
0.39*
Tr
4.65
4.58
5.71
5.92
0.14
1 42***
Tx
5.00
4.84
4.86
6.05
0.68**
2.08***
5. Refrigerator
Sh
6.00
2.00***
2.40***
3.95***
2.80***
2.20***
Tr
5.31
4.22**
4.19**
5.70
1.28***
2.03***
Tx
6.03
2.23***
2.39***
3.34***
2.34***
1.32***
6. Telephone
Sh
5.30
2 70***
4.29**
3.48***
0.90***
0.94**
Tr
5.44
5.16+
4.19***
4.18***
0.15
0.21+
Tx
5.74
5.18**
4 75***
6.02
0.90***
1 41***


schema). Furthermore, it is suggested that aesthetic responses influence
consumers attitudes toward products, perceptions of product quality, and price
expectations.
Three experiments are conducted in order to examine the role of unity
and prototypicality in influencing aesthetic responses and product perceptions.
The first study examines the relationship between the two factors (i. e., unity
and prototypicality) and consumers responses to product designs. This study
provides evidence that there is a positive effect of unity on aesthetic response
and that the favorable aesthetic response generated by consistency with the
unity principle also influences "non-aesthetic" product perceptions. The study
also finds that the unity factor provides a better explanation of aesthetic
response than does prototypicality.
The second and third experiments examine whether or not the context
of the evaluation situation (i. e., presence of other similar products, written
product descriptions) moderates the influence of unity. These studies find that
even though the evaluation context can reduce the magnitude of the unity
effect, the unity effect is quite robust and is pervasive.
Taken together, the results of the three experiments indicate that the
unity visual organization principle is a very important design factor and that
aesthetic responses can have a significant impact on consumers responses to
product designs. The results highlight the relationship between product
aesthetics (i. e., design) and consumer behavior.
vii


79
TABLE III 21
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
48.662
1.55
0.2149
Order
1
28.494
1.82
0.1795
Version Order
2
45.063
1.44
0.2405
Error Sub (Version Order)
179
2808.080
Within Subiect Effects
Product
8
1089.9719
38.16
0.0001
Trans
2
58.7083
8.22
0.0010
Prod Version (Residual)
14
137.2604
2.75
0.0001
Prod Order
8
182.0655
6.37
0.0001
Trans Order
2
8.0139
1.12
0.2500
Prod Version Order (Residual)
14
76.0911
1.52
0.1000
Error Sub (Product)
1432
5112.3283
Feature A
1
250.9383
56.16
0.0001
A Version
2
83.3609
9.33
0.0001
A Order
1
1.9605
0.44
0.5086
A Version Order
2
4.8154
0.54
0.5844
Error Sub (A)
179
799.8424
Feature B
1
321.3547
86.02
0.0001
B Version
2
130.0014
17.40
0.0001
B Order
1
2.3247
0.62
0.4312
B Version Order
2
11.9936
1.61
0.2037
Error Sub (B)
179
668.6915
A B
1
193.1558
57.67
0.0001
A B Version
2
26.2598
3.92
0.0216
A B Order
1
0.0024
0.00
0.9785
A B Version Order
2
4.2203
0.63
0.5337
Error Sub (A B)
179
599.4864
Prod A B
8
200.3073
15.86
0.0001
Prod A B Version
16
129.5848
5.13
0.0001
Prod A B Order
8
9.5617
0.76
0.6409
Prod A B Version Order
16
27.3044
1.08
0.3680
Error Sub (Prod A B)
1432
2260.7715


46
subtracting from this the ratings for both of the Non-unity/Atypical conditions.
The resulting score provides a straight-forward measure of the main effect of
unity. The third way of forming the dependent measures involved a
modification of the interaction score. This variation entailed subtracting the
product rating of the lowest rated Non-unity/Atypical condition from the
product rating for the Unity/Atypical condition. This score allows a non-linear
model of prototypicality to be examined (see Figure III-3). If prototypicality
dominates unity then product versions that are more similar to the category
prototype should receive higher ratings than less typical product versions (i. e.,
the minimum rating for either of the Non-unity/Atypical conditions should
always be greater than the rating for the Unity/Atypical condition since the
former always has more shared features with the category prototype than does
the latter).
Both the unity linear score (U-jO and the unity nonlinear score (Ujj) are
needed in order to get a (more) complete sense of the effects of unity and
prototypicality. While the score does provide an indication of the main
effect of unity, it does not examine a (decreasing) non-linear model of
prototypicality. The score does examine a non-linear model of
prototypicality; however, the U-^ score reflects some random variation in the
subjects ratings (i. e., the lowest rated Non-unity/Atypical condition is always
used to construct the score). Parallel analyses were conducted on each of
the three variations of dependent measures (i. e., ratings scores, linear unity


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Flashlights
Dislike
2 3
- 5
6 7
8
9
Like
Dislike
2
3
4
5
6
7
8
9
Like
Ugly
*> ^
- 5
6 7
8
9
Beautiful
Ugly
2
3
4
5
6
7
8
9
Beautiful
Familiar
2 3
4 5
6 7
8
9
Unfamiliar
Familiar ]
2
3
4
5
6
7
8
9
Unfamiliar
High Quality
- 5
6 7
8
9
Low Quality
High Quality
p
3
4
5
6
7
8
9
Low Quality
Low Price
2 3
4 -
6 7
S
9
High Price
Low Price
2
3
4
5
6
7
8
9
High Pnce
Dislike
2 '
- ;
6
7
8
9
Like
Dislike
2
3
4
5
&
7 8
Q
Like
Ugly
2 3
4 5
6
7
8
9
Beautiful
Ugly
2
3
4
5
6
7 8
9
Beautiful
Familiar ]
2 3
-
5
7
8
9
Unfamiliar
Familiar ]
n
3
4
5
6
" 8
9
Unfamiliar
High Quality
2 3
- 5
6
7
8
9
Low Quality
High Quality
2
3
4
5
6
7 8
9
Low Quality
Low Pnce
2 3
4 5
6
7
8
9
High Price
Low Price
2
3
4
5
6
7 8
9
High Price


47
Linear Model
Aesthetic
(or Dervied)
Response
C + + ) ( + ) (- + ) (--)
TJnity/Prototypicality
Nonlinear Model
Aesthetic
(or Derived)
Response
( + + )
( + ) ( + )
U nity/Pr ototypicality
FIGURE III-3
LINEAR AND NONLINEAR "UNITY" SCORES


45
"unity" hypothesis, however, suggests that if the changes in the features are
made in a way that increases unity then aesthetic responses should be higher
in this case (as opposed to the non-unity cases) since the product version
exhibits consistency with the visual organization principle of unity. This is in
effect a theory test (for a linear model) which corresponds to a test of the
Feature A x Feature B interaction. This test can be extended to the
investigation of the influence of unity and prototypicality on derived responses.
Analyses
The data were analyzed using mixed ANOVA designs with Version and
Order being treated as between-subjects factors and Feature A, Feature B, and
Products being treated as within-subjects factors. Transformation was also
treated as a within-subjects factor because each subject saw products that had
been altered using the three different transformation types.
Subjects product ratings on the five 9-point semantic differential scales
were treated in three different ways in order to compute the dependent
measures that were analyzed. The first approach was to perform the analysis
directly on subjects ratings of the products. A second approach involved
computing the linear contrast that reflected the interaction of Feature A and
Feature B (i. e., the main effect of unity) for each design set. Under this
approach difference or "interaction" scores were formed by adding subjects
ratings for the Unity/Prototypical and Unity/Atypical conditions and


26
examined the price-quality relationship have been somewhat mixed and the
findings imply that price may not be the dominant cue in quality perception
(Monroe 1977; Zeithaml 1988). Moreover, there seem to be cases where
perceptions of (high) quality are formed without being diminished by (low)
price and the quality perceptions subsequently influence price (e. g., the
perceived quality of Japanese automobiles despite their initial low prices).
In cases where consumers initially do not have price information it
seems likely that they might form impressions about a product based on non
price information (e. g., physical composition of the product, packaging, brand
name, etc.) and that these impressions could influence price expectations.
Thus, it is conceivable that the same design factors that influence aesthetic
responses and thereby perceptions of product quality may also influence price
expectations.
H9: Price expectations (ratings) are higher for products that
receive more positive aesthetic ratings than for products that
receive less positive aesthetic ratings.
Ultimately, the price expectations fostered by product design/aesthetics
may play a role determining consumers "price thresholds" (Monroe 1973) for
a particular product within a category.
If indeed unity and prototypicality systematically influence aesthetic
responses and these, in turn, influence product preferences and product


11
other tend to form groups), similarity (i. e., elements that are similar tend to
form groups), and common destiny (i. e., parts of a figure that have a common
destiny tend to form units) may be viewed as ways to achieve unity (Katz 1950;
Lauer 1979). According to the Gestalt psychologists beauty is dependent on
the degree to which an object displays relations consistent with the Gestalt
laws of organization. Koffka (1935) clearly suggested this when he discussed
how violations of such laws as "good continuation" and "good shape" are not
only felt as violations, they conflict with our feeling of "fit" and "hurt our sense
of beauty" (Koffka 1935, p. 175).
Although the visual organization principle of unity would seem to be a
likely factor in influencing aesthetic responses, there is little in the way of
research that relates visual organization principles (e.g., design principles,
Gestalt laws) such as unity to aesthetic responses. There is, however, a
limited amount of research that indicates that there are general guidelines or
principles for combining visual elements in order to maximize aesthetic
responses. A study by Bell, Holbrook, and Solomon (1991), which examined
the impact of gestalt-like ensemble effects and the influence of personality
factors on product evaluations, provides support for the view that unity may
systematically influence aesthetic responses. In their study, subjects were
asked to look at one of thirty-two color photographs containing various
combinations of traditional and contemporary styles of five types of living-room
furniturespecifically, a chair, a table, a piece of sculpture, a floor lamp, and


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
...Refrigerators.
Dislike
? -1
5 6
7
8
9
Like
Dislike
:
3
4
5
6
7
8
9
Like
Ugly
3 4
5 6
7
8
9
Beautiful
Ugly
2
3
4
5
6
7
8
9
Beautiful
Familiar
3 4
5 6
7
8
9
Unfamiliar
Familiar 1
2
3
4
5
6
7
8
9
Unfamiliar
High Quality
*>
3 4
5 6
7
8
9
Low Quality
High Qualirv
3
4
5
6
7
8
9
Low Quality
Low Price
2
3 4
5 6
7
8
9
High Pnce
Low Price
o
3
4
5
6
7
8
9
High Pnce
Dislike
2
3 -
5
6
7
8
9
Like
Dislike
"i
3 4
5
6
-
8
9
Like
Ugly
2
3 9
5
6
7
8
9
Beautiful
Ugly
3 4
5
6
7
9
Beautiful
Familiar 1
i
3 4
5
A
7
8
9
Unfamiliar
Familiar ]
o
3 4
S
6
7
8
9
Unfamiliar
High Quality
*>
3 4
5
6
7
8
9
Low Quality
High Quality
->
3 4

6
7
8
9
Low Quality
Low Price
i
3 4
5
6
7
8
9
High Price
Low Pnce
2
3 4
5
6
7
8
9
High Price
143


THE INFLUENCE OF UNITY AND PROTOTYPICALITY ON
AESTHETIC RESPONSES TO NEW PRODUCT DESIGNS
By
ROBERT W. VERYZER, JR.
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1993

Copyright 1993
by
Robert W. Veryzer, Jr.

ACKNOWLEDGEMENTS
A number of people deserve credit for their role in helping me to
accomplish the goal of finishing this dissertation. First and foremost is
Dr. J. Wesley Hutchinson, the chairman of my committee, to whom I am
extremely grateful. The time and effort that he has contributed to this
endeavor are greatly appreciated. This work would not have been possible
without his benevolent guidance.
I am very grateful to Dr. Richard Lutz, Dr. Chris Janiszewski, and
Dr. David Mick for their- guidance in refining my ideas and for the
encouragement that they have given me. I would also like to thank
Dr. John Lynch, Jr. for his guidance on the statistical analyses and
Dr. Jonathan Hamilton for contributing an economists point of view.
I am indebted to my parents, Robert and Marion Veryzer, for their love
and support, and for always being there when I need them.
Finally, I would like to thank my brother David who has been a
tremendous source of inspiration to me.
in

TABLE OF CONTENTS
ACKNOWLEDGEMENTS iii
ABSTRACT vi
CHAPTERS
I INTRODUCTION 1
II CONCEPTUAL BACKGROUND ON THE
FACTORS THAT INFLUENCE AESTHETIC
RESPONSES 6
Introduction 6
Factors Affecting Aesthetic Responses 9
Visual Organization Principles 9
Prototypicality 14
Derived Responses 21
III EXPERIMENT 1 28
Overview 28
Stimuli 28
Experimental Design 39
Experimental Procedure 40
Results and Discussion 41
Summary 81
IV EXPERIMENT 2 84
Overview 84
Experimental Design 84
Experimental Procedure 86
Results and Discussion 87
Summary 101
V EXPERIMENT 3 104
Overview
Stimulus Materials ,
Experimental Design
104
105
106

Experimental Procedure 108
Results and Discussion 108
Summary 114
VI GENERAL DISCUSSION 116
APPENDICES
A LIST OF PRODUCTS EXPLORED FOR USE AS
STIMULI 127
B STIMULUS SETS 129
C DIAGRAM OF EXPERIMENT 1 157
D DIAGRAM OF EXPERIMENT 2 159
E STRONG AND WEAK PRODUCT DESCRIPTIONS 161
REFERENCE LIST 164
BIOGRAPHICAL SKETCH 171
v

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
THE INFLUENCE OF UNITY AND PROTOTYPICALITY ON
AESTHETIC RESPONSES TO NEW PRODUCT DESIGNS
By
Robert W. Veryzer, Jr.
August 1993
Chairman: Dr. J. Wesley Hutchinson
Major Department: Marketing
This dissertation investigates the influence of the aesthetic aspects of
product appearance on consumers product preferences and product
evaluations. Two factors, unity and prototypicality, are identified and
discussed as being the principal factors that influence consumers aesthetic
responses to product designs. The research provides a basis for a theory of
consumer aesthetics and has implications for new product development,
product quality, and marketing strategy.
The dissertation postulates that consumers responses to product designs
are influenced by the designs consistency with the visual organization
principle of unity (i. e., a congruity among the elements of a design) and its
level of prototypicality (i. e., familiarity; shared features with the category
vi

schema). Furthermore, it is suggested that aesthetic responses influence
consumers attitudes toward products, perceptions of product quality, and price
expectations.
Three experiments are conducted in order to examine the role of unity
and prototypicality in influencing aesthetic responses and product perceptions.
The first study examines the relationship between the two factors (i. e., unity
and prototypicality) and consumers responses to product designs. This study
provides evidence that there is a positive effect of unity on aesthetic response
and that the favorable aesthetic response generated by consistency with the
unity principle also influences "non-aesthetic" product perceptions. The study
also finds that the unity factor provides a better explanation of aesthetic
response than does prototypicality.
The second and third experiments examine whether or not the context
of the evaluation situation (i. e., presence of other similar products, written
product descriptions) moderates the influence of unity. These studies find that
even though the evaluation context can reduce the magnitude of the unity
effect, the unity effect is quite robust and is pervasive.
Taken together, the results of the three experiments indicate that the
unity visual organization principle is a very important design factor and that
aesthetic responses can have a significant impact on consumers responses to
product designs. The results highlight the relationship between product
aesthetics (i. e., design) and consumer behavior.
vii

CHAPTER I
INTRODUCTION
There is a growing recognition that product design, and particularly
aesthetic aspects of product design, is emerging as a key marketing element
(Kotler and Rath 1984). Throughout the 1980s companies updated their
manufacturing methods, quality control, distribution networks, customer
service, and labor/management relations in order to compete with other foreign
and domestic manufacturers. Today, as more and more manufacturers are
able to achieve similar levels of price, quality, reliability, and technology,
product appearance is increasingly being acknowledged as the major difference
around which consumers can exercise a choice (Oakley 1990, p. 4).
Product design, or "competitive aesthetics" as it has been called (Reed
1990), is gaining recognition as a strategic activity that companies can use to
gain a sustainable competitive advantage (Kotler and Rath 1984; Whitney
1988). It is being used for every sort of product from Apple computers to La
Croix sparkling water cans. As Bruce Nussbaum pointed out in Business
Week (June 17, 1991, p. 62): "Recently, business has grown increasingly aware
that design sells. U. S. Companies, in particular, are rediscovering that good
design translates into quality products, greater market share, and heftier
profits." Although the technical aspects of products remain vitally important,
1

2
more and more manufacturers of all kinds of products are having to come to
terms with the reality that product appearance is often the major factor
influencing (consciously or unconsciously) what people buy and how much they
are willing to pay (Oakley 1990, p. 5).
The influence of aesthetic factors on product preferences and perceptions
is an important but often neglected area of study in consumer research. The
design of products inherently involves aesthetics. The aesthetic aspects of a
product give rise to the registering of affect or pleasure due to the conscious
or unconscious influences of the characteristics of the product (Holbrook and
Zirlin 1985). The response arising from the interaction between the aesthetic
aspects of an object and the perceiver of the object has been termed the
"aesthetic response" (Olson 1981). There is a growing recognition among
marketing researchers that aesthetic responses can significantly affect
consumer behavior and thus aesthetics is gaining recognition as an important
marketing variable (Kotler and Rath 1984; Wallendorf 1980). The influence
of aesthetics is increasingly being acknowledged as an important part of new
product development (Whitney 1988), marketing strategy (Kotler and Rath
1984), product quality (Garvin 1984; Zeithaml 1988), product differentiation
(Dickson and Ginter 1987), and competitive advantage (Holt 1985; Kotler and
Rath 1984).
Despite the growing awareness of the significant influence that the
aesthetic aspects of products can have on product preferences and perceptions,

3
surprisingly little in the way of design/aesthetic theory has been offered that
aids in our understanding of how aesthetic responses are formed (Berlyne
1974, p. 5; Gorski 1987, p. 68; Holbrook and Zirlin 1985; Pye 1978, p. 88).
Even though disciplines concerned with basic research such as experimental
aesthetics remain dormant or in a "protracted infancy" (Berlyne 1971), other
disciplines such as industrial (product) design look to them to address the lack
of design theory (Gorski 1987, p. 68). The end result is that managers
continue to exhibit much "unease" when it comes to making decisions about
design and managing design projects (Oakley 1990, p. 7). This is particularly
problematic for managers charged with transferring new technology out of the
laboratory and into the market. It is at this point that the appearance of a
product plays a crucial role in communicating the products identity and use to
consumers.
Although consumer research seems to be ideally suited for the study of
aesthetic response due to its unique combination of scientific research methods
and its tangible research context (i. e., consumer product focus), consumer
researchers have taken only tentative steps toward exploring aesthetics and
its relationship to consumers behavior. Much of this work has been concerned
with debating the definition and scope of consumer aesthetics rather than
examining the nature and influence of aesthetics and aesthetic responses
(Holbrook and Zirlin 1985). Thus, aesthetic research has proceeded without
the benefit of a conceptual foundation (Holbrook and Zirlin 1985; Olson 1981).

4
Consumer researchers have not developed one, and they have not found one
in any of the other disciplines to adopt as a starting point for building
aesthetic theory.
This dissertation attempts to address this void in aesthetic theory by
examining the influence of the aesthetic aspects of product appearance on
consumers product preferences and perceptions. This research represents
important steps toward understanding the implications of aesthetics and
product design for consumer research and the formulation of a theory of
consumer aesthetics.
In Chapter II the two principal factors, unity and prototypicality, that
influence consumers aesthetic responses to a products design are discussed
and the supporting research is reviewed. Several hypotheses are developed
that involve the relationship of the factors to aesthetic and derived responses.
Chapters III and IV each present the method and results of an
experiment designed to investigate the influence of the unity and
prototypicality factors on aesthetic and derived responses. Chapter III also
includes a detailed discussion of how the stimuli employed in all of the
experiments were developed.
Chapter V presents the method and results of an experiment that
focuses on derived responses.
Chapter VI provides a general discussion of the findings and
implications of these results for research that examines the influence of

5
product aesthetics on consumer behavior. The managerial implications of
these findings are also discussed.

CHAPTER II
CONCEPTUAL BACKGROUND ON THE FACTORS THAT INFLUENCE
AESTHETIC RESPONSES
Introduction
Even though there is a growing awareness of the important role that
aesthetics play in influencing product preferences, our understanding of
aesthetic responses is extremely limited. Our primitive understanding of
aesthetic response phenomena may be due to the highly fragmented approach
that has characterized its study. Aesthetics have been studied in a number of
fields including philosophy, art history, psychology, experimental aesthetics,
industrial design, and more recently consumer behavior. Each field has
contributed to our comprehension of aesthetics and yet, little progress has been
made in understanding the specifics of aesthetic response (Berkowitz 1987;
Wohlwill 1981). Within the many conceptualizations of aesthetic response that
have emerged from these different fields of study, there is little agreement and
limited insight regarding why a particular object is perceived as pleasurable
or beautiful while another is viewed as unattractive. The question concerning
what makes an object aesthetically pleasing, regardless of whether it is a
painting or a portable stereo, has received diverse and typically vague answers
from the disciplines that have studied it. While a lack of convergence is not
6

7
uncommon in interdisciplinary research, the different orientations of the
disciplines that are or have been concerned with aesthetic phenomena and the
vagueness of the theories that have been offered to explain aesthetic responses
have severely inhibited progress toward gaining an understanding of aesthetic
responses.
Some aesthetic response theories, which seem to view aesthetic response
as idiosyncratic, maintain that their are no laws or principles of aesthetics
(Mothersill 1989), while others suggest that inner tendencies of the visual
system result in laws that govern perception and thereby influence aesthetic
response (Katz 1950; Koffka 1935). There are aesthetic theories that point to
fashion trends or the influence of culture as the determinant of systematic
aesthetic responses for all classes of products (Pleydell-Pearce 1970; Sproles
1981). There are views of aesthetic response that would suggest that the
consumers preference for a specific product was determined by a desire for
unity in variety (Auld 1981; Berlyne 1971; Lauer 1979). Other views maintain
that preference is related to prototypicality (Loken and Ward 1990; Nedungadi
and Hutchinson 1985). There are also those that periodically suggest that the
term "aesthetic response" applies only to works of art (Holbrook 1981), only to
be contradicted by those who maintain that all objects have an aesthetic
component (Berlyne 1974; Wallendorf 1980). Even the artists and product
designers who determine the forms of the objects seem to have trouble
agreeing on the basics of design (Lauer 1979) and as yet "have not formulated

8
what they know" (Pye 1978, p. 11). In fact, the well worn maxim that "form
follows function" has even been challenged (Lewalski 1988; Pye 1978).
Interestingly enough, designers are beginning to look to the disciplines of
experimental psychology and consumer behavior in order to understand the
cognitive condition of design (Zaff 1987).
The lack of understanding concerning what makes an object
aesthetically pleasing is particularly evident in the area of new product
development. The issue of aesthetic response and the factors that influence it
are frequently overlooked in many discussions of the new product development
process. Much of the work in this area seems to sidestep these considerations
by either ignoring the role of industrial design (i. e., the process of shaping or
giving form to goods that are to be mass produced) in the new product
development process or subsuming the industrial design function under the
engineering function (e. g., Gruenwald 1985; Urban and Hauser 1993). In
either case, the result is that the role of product appearance in the success of
a new product is not explicitly acknowledged or addressed. Given the
increasing importance of product aesthetics this omission is a rather serious
deficiency. Fortunately, the distinct role of industrial design and its
relationship to the engineering and marketing functions is beginning to receive
the attention that it merits (e. g., Lorenz 1986). However, most of the work
that addresses the role of industrial design in the new product development

9
process stops short of identifying and investigating the factors that influence
aesthetic responses to products.
Factors Affecting Aesthetic Response
Aesthetic response is a complex phenomenon that is not yet well
understood. As with any complex phenomenon it is likely to involve a number
of factors; however, prior research seems to suggest that there are two
principal factors: the visual organization principle of unity and prototypicality,
that may significantly influence aesthetic responses. Some of the significant
prior research that suggests that these factors may play an important role in
influencing aesthetic responses will now be reviewed. The influence of
aesthetics/aesthetic response on non-aesthetic perceptions (i. e., derived
responses) will also be discussed.
Visual Organization Principles
The greatest void in aesthetic theory and research (and the greatest
opportunity) concerns the identification of specific factors that systematically
influence aesthetic responses. Although very general "rules" or "principles"
have sometimes been offered (e. g., "unity in variety" Hutcheson 1725), the
rules are usually vague and unspecified. Even though aestheticians seem to
rely on principles such as "unity" and "balance" and aesthetics research has
frequently employed scales that attempt to measure dimensions such as

10
disorderly/orderly, weak/powerful, somber/bright, etc. as they pertain to
aesthetic stimuli (e. g., paintings), the specific contributions of these
dimensions to aesthetic response are not yet well understood. Although
principles such as "unity" and "balance" may be understood in terms of visual
organization principles such as the Gestalt laws of perception (Lewalski 1988;
Pickford 1972, p. 31), no reported research has been found that relates Gestalt
laws to general rules regarding aesthetic responses (see Veryzer 1993 for an
exception).
Unity
Design Principles such as "unity" describe perceived spatial relations
between the parts of a visual display. The design principle of unity refers to
a congruity among the elements of a design such that they look as though they
belong together or as though there is some visual connection beyond mere
chance that has caused them to come together (Lauer 1979). The tendency to
perceive groupings of constitutive elements in certain ways or as integrated
entities is an important aspect of perception. These tendencies are described
by Gestalt laws (Koffka 1935) and design principles (Ching 1979; Lauer 1979).
Although design principles are more general than the Gestalt laws of
perception, the two sets of rules of perception are related and in some
instances a set of Gestalt laws may be used to describe a design principle. For
example, the Gestalt laws of proximity (i. e., elements that are closest to each

11
other tend to form groups), similarity (i. e., elements that are similar tend to
form groups), and common destiny (i. e., parts of a figure that have a common
destiny tend to form units) may be viewed as ways to achieve unity (Katz 1950;
Lauer 1979). According to the Gestalt psychologists beauty is dependent on
the degree to which an object displays relations consistent with the Gestalt
laws of organization. Koffka (1935) clearly suggested this when he discussed
how violations of such laws as "good continuation" and "good shape" are not
only felt as violations, they conflict with our feeling of "fit" and "hurt our sense
of beauty" (Koffka 1935, p. 175).
Although the visual organization principle of unity would seem to be a
likely factor in influencing aesthetic responses, there is little in the way of
research that relates visual organization principles (e.g., design principles,
Gestalt laws) such as unity to aesthetic responses. There is, however, a
limited amount of research that indicates that there are general guidelines or
principles for combining visual elements in order to maximize aesthetic
responses. A study by Bell, Holbrook, and Solomon (1991), which examined
the impact of gestalt-like ensemble effects and the influence of personality
factors on product evaluations, provides support for the view that unity may
systematically influence aesthetic responses. In their study, subjects were
asked to look at one of thirty-two color photographs containing various
combinations of traditional and contemporary styles of five types of living-room
furniturespecifically, a chair, a table, a piece of sculpture, a floor lamp, and

12
a framed picture (each photograph contained all five types of living-room
furniture but differed with respect to the mix of traditional and contemporary
styles). Subjects rated the furniture in the randomly assigned photo in terms
of perceived unity (i. e., a visual connection among elements that suggests that
something beyond mere chance has caused them to come together), aesthetic
response, social impression, general liking, and intention to own. Subjects also
provided ratings on a number of items which measured personality and
motivation variables. Bell et al. (1991), found that aesthetic response did, in
o
fact, depend on perceived unity (R = .05, p < .001) and that perceived unity
depended on the product styles (R = .19, p < .001 with significant
contributions of chair, table, and ensemble). These findings suggest that the
principle of unity, which was operationalized in this study as the mix of the
two styles shown in each photograph, does seem to influence aesthetic
response.
The influence of unity has also been examined in the context of social
information processing studies. For example, Lennon (1990) investigated the
effects of clothing attractiveness on perceptions. In order to determine the
degree to which people perceive others differentially as a function of the
attractiveness of their clothing, slides of six different models in business attire,
three wearing attractive clothing and three wearing unattractive clothing, were
prepared. In the unattractive clothing condition, models wore garments and
accessories that did not match either in color, style, or pattern. In the

13
attractive clothing condition, models wore clothing that was well matched and
wore accessories to complement their clothing (two pilot studies were
conducted to get a consensus regarding clothing attractiveness). Fifty-eight
female subjects listened to a pre-recorded audio tape consisting of thirty
suggestions relative to marketing a perfume. As a comment was heard, a slide
of the woman purported to have made the comment was projected. Subjects
rated the women on competence, work comfort (a measure of the extent to
which the respondent would feel comfortable working with the woman shown),
and sociability. A repeated measures analysis of variance revealed that there
was a main effect for clothing attractiveness on perceived competence (F(l,58)
= 52.14, p = .000). Models dressed in attractive clothing (M = 55.15) were
perceived to be more competent than models dressed in unattractive clothing
(M = 46.49). There was also a main effect for clothing attractiveness on
perceived work comfort (F(l,58) = 24.18, p = .000). Respondents indicated that
they would feel more comfortable working with attractively dressed models (M
= 22.02) than with unattractively dressed models (M = 19.10). Finally, there
was a main effect for clothing attractiveness on perceived sociability (F(l,58)
= 5.28, p = .025). Attractively dressed models (M = 11.41) were perceived to
be more sociable than those models who were unattractively dressed (mean =
10.86). These results provide further support for the view that configurations
of aesthetic elements (i. e., visual organization rules such as unity, color

14
harmony, repetition, etc.) can significantly influence perceptions. The
preceding discussion suggests the following hypothesis:
HI: Aesthetic responses are more positive for objects (products)
exhibiting high consistency with the visual organization principle
of unity than they are for objects (products) that are not
consistent with this visual organization principle.
Prototypicality
Another factor that seems to exert an influence on aesthetic response is
that of experience. Familiarity has been shown to lead to positive affect
(Kunst-Wilson and Zajonc 1980). This would seem to suggest that more typical
or familiar items should be better liked (Loken and Ward 1990). Typicality,
or prototypicality, is concerned with the degree to which an object is
representative of a category. A prototype can be defined as the central
representation of a category, as possessing the average or mean value of the
attributes of that category and as representing the averaged members of the
class (Langlois and Roggman 1990; Rosch 1978). According to the prototypi
cality view, people respond most favorably to objects that are highly protypical
and less favorably to objects that are less prototypical (Glass and Holyoak
1986, p. 170; Langlois and Roggman 1990). The prototypicality explanation of
preference maintains that more prototypical examples tend to be better liked
(Loken and Ward 1990; Nedungadi and Hutchinson 1985). Although prototypi-

15
cality theories were not developed to specifically address aesthetic issues, the
absence of a theory of aesthetic response has led consumer researchers as well
as others to rely on prototypicality as a default theory for explaining aesthetic
response.
A number of explanations have been suggested for the relationship
between prototypicality and preference. One explanation for the relationship
between prototypicality and preference/attitude suggests that more prototypical
items are more familiar and therefore better liked. Familiarity refers to either
an items meaningfulness (i. e., perceived knowledge about an item) or the
frequency of exposure to the item (Loken and Ward 1990). Another
explanation suggests that more prototypical category members are preferred
because they have more valued attributes. This explanation does not hold that
prototypically per se leads to product preference, but rather maintains that as
product categories evolve one or a few products tend to become market-share
leaders because they have attributes widely desired by consumers who buy the
product. Competitive brands are designed to appeal to the same segment(s)
of consumers so they are similar in many ways to market leaders (Loken and
Ward 1990). It has also been suggested that the link between prototypicality
and preference may in part be due to the information theory notion of
redundancy in that prototypes appear to be just those members of a category
that most reflect the "redundancy structure" of the category as a whole (Rosch
1978, p. 37). The preceding discussion suggests the following hypothesis:

16
H2: Objects (products) that are more prototypical (i. e., have
more shared features with the category schema) will receive more
favorable aesthetic responses than objects that are less
prototypical (i. e., fewer shared features with the category
schema).
Although prototypicality/familiarity seems to provide a satisfactory
explanation of aesthetic response in some cases (e. g., Kunst-Wilson and Zajonc
1980; Loken and Ward 1990), such an explanation does not seem adequate in
others (e. g., Meyers-Levy and Tybout 1989). In fact, in some cases it is the
converse of prototypicality (i. e., novelty, distinctiveness) that seems to account
for positive aesthetic response (Woll and Graesser 1982). Some research that
examined novelty and complexity may provide insight into this apparent
inconsistency. In a study that examined novelty ratings of simple and complex
shapes, Eisenman (1968) found that more complex polygons (i. e., those with
more sides) were rated as being more novel. In a related study, Berlyne (1970)
examined the effect of repeated presentation on hedonic value for simple and
complex (i. e., novel) patterns. Berlyne had subjects rate two simple and two
complex patterns six times on a 7-point pleasingness scale. Between
consecutive tests the subjects saw each of the patterns eight times without
having to record a judgement. The results confirmed that ratings of complex
(i. e., novel) patterns rose and then fell after reaching a maximum at the third
test. Ratings of the simple (i. e., less novel) patterns, which were initially

17
higher than those for the more complex (novel) patterns, fell throughout the
tests until they finally flattened out (see Figure II 1). This would seem to
explain why (and roughly when) prototypicality will be liked better in some
cases and novelty will be better liked in others.
Mandler (1982) has theorized that the level of congruity between a
product and a more general product category schema may influence the nature
of information processing and thus product evaluations. Products that are
moderately incongruent with their associated category schemas are said to
stimulate processing that leads to a more favorable evaluation relative to
products that are either congruent or extremely incongruent with the category
schema. Mandler suggests that moderate incongruities are those that can be
successfully and readily resolved by the processor.
Meyers-Levy and Tybout (1989) conducted a series of experiments to test
Mandlers schema (in)congruity hypothesis. The general method, which was
modified slightly over the course of the three studies that were conducted,
consisted of presenting subjects with descriptions and samples of beverage/soft
drink products and having them evaluate the products along dimensions such
as appeal, taste, quality, interest in trial, etc. In their design, schema
congruity and schema incongruity were manipulated by varying a single
attribute in the product description (high preservatives vs. all natural
ingredients) so as to alter the structural and descriptive congruence of the
product description with the schema activated by the category label (beverage

18
Tests
Mean pleasingness ratings for complex (C) and simple (S) patterns in
successive test. Consecutive tests were separated by eight presentations
of each pattern. (Source: Berlyne 1970)
FIGURE II 1
BERLYNES EXPERIMENTAL RESULTS

19
vs. soft drink) that was specified in the first sentence of the product description
(e. g., subjects were told that they were going to be evaluating either a new
"beverage" or a new "soft drink" and later they were told that the product had
either "high preservatives" or had "all natural ingredients"). Across their three
experiments Meyers-Levy and Tybout found support for Mandlers view that
the process of responding to levels of schema congruity influences evaluations,
and that moderate schema incongruity enhances evaluations. Moderate
schema incongruity led to more favorable evaluations than either schema
congruity or extreme schema incongruity. This discussion suggests the
following hypotheses:
H3: Moderate schema incongruity (i. e., distinctiveness, novelty)
leads to a more favorable aesthetic response than does complete
congruity between a product and its product class (i. e.,
prototypicality).
H4: Moderate schema incongruity leads to more favorable
aesthetic responses than does extreme schema incongruity.
It is also possible that people may prefer more novel products due to
variety seeking (Holbrook and Hirschman 1982) or perhaps because of a
products salience relative to other products (Loken and Ward 1990; Woll and
Graesser 1982).

20
H5: Objects (products) that are more novel (i. e., atypical) will
receive more favorable aesthetic response ratings than objects
that are less novel (i. e., more prototypical).
These hypotheses would predict that a product that exhibits a singular
change from the category prototype should receive more positive aesthetic
ratings than either the category prototype or a product that exhibits multiple
changes (i. e., extreme schema incongruity). While this result might be
expected if familiarity is entirely a function of memory, it is possible that the
unity design factor may influence feelings of "perceived familiarity." In an
instance where two attributes of a stimulus were altered such that the
stimulus was no longer prototypical but it did exhibit unity, the relational
similarity (i. e., unity) to the category prototype may generate a sense of
"perceived" familiarity (Goldstone, Medin, and Gentner 1991). This discussion
suggests the following hypothesis:
H6: Consistency with the visual organization principle of unity
is positively related to perceived familiarity.
In addition to the hypotheses that have already been presented
concerning the influence of specific factors on aesthetic response, there are
several other important hypotheses concerning the influence of the aesthetic
responses that are fostered by these factors. These hypotheses are developed
and presented in the section that follows.

21
Derived Responses
Attitude
Aesthetic responses seem to influence derived responses (i. e., non-
aesthetic evaluations) although their influence is often attributed to other
factors (e. g., Berkowitz 1987). One area that seems to be influenced by
aesthetic responses is that of attitude. Bell et al. (1991) have suggested that
aesthetic responses are key determinants of general liking. In their study,
which was described earlier, they found evidence that general liking was
influenced by aesthetic response (recall that aesthetic response in their study
was influenced by perceived unity which was manipulated by means of the
ensemble that was depicted).
In general, "attitudes" may be conceptualized as evaluative judgements.
The term is usually used to refer to an individuals disposition to respond
favorably or unfavorably to an object, person, institution, or event, or to any
other discriminable aspect of an individuals world (Ajzen 1989, p. 241). Ajzen
(1989) has noted that affective reactions may feed into the overall evaluative
response to an attitude object and thus may be at least partly responsible for
the evaluative direction and intensity of a persons beliefs.
The influence of affect due to aesthetic factors (i. e., happy or angry
looking faces, taste, smell) on attitude formation and change was examined in
a study by Edwards (1990). In two experiments, Edwards examined the
hypotheses that the sequence of affect and cognition in an attitudes formation

22
is an important determinant of its subsequent resistance to affective and
cognitive means of persuasion. Affect-based and cognition-based attitudes were
induced and subsequently challenged by either affective or cognitive means of
persuasion. The procedure used to create the two types of attitudes and the
means of persuasion involved varying the sequence of affect and cognition
while holding the content of communications constant. Edwards found that
affect-based attitudes exhibited more change under affective means of
persuasion than under cognitive means of persuasion. Cognition-based
attitudes, on the other hand, exhibited equal change under both forms of
persuasion. In addition, it was found that affect-based attitudes were
expressed with greater confidence than their cognition-based counterparts.
These findings demonstrate that aesthetic responses can play an important
role in determining attitudes.
H7: Attitudes (e. g., general liking) towards products are more
favorable for products receiving more positive aesthetic ratings
than for products receiving less positive aesthetic ratings.
Perceived Quality
Another important area where aesthetic responses seem to play an
important role in influencing non-aesthetic evaluations is that of product
quality. Garvin (1984) has identified aesthetics as one of the eight dimensions
of quality. Zeithaml (1988) has suggested that "intrinsic cues," which involve

23
the physical composition of the product (e. g., texture, color, flavor, etc.), are
very important in signaling perceived quality to the consumer. In a study by
Berkowitz (1987), consumers seemed to make unconscious inferences
concerning freshness, taste, and quality based on the shapes of the products.
Berkowitz examined consumer reaction to a food product frozen corn on the
cob of two shapes (full ears with squared-off ends and full ears with
untrimmed ends), in order to determine: (1) whether the shape of the product
would influence preference; and (2) whether preference levels would vary with
involvement and experience with the product category. The experimental
design involved paired comparison tests at laboratory kitchens in enclosed
malls and sequential monadic tests in subsequent home placements. Test
panelists included 286 female homemakers of which 184 currently purchased
the frozen variety of corn and 102 bought only the fresh variety. The findings
showed a marked preference for the untrimmed shape. Preference ratios
comparing preference scores for the rounded, untrimmed shape to those for the
squared-off shape were:
- Laboratory test 1.1: 1 frozen users; 2.0: 1 fresh only
- Home placement 1.8: 1 frozen users; 2.2: 1 fresh only
The results were statistically significant at the .01 level. Nearly four out of
five consumers said the reason for their choice in the home test was better
taste, about half said the untrimmed was a more natural product, and half
reported better texture. Visual appeal or a more pleasing shape, per se, were

24
very minor motivations. Ratings on ten attributes showed the basis of
consumer preference in a more systematic way. Overall preference ratios were
as follows: more like fresh 3.2; more natural 3.1; taste/flavor 2.3;
quality -- 2.3; size -- 1.7; texture -- 1.5; shape 1.5; and appearance 1.1.
Panelists did not misperceive criteria like ease of preparation (1.0) and ease
of holding (.8) for which shape had objectively little impact.
Berkowitz suggests that the data seem to indicate a chain of interrelated
inferences which stem from the shape of the product rather than a single
direct linkage. He notes that the findings indicate that an attribute
communicated and presumably noticed may not be considered by consumers
to be discriminating, but the attribute(s) that it triggers may be considered to
be discriminating. Thus, the squared-off shape of the one test item may have
fostered an association (or cued categorization) with processing or processed
products while the more natural looking product may have been associated
with freshness or "fun experiences such as summer family barbecues when
fresh corn was served" (Berkowitz 1987, p. 559). Berkowitz notion of
"interrelated inferences" is similar to the concepts of perceptual categorization
and perceptual inferences (Wilkie 1990). Wilkie (1990) has pointed out how
these processes, which translate sensory inputs into a mental "identification"
of a particular stimulus and develop (i. e., construct) beliefs concerning the
stimulus based on other information such as stimulus properties, lead to

25
consumer inferences and thus play a major role in directing consumer behavior
(Wilkie 1990, p. 267).
Evidence for the influence of aesthetic response on quality and
ability/performance evaluations can also be found in the social perception
literature. Landy and Sigall (1974) found significant main effects for writer
attractiveness on evaluations of a writer and her work. Similarly, Lennon
(1990) found a significant effect of clothing attractiveness (clothing and
accessories that matched vs. clothing and accessories that did not match in
color, style or pattern) on perceived competence. The findings of these studies
suggest that the aesthetic aspect of products (i. e., objects) and the aesthetic
responses that they give rise to may exert an influence on non-aesthetic
aspects of products such as quality.
H8: Quality ratings are higher for products that receive more
positive aesthetic ratings than for products that receive less
positive aesthetic ratings.
Price
The possibility that aesthetic response may influence perceived product
quality raises some interesting questions with regard to the price-quality
relationship. This relationship has been examined primarily in terms of price
as cue to quality (Monroe 1973; Zeithaml 1988). Even though a positive price-
perceived quality relationship does appear to exist, results of studies that have

26
examined the price-quality relationship have been somewhat mixed and the
findings imply that price may not be the dominant cue in quality perception
(Monroe 1977; Zeithaml 1988). Moreover, there seem to be cases where
perceptions of (high) quality are formed without being diminished by (low)
price and the quality perceptions subsequently influence price (e. g., the
perceived quality of Japanese automobiles despite their initial low prices).
In cases where consumers initially do not have price information it
seems likely that they might form impressions about a product based on non
price information (e. g., physical composition of the product, packaging, brand
name, etc.) and that these impressions could influence price expectations.
Thus, it is conceivable that the same design factors that influence aesthetic
responses and thereby perceptions of product quality may also influence price
expectations.
H9: Price expectations (ratings) are higher for products that
receive more positive aesthetic ratings than for products that
receive less positive aesthetic ratings.
Ultimately, the price expectations fostered by product design/aesthetics
may play a role determining consumers "price thresholds" (Monroe 1973) for
a particular product within a category.
If indeed unity and prototypicality systematically influence aesthetic
responses and these, in turn, influence product preferences and product

27
perceptions, then one would expect these design factors to influence the ratings
of products in accordance with (the level of) their presence (or absence) in the
products. Aesthetic responses would be expected to be more positive for objects
(products) exhibiting high unity than they would for objects (products) that
were not consistent with this visual organization principle. In general,
prototypicality would be expected to lead to more favorable aesthetic responses
than would atypicality; however, there is some question about the nature of
this relationship with regard to the level of prototypicality (i. e., prototypical
or moderately atypical) that maximizes positive aesthetic response. The unity
and typicality factors would also be expected to influence non-aesthetic
evaluations or derived responses such as attitude, perceived quality, and price
expectations.
In the chapters that follow a series of three experiments that examine
the hypotheses that have been presented here are discussed. Experiments 1
and 2 examine the influence of unity and prototypicality on aesthetic responses
(i. e., HI, H2, H3, H4, and H5) and derived responses (i. e., H7, H8, and H9).
The influence of unity on perceived familiarity is also examined (i. e., H6).
Experiment 3 focuses on the influence of product aesthetics on the attitude and
perceived quality derived responses (i. e., H7 and H8). In this experiment,
additional information is presented during the evaluation task in order to
examine whether or not the presence of the information moderates the
influence of aesthetic response on the derived responses.

CHAPTER III
EXPERIMENT 1
Overview
The hypotheses developed in Chapter II concern the role of design
factors in influencing aesthetic and derived responses. It was hypothesized
that aesthetic and derived responses would be more positive for products
exhibiting high unity than they would be for products that were not consistent
with the unity visual organization principle. It was also suggested that
although there is some question about the level of prototypicality that
maximizes positive aesthetic response, prototypicality, in general, would be
expected to lead to more favorable aesthetic and derived responses. This
chapter first discusses the stimuli and methodology that were used in
Experiment 1 to examine these hypotheses. The analyses and results of
Experiment 1 are then presented and discussed.
Stimuli
In order to examine the hypotheses concerning the influence of unity and
prototypicality stimulus sets were created that allowed these factors to be
manipulated independently. This discussion will first describe how these
28

29
manipulations were accomplished. The manipulations will then be related to
the basic design employed in the experiment. This will be followed by a
general discussion of how the stimulus products employed in this experiment
were selected and developed.
Stimulus Manipulations
The stimulus sets (i. e., design sets) were constructed by first selecting
a prototypical product (i. e., product form) from a product category. Two
prominent parts of the prototypical product were then selected for
manipulation. Three variations of the prototypical product were then produced
by altering either one or the other of the two selected features or both features.
This produced a stimulus set consisting of four product variations (i. e., the
original prototypical product variation, two variations that contained one
altered feature, and one variation in which both features had been altered).
The two product variations that shared one of the selected features with the
prototypical variation (i. e., only one feature had been altered) were moderately
atypical products. The product variation that did not share either of the two
selected features with the prototype (i. e., both features had been altered) was
the most atypical product of the set. Thus, within the stimulus set of four
products three levels of prototypicality were represented (i. e., prototypical,
moderately atypical, and extremely atypical).

30
The transformation of the two product features was done in a way that
also manipulated each product variations consistency with the unity visual
organization principle. In the case of the prototypical variation the two
selected features displayed a visual connection with each other (e. g., a
repetition of the same shape). When one of the two selected features was
altered it was done in a way that decreased unity (e. g., did not display a
visual connection with other parts of the product) and therefore decreased the
unity exhibited by the product variation. This was the case for each of the
variations in which only one feature had been altered (i. e., the moderately
atypical variations). In the case of the product variation where both features
were altered unity was again achieved since the altered features displayed a
visual connection to each other even though the features were very different
from the features of the prototypical variation. Thus, changes in the two
selected product features resulted in two levels of unity (unified and un
unified), and three levels of prototypicality (prototypical, moderately atypical,
extremely atypical).
An example of how unity and prototypicality were manipulated by
transforming two of a products features is shown in Figure III-l. In this
example two features of a telephone (handset and base) are altered using the
shape transformation in order to make the product either more or less
prototypical and more or less unified. The upper left-hand cell of the figure
contains the most prototypical (++) form of the product. The prototypicality of

Feature A
Base
31
(++) Prototypical Handset, Prototypical Base
(+-) Prototypical Handset, Atypical Base
(-+) Atypical Handset, Prototypical Base
(--) Atypical Handset, Atypical Base
FIGURE III-1
EXAMPLE OF UNITY AND PROTOTYPICALITY MANIPULATIONS

32
all of the products that were used as stimuli was established through an
examination of the products available on the market and was confirmed in
pilot tests.1 This product form is also unified in that there is consistency or
compatibility among the shapes of the base (Feature A) and the handset
(Feature B). In this (prototypical) case the number of shared features with the
category schema is high (positive) and unity or a visual connection or
consistency among product features is also high (positive). In the upper right-
hand cell of Figure III-l, the base (Feature A) of the product has been altered
using a shape transformation so that it is no longer the same as that of the
category prototype. This change has at the same time decreased the unity
between the handset feature and the base feature. Thus, this version of the
telephone product is atypical (+-) with respect to one of the two features being
manipulated and is no longer unified. The same is true of the product in the
lower-left cell of Figure III-l. In this case, the shape of the handset was
altered in order to effect the manipulations of prototypicality and unity that
would produce an atypical (-+) stimulus on one product feature and an "un
unified" appearance. The base and handset features of the product in the
lower-right corner of Figure III-l have both been altered. In this case the
product no longer shares the shape of either Feature A or Feature B with the
1 The aggregate mean ratings for the familiarity of the eleven stimulus
sets that were pilot tested were as follows: Unity/Prototypical (M =
6.46), Non-unity/Atypical (M = 4.86), Non-unity/Atypical (M = 4.96), and
Unity/Atypical (4.76). Product variations were rated on 9-point semantic
differential scales with 9 being the most familiar.

33
category prototype. Thus, this product is quite atypical (-) with respect to
prototypicality. This version does, however, exhibit unity since there is a
visual connection or consistency among the two product features. As can be
seen in the figure, this manipulation of the two product features results in
three levels of prototypicality and two levels of unity being produced.
Four possible patterns of results are shown in Figure III-2. These
patterns have been labeled to indicate the explanation of aesthetic response
that each supports. The unity explanation of aesthetic response would be
supported by results showing that both unity conditions were more highly
rated than both of the non-unity conditions (Figure III-2a). Such results would
suggest that a product variations consistency with the unity visual
organization principle positively influenced subjects aesthetic responses. The
predictions of the schema incongruity hypotheses are the opposite of those for
the unity explanation of aesthetic response. The schema incongruity
hypotheses predict that moderately atypical stimuli will be preferred to stimuli
that are prototypical or extremely atypical (Figure III-2b). The prototypicality
explanation of aesthetic response would be supported by the results depicted
in Figure III-2c. Here, the more prototypical a product variation is the higher
its ratings. A pattern of results exactly opposite those for the prototypicality
explanation would suggest an effect of novelty (Figure III-2d). In this way, the
basic 2(Feature A) x 2(Feature B) design allows the hypotheses presented
earlier to be examined.

34
Unity
Feature B
Feature A
Feature B
(a)
Schema Incongruity
Feature A
+
(b)
Feature B
Feature A
+
0
0
-
Feature B
( c )
Novelty
Feature A
-
0
0
-h
( d)
+ indicates the highest (i. e., most positive) rating(s) for a product version
contained in the stimulus set.
- indicates the lowest rating(s) for a product version contained in the set.
0 indicates a rating between the highest and lowest ratings.
FIGURE III-2
POSSIBLE PATTERNS OF RESULTS

35
Stimulus Development
The construction of stimuli that are to be used in research that examines
aesthetic influences requires a great deal of care. Many aspects of an objects
appearance have the potential to affect aesthetic responses. The influence of
aesthetic aspects of a stimulus object (e. g., a product or a picture of a product)
such as color, perspective, shading, etc. that could affect an aesthetic response
must be eliminated or controlled. In addition to the problems and limitations
inherent in controlling for extraneous aesthetic influences, the construction of
visual (i. e., pictorial) stimuli for aesthetic research is further complicated by
the fact that it is usually difficult to precisely determine the strength of
variables of interest (Nunnally 1981).
Although there are quite a number of studies that have manipulated
prototypicality (e. g., Hutchinson and Alba 1991), there is little in the way of
precedent for constructing stimuli that simultaneously exhibit different levels
of prototypicality and unity. The construction of sets of stimuli for this
experiment necessitated an exploration of the ways that unity and
prototypicality could be manipulated across a range of products. There were
four principal requirements that guided the development of the stimulus sets.
The first requirement, which pertained to the selection of particular products,
was that the product class had to have a strong category prototype. In order
to investigate the hypotheses concerning prototypicality it was necessary for
there to be a strong prototypical product design (i. e., form, configuration) for

36
the product category. A second requirement that directed the selection of
products for the stimuli sets concerned the (non)existence of a product that was
atypical but unified. Product categories that contained instances of products
that were atypical (i. e., novel) and unified were considered problematic
because in such cases people may be more favorably disposed (i. e., receptive)
toward atypical (but unified) product designs due to prior experience with or
knowledge of the product. If this were the case it could lead to results that
overstated the influence of the unity design factor. This is especially important
since Experiment 1 effectively provides a theory test between a prototypicality
explanation of aesthetic response and a unity explanation. A third
requirement involved product conduciveness to the manipulation of unity. In
order for a product to exhibit unity (i. e., a visual connection among elements,
repetition of form or pattern) or disunity the product had to have parts that
could be perceived and manipulated separately. The fourth requirement that
guided the development of the stimulus sets concerned the medium used to
create the stimuli and stimulus communicability. A medium was needed that
would allow the creation and presentation of stimulus products that were
drastically different from the products that currently existed. This made it
difficult to construct stimuli by altering (either photographically or using
computer scanned images) existing products or pictures of existing products
because such modification often results in introducing "aesthetic confounds"

37
(e. g., inconsistent perspective) and unintended degradation (e. g., blurry edges,
cloudy surfaces) into a stimulus.
The need for a method that afforded the construction of previously
"uncreated" (i. e., nonexistent) products that were markedly different from
existing products and the need to control for extraneous aesthetic influences
led to the use of line drawings. The use of drawings does entail a trade-off of
realism for "producibility" and greater experimental control. While this
reduction in realism is unfortunate, it is not uncommon for research that is
conducted in a laboratory setting. The use of drawings was also more practical
for the purpose of reproducing the stimuli for inclusion in the booklets used in
the experiment. This method of constructing stimuli necessitated the selection
of products that could be clearly communicated through simple line drawings.
A number of products in a wide variety of product categories were
examined in catalogs and in stores in order to determine their suitability for
use as stimuli in this experiment. An initial series of product, "studies" (i. e.,
sketches) was done to explore ways of manipulating products along the unity
and prototypicality dimensions. A number of products that seemed to meet the
four requirements were then selected for further development in product
2
The possibility of modifying computer scanned images was explored.
This approach proved to be unsatisfactory for the purposes of creating
the stimuli for this particular experiment because the radical changes
that had to be made to the computer scanned images introduced
aesthetic confounds into the stimuli and often resulted in severe
degradation of the image.

38
studies. A list of these products is presented in Appendix A. These product
studies led to the identification of three basic ways to transform the features
of the products so as to simultaneously accomplish the prototypicality and
unity manipulations. These transformations involved altering product features
by changing the shape of parts of the product, adding texture to parts of the
product, or adding trim to part of the product. Additional product studies were
then undertaken to determine which products could be manipulated using each
of the three transformations (shape, trim, and texture). These studies involved
producing four versions of each product (for each of the three types of
transformations) by systematically transforming two product features in order
to accomplish the unity manipulation (i. e., a visual connection among product
features) at a particular level of prototypicality (i. e., shared features with the
category prototype).
In order to increase the generalizability of the results and demonstrate
the robustness of the effects of interest, nine product categories were selected
for use as stimuli. These nine product categories were: alarm clocks, bathroom
scales, dressers, flashlights, hair dryers, lamps, refrigerators, telephones, and
television remote controls. This resulted in a total of twenty-seven replications
(nine product categories x three types of transformation). Each of these
products had to be produced in four versions (Unity/Prototypical ++, Non
unity/Atypical +-, Non-unity/Atypical -+, and Unity/Atypical ). Thus, the

39
entire stimuli set contained one hundred and eight drawings of products.
Examples of these stimuli sets are presented in Appendix B.
Experimental Design
The overall design of the experiment is a 2(Feature A) x 2(Feature B) x
9(Products) x 3(Version) x 2(Order) mixed factorial design. As was pointed out
in the preceding discussion the manipulation of features A and B result in the
manipulation of prototypicality and unity. The twenty-seven replicate sets of
stimuli were organized into three questionnaire versions. Each of the three
questionnaire versions contained all nine of the product categories (each
product category was made up of a set of four variations of the product i. e.,
Feature A x Feature B) but for only one of the three transformation types
(shape, trim, or texture). So for example, questionnaire version one contained
the telephone stimulus set (set of four product variations) that were altered
using the trim transformation; questionnaire version two contained the
telephone stimulus set altered using the texture transformation; and
questionnaire version three contained the telephone stimulus set altered using
the shape transformation. While each questionnaire contained each of the nine
product categories, in three of the nine cases the trim transformation was
utilized, in three of the nine cases the texture transformation was utilized, and
in three of the nine cases the shape transformation was utilized. A diagram
of this design is presented in Appendix C. Thus, Feature A, Feature B,

40
Products, and Transformation were all within-subjects factors. Particular
transformations of the nine product categories (i. e., stimulus sets) were
contained in the questionnaire Versions. Questionnaire Version and Order
were between-subjects factors. The order in which the stimulus products
(stimulus sets) were presented was reversed for half of the subjects.
Experimental Procedure
One-hundred and ninety-seven volunteer subjects enrolled in the
Introductory Marketing course at the University of Florida participated in this
O
experiment. Subjects received extra credit for their participation. The
subjects were run in groups of 10-20 participants. The stimulus materials
were contained in a booklet. The introductory page informed the subjects that
they would be shown drawings of products and asked to evaluate the appeal
of the product ideas based on their appearance. The subjects were told that
the purpose of the study was to obtain consumers reactions to products that
companies were considering for introduction. It was also explained to the
subjects that the products they were going to be evaluating were in the early
stage of the product development process and for this reason the drawings
were in a very rough (unfinished) form. Subjects were then told that each of
the versions of a product performed equally well and that they were to rate all
O
Power estimates using pilot data and the procedure suggested by Cohen
(1977, pp. 364-379) for factorial designs indicated that this sample size
would be sufficient for estimating effect sizes at the p < .05 level.

41
four of the product versions that were shown on each page. Subjects were then
allowed to proceed through the task of rating the products at their own pace.
The subjects rated each product design on 9-point semantic differential scales
that measured aesthetic response (beautiful/ugly), familiarity
(familiar/unfamiliar), attitude (like/dislike), quality (high quality/low quality),
and price (high price/low price).^ Following the rating tasks, subjects were
asked to write about how they determined the ratings that they gave to the
proposed products. The entire procedure took approximately one-half hour to
complete.
Results and Discussion
The manipulation check of the effect of a change in either Feature A or
Feature B on aesthetic responses found that there was, in fact, a reduction in
ratings of the appearance of the products (compared to those of the prototypical
product version) as measured on the semantic differential scale anchored by
Beautiful/Ugly. This is to say that in the case of the product variations that
were altered so that one feature (either Feature A or Feature B) was no longer
prototypical or unified the effect of the change was a reduction in positive
aesthetic response. The average effect of a change in Feature A on aesthetic
responses was -1.24. Likewise, the average effect of a change in Feature B on
4
The direction of some of these scales was reversed in order to reduce
possible response bias on the part of subjects.

42
aesthetic response was a -1.34. This pattern was consistent across most of the
sets of stimuli for each product transformation (i. e., stimulus sets).5 The
effects of a change in Feature A or Feature B from the category prototype are
shown in Table III-l. In twenty-one of the twenty-seven stimulus sets the
change of either Feature A or Feature B resulted in a lower beauty rating for
the product variation as compared with the rating for the category prototype.
In two stimulus sets the effects of a change in the product features were
mixed. That is, a change in one feature (either A or B) had a negative effect
on aesthetic response but a change in the other feature (B or A) had a positive
effect (see the lamp/trim and lamp/texture stimulus sets in Table III-l). In
only four of the twenty-seven stimulus sets were the effects of a change in
Feature A and Feature B both (separately) in the opposite (i. e., positive)
direction (see the bathroom scale/trim, bathroom scale/texture, TV remote
control/trim, and the hair dryer/trim stimulus sets in Table III-l). Overall, the
manipulation check demonstrates that for the majority of the stimulus sets a
reduction in favorable aesthetic response does occur when the products are
altered in such a way as to be both less prototypical and less unified than the
category prototype. It should be noted that there is variation in the reduction
in favorable aesthetic response across the twenty-seven stimulus sets. This is
to be expected given the difficulties of precisely estimating the strength of each
5 "Stimulus set" will be used throughout this discussion to refer to one of
the transformation types (shape, trim, or texture) of a product category
(e. g., telephone/shape transformation set).

43
TABLE III 1
MANIPULATION CHECK FOR THE EFFECT OF A CHANGE
IN FEATURE A OR B ON BEAUTY RATINGS
Product
Feature A
Feature B
1. Bathroom scale
Sh
-0.14
-0.17
Tr
1.00
0.14
Tx
1.77
0.86
2. TV Control
Sh
-0.73
-0.22
Tr
0.65
0.97
Tx
-0.98
-0.84
3. Flashlight
Sh
-1.83
-1.16
Tr
-3.04
-3.55
Tx
-2.36
-3.31
4. Lamp
Sh
-0.16
-0.14
Tr
0.02
-1.28
Tx
-0.07
1.06
5. Refrigerator
Sh
-4.00
-3.60
Tr
-1.09
-1.12
Tx
-3.80
-3.64
6. Telephone
Sh
-2.60
-1.01
Tr
-0.28
-1.25
Tx
-0.56
-0.99

44
TABLE III 1 continued
MANIPULATION CHECK FOR THE EFFECT OF A CHANGE
IN FEATURE A OR B ON BEAUTY RATINGS
Product
Feature A
Feature B
7. Hair dryer
Sh
-1.59
-3.72
Tr
0.33
0.33
Tx
-3.58
-3.77
8. Dresser
Sh
-0.13
-0.54
Tr
-1.88
-1.75
Tx
-2.07
-1.20
9. Clock
Sh
-3.33
-3.00
Tr
-1.43
-1.44
Tx
-1.60
-1.72
Mean effect of change
across products
-1.24
-1.34
manipulation (i. e., change in Features A and B via the transformation type)
as it is adapted for each of the stimulus sets.
The Influence of Unity and Prototypicality on Aesthetic and Derived Responses
The critical test concerns what happens to aesthetic responses in the
case where both Feature A and Feature B are altered so that they are less
prototypical but more unified. The prediction consistent with the
"prototypicality" hypothesis is that in such a case the aesthetic response
ratings should go down since the resulting stimulus is quite atypical. The

45
"unity" hypothesis, however, suggests that if the changes in the features are
made in a way that increases unity then aesthetic responses should be higher
in this case (as opposed to the non-unity cases) since the product version
exhibits consistency with the visual organization principle of unity. This is in
effect a theory test (for a linear model) which corresponds to a test of the
Feature A x Feature B interaction. This test can be extended to the
investigation of the influence of unity and prototypicality on derived responses.
Analyses
The data were analyzed using mixed ANOVA designs with Version and
Order being treated as between-subjects factors and Feature A, Feature B, and
Products being treated as within-subjects factors. Transformation was also
treated as a within-subjects factor because each subject saw products that had
been altered using the three different transformation types.
Subjects product ratings on the five 9-point semantic differential scales
were treated in three different ways in order to compute the dependent
measures that were analyzed. The first approach was to perform the analysis
directly on subjects ratings of the products. A second approach involved
computing the linear contrast that reflected the interaction of Feature A and
Feature B (i. e., the main effect of unity) for each design set. Under this
approach difference or "interaction" scores were formed by adding subjects
ratings for the Unity/Prototypical and Unity/Atypical conditions and

46
subtracting from this the ratings for both of the Non-unity/Atypical conditions.
The resulting score provides a straight-forward measure of the main effect of
unity. The third way of forming the dependent measures involved a
modification of the interaction score. This variation entailed subtracting the
product rating of the lowest rated Non-unity/Atypical condition from the
product rating for the Unity/Atypical condition. This score allows a non-linear
model of prototypicality to be examined (see Figure III-3). If prototypicality
dominates unity then product versions that are more similar to the category
prototype should receive higher ratings than less typical product versions (i. e.,
the minimum rating for either of the Non-unity/Atypical conditions should
always be greater than the rating for the Unity/Atypical condition since the
former always has more shared features with the category prototype than does
the latter).
Both the unity linear score (U-jO and the unity nonlinear score (Ujj) are
needed in order to get a (more) complete sense of the effects of unity and
prototypicality. While the score does provide an indication of the main
effect of unity, it does not examine a (decreasing) non-linear model of
prototypicality. The score does examine a non-linear model of
prototypicality; however, the U-^ score reflects some random variation in the
subjects ratings (i. e., the lowest rated Non-unity/Atypical condition is always
used to construct the score). Parallel analyses were conducted on each of
the three variations of dependent measures (i. e., ratings scores, linear unity

47
Linear Model
Aesthetic
(or Dervied)
Response
C + + ) ( + ) (- + ) (--)
TJnity/Prototypicality
Nonlinear Model
Aesthetic
(or Derived)
Response
( + + )
( + ) ( + )
U nity/Pr ototypicality
FIGURE III-3
LINEAR AND NONLINEAR "UNITY" SCORES

48
scores, and nonlinear unity scores) in order to insure that the results reported
were not simply an artifact of a particular analysis scheme.
Hypothesis Testing
The critical test for examining the hypotheses concerning the influence
of unity and prototypicality on aesthetic and derived responses involves the
Feature A x Feature B interaction. In this design, the Feature A x Feature B
interaction is the main effect of unity. This interaction "tests" the competing
predictions of the prototypicality and unity hypotheses. If the
Unity/Prototypical (++) products of this design receive higher ratings than the
Non-Unity/Atypical (+-) and Non-unity/Atypical (-+) products and these in turn
are more highly rated than the Unity/Atypical (--) products then the
prototypicality explanation of aesthetic (and derived) response would seem to
be supported (Hypothesis 2). If, however, the Unity/Prototypical (++) products
and the Unity/Atypical (--) products were both rated significantly higher than
the Non-unity/Atypical (+-, -+) products then support for the unity hypothesis
would be indicated (Hypothesis 1). The schema incongruity explanation of
aesthetic response would be supported if the Non-unity/Atypical (+-, -+)
products were rated significantly higher than the prototypical products (++)
and/or the extremely atypical products (). If aesthetic response was found to
be positively related to atypicality than this would suggest a novelty effect.

49
The means for the product beauty ratings for each of the twenty-seven
stimulus sets are presented in Table III-2. Across the twenty-seven stimulus
sets the products that did not exhibit unity (M = 4.06) were rated lower than
the products that did exhibit unity (M = 5.12). The analysis of variance
presented in Table III-3 indicates that this difference which is captured in the
Feature A x Feature B interaction is significant F (1,184) = 338.16, p < .0001.
This supports the unity hypothesis which predicted that aesthetic responses
are more positive for objects (products) that exhibit high consistency with the
visual organization principle of unity than they are for objects (products) that
are not consistent with the unity visual organization principle.
The significance of the Factor A x Factor B interaction (i. e., main effect
of unity) was also tested in parallel analyses that utilized the linear unity
scores (U--_) formed from subjects ratings of all four versions of each stimulus
set [i. e., (Unity/Prototypical + Unity/Atypical) (Non-unity/Atypical + Non-
Unity/Atypical)] and the nonlinear unity scores (U^) [i. e., (Unity/Atypical) -
minimum (Non-unity/Atypical)]. The analysis of variance tables for each of
these approaches are shown in Table III-4 and Table III-5. The unity
hypothesis (i. e., Feature A x Feature B interaction) is tested by determining
whether or not the intercept is zero. That is, if the intercept were actually
zero, what would the probability be of obtaining, by chance alone, a value as
large or larger than the one actually obtained? The results reported in Table

TABLE III 2
MEAN BEAUTY RATINGS AND UNITY SCORES
50
Product
(++)
(+-)
(-+)
(-)
uT,
Un
1. Bathroom Scale
Sh
5.30
5.16
5.13
5.18
0.85
0.88***
Tr
4.42
5.42
4.56
5.81
0.12*b
1.38***
Tx
4.03
5.80
4.89
4.81
-0.89
0.43+
2. TV Control
Sh
4.82
4.09*a
4.60
4.95
0.51**
1 41***
Tr
3.98
4.63
4.95
6.42
0.44*
2.23***
Tx
5.36
4.38**
4.52**
5.28
0.84**
1.36***
3. Flashlight
Sh
6.31
3.95***
3.00***
3.05***
1.20***
0.55**
Tr
5.57
3.74***
4r 41***
3 97***
0 70***
0.61*
Tx
5.89
2.85***
2.34***
5.15*
2 94***
3.00***
4. Lamp
Sh
5.64
5.66
4.36***
4.36**
0.00
0.39*
Tr
4.65
4.58
5.71
5.92
0.14
1 42***
Tx
5.00
4.84
4.86
6.05
0.68**
2.08***
5. Refrigerator
Sh
6.00
2.00***
2.40***
3.95***
2.80***
2.20***
Tr
5.31
4.22**
4.19**
5.70
1.28***
2.03***
Tx
6.03
2.23***
2.39***
3.34***
2.34***
1.32***
6. Telephone
Sh
5.30
2 70***
4.29**
3.48***
0.90***
0.94**
Tr
5.44
5.16+
4.19***
4.18***
0.15
0.21+
Tx
5.74
5.18**
4 75***
6.02
0.90***
1 41***

51
TABLE III 2 continued
MEAN BEAUTY RATINGS
Product
(++)
(+-)
(-+)
(-)
UL
uN
7. Hair dryer
Sh
6.57
4 gg***
2.85***
3.30***
0.90***
0.61**
Tr
4.64
4.98
4.98
5.42
0.03
0.92***
Tx
6.33
2.75***
2.56***
5.17**
3.06***
2.94***
8. Dresser
Sh
5.08
4.95
4.54*
5.66
0.62***
1.80***
Tr
5.44
3.56***
3.69***
6.61
2 4^***
3.26***
Tx
5.64
3.57***
4 44**
4.54**
0.99***
1.42**
9. Clock
Sh
5.72
2.39***
2.72***
4.75**
2.65***
2.58***
Tr
4.92
3.49***
3.48***
4.26**
1 07***
1.03***
Tx
5.17
3.57***
3.45***
5.00
1.56***
1 71***
Totalc
5.34
4.10***
4.01***
4 90***
1.05***
1.48***
a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
b Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.

52
TABLE III 3
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
42.9350
1.63
0.1991
Order
1
13.7790
1.05
0.3080
Version Order
2
107.6630
4.08
0.0184
Error Sub (Version Order)
184
2425.7720
Within Subiect Effects
Product
8
1081.1417
40.78
0.0001
Trans
2
165.0833
24.91
0.0001
Prod Version (Residual)
14
566.1260
12.02
0.0001
Prod Order
8
199.3217
7.52
0.0001
Trans Order
2
5.4222
0.82
0.2500
Prod Version Order (Residual)
14
43.1955
0.93
0.2500
Error Sub (Product)
1472
4878.0299
Feature A
1
124.2720
23.97
0.0001
A Version
2
252.2238
24.33
0.0001
A Order
1
10.6083
2.05
0.1542
A Version Order
2
6.1115
0.59
0.5556
Error Sub (A)
184
953.7624
Feature B
1
52.8499
11.86
0.0007
B Version
2
114.6225
12.86
0.0001
B Order
1
21.7458
4.88
0.0284
B Version Order
2
17.8531
2.00
0.1378
Error Sub (B)
184
819.8276
Feature A Feature B
1
1941.8813
338.16
0.0001
A B Version
2
67.4647
5.87
0.0034
A B Order
1
3.1291
0.54
0.4614
A B Version Order
2
5.6170
0.49
0.6140
Error (A B)
184
1056.6273
Prod A B
8
933.9437
53.26
0.0001
Prod A B Version
16
787.4382
22.45
0.0001
Prod A B Order
8
94.0783
5.37
0.0001
Prod A B Version Order
16
35.9705
1.03
0.4255
Error Sub (Prod A B)
1472
3226.3444

53
TABLE III 4
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
LINEAR SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
7767.5254
338.16
0.0001
Version
2
269.8589
5.87
0.0034
Order
1
12.5163
0.54
0.4614
Version Order
2
22.4679
0.49
0.6140
Error Sub (Version Order)
184
4226.5090
Within Subiect Effects
Product
8
3735.7748
53.26
0.0001
Trans
2
549.8955
31.36
0.0001
Prod Version (Residual)
14
2599.8573
21.18
0.0001
Prod Order
8
376.3133
5.37
0.0001
Trans Order
2
40.2259
2.29
0.2500
Prod Version Order (Residual)
14
103.6562
0.84
0.2500
Error Sub (Prod)
1472
12905.3777
III-4 and Table III-5 clearly indicate support for the unity hypothesis. The
intercept using the linear unity score (Ut) as the dependent measure was
significantly different from zero F (1,184) = 338.16, p < 0.0001. The
intercept as tested using the nonlinear unity score (U-^) approach was also
significantly different from zero F(l,190) = 473.48, p < 0.001.^ The fact that
there is a small, but significant t (190 = 6.04, one-tailed p < .001) difference
between the means of the Unity/Prototypical (++) products (M = 5.34) and the
/2
The results for the linear unity scores are identical to the results for the
ratings score approach.
7
The difference in the denominator degrees of freedom are due to missing
data that caused several subjects to be eliminated from the "unity score"
analysis.

54
TABLE III 5
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
NONLINEAR UNITY SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subject Effects
Intercept
1
3873.5708
473.48
0.0001
Version
2
340.8709
20.83
0.0001
Order
1
3.2588
0.40
0.5287
Version Order
2
1.3466
0.08
0.9210
Error Sub (Version Order)
190
1554.4164
Within Subject Effects
Product
8
298.1844
9.67
0.0001
Trans
2
63.3658
8.21
0.0001
Prod Version
14
484.9146
8.98
0.0001
Prod Order
8
123.7531
4.01
0.0100
Trans Order
2
41.8300
5.42
0.0100
Prod Version (Residual)
14
75.8172
1.40
0.2500
Error Sub (Prod)
1520
5861.4233
Unity/Atypical () products (M = 4.90) suggests that prototypicality does have
some effect on aesthetic responses.
The analysis of variance presented in Table III-3 as well as those shown
in Table III-4 and Table III-5 also show that there were a number of factors
and interactions among factors that resulted in significant effects in this
design. Most of these significant effects can be attributed to differences in the
strength of the manipulations used to create each stimulus set and the
differences across product categories. The effect of Product was significant,
F (8,1472) = 40.78, p < .0001, as was the effect of Transformation type
F (2,1472) = 24.91, p < .0001. Since stimulus sets (i. e., Transformation types)
are contained in each of the versions of the questionnaire the effects of these

55
factors are also indicated in the Product x Version (Residual) interactions.8
The differential strength of the manipulations across stimulus sets was further
examined in order to ensure that the effects were in the right direction across
the twenty-seven stimulus sets. This was done by conducting t-tests using the
"unity" scores for each of the stimulus sets in order to assess the degree to
which the effect of unity was significant and in the proper direction across the
twenty-seven stimulus sets despite the differences in the strength of the
manipulations. The analysis for the linear unity scores is shown in Table III-6
(the unity scores were also presented in Table III-2). In twenty-one out of the
twenty-seven stimulus sets the linear unity scores are significantly different
from zero (p < .05; p < .01 for nineteen of the stimulus sets). In six cases the
interaction scores were not significant and in only one case was a score
negative and significant (i. e., in the "wrong" direction).^ The findings are
similar when the analysis is conducted using nonlinear unity scores. This
analysis is presented in Table III-7. When the differential strength of the
Q
In this design the main effect of Transformation type is examined using
a Latin square type orthogonal fraction of the complete five factor
design. This is consistent with the treatment suggested by Winer,
Brown, and Michels (1991, pp. 706-711) for Latin squares and related
designs.
Q
This negative case occurred for the texture manipulation of the
bathroom scale stimulus set. All three of the bathroom scale
product/cases were found to be somewhat problematic because this
product was not easily divided into two separate parts (e. g., telephone
stimulus handset, base). This was particularly true for the texture
manipulation of the bathroom scale (see bathroom scale/texture stimuli
in Appendix B).

56
TABLE III 6
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES -
LINEAR UNITY SCORES
Product
UtiScore for
Tleauty
S. E.
t-test
1. Bathroom scale
Sh
0.8335
0.2032
0.8201
Tr
0.1212
0.1449
1.6726*
Tx
-0.8923
0.4442
-4.0172
2. TV Remote Control
Sh
0.5076
0.3853
2.6345**
Tr
0.4384
0.3913
2.2410*
Tx
0.8409
0.4711
3.5696**
3. Flashlight
Sh
1.2032
0.3527
6.8233***
Tr
0.6969
0.3547
3.9298***
Tx
2.9394
0.4888
12.0267***
4. Lamp
Sh
0.0000
0.2412
0.0000
Tr
0.1385
0.2787
0.9938
Tx
0.6770
0.4214
3.2124**
5. Refrigerator
Sh
2.7955
0.4302
12.9961***
Tr
1.2770
0.4308
5.9281***
Tx
2.3359
0.4761
9.8119***
6. Telephone
Sh
0.9000
0.4309
4.1773***
Tr
0.1485
0.2062
1.4401
Tx
0.9015
0.4160
4.3341***

57
TABLE III 6 continued
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES -
LINEAR UNITY SCORES
Product
Uj Score for
U3eauty
S. E.
t-test
7. Hair dryer
Sh
0.8940
0.4167
4.2902***
Tr
0.0303
0.3029
0.2000
Tx
3.0539
0.5295
11.5353***
8. Dresser
Sh
0.6137
0.3105
3.9526***
Tr
2.4077
0.4661
10.3311***
Tx
0.9925
0.5844
3.3966***
9. Clock
Sh
2.6462
0.4438
11.9243***
Tr
1.0682
0.4164
5.1308***
Tx
1.5531
0.3905
7.9536***
*** Indicates differences significant at p < .001 in one-tailed-tests
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.
manipulations is examined by looking at the difference between the
Unity/Atypical product and the lowest rated Non-Unity/Atypical product for
each stimulus set, all of the scores are positive and the scores for twenty-five
of the stimulus sets are significant at the p < .05 level (twenty-two of these are
significant at the p < .01 level or higher). These analyses suggest that even
though there are differences in the strengths of the manipulations due to
transformation type and product class, the effect of unity is relatively
consistent (i. e., significantly positive) across both products and
transformations.

58
TABLE III 7
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES -
NONLINEAR UNITY SCORES
Product
Ujt Score for
Beauty
S. E.
t-test
1. Bathroom scale
Sh
0.8788
0.1985
4.4272***
Tr
1.3788
0.1988
6.9359***
Tx
0.4308
0.2816
1.5296+
2. TV Remote Control
Sh
1.4091
0.2653
5.3113***
Tr
2.2308
0.2287
9.7551***
Tx
1.3636
0.2981
4.5741***
3. Flashlight
Sh
0.5539
0.2352
2.3551**
Tr
0.6061
0.2678
2.2628*
Tx
3.0000
0.3260
9.2017***
4. Lamp
Sh
0.3939
0.1753
2.2479*
Tr
1.4242
0.2275
6.2592***
Tx
2.0769
0.2768
7.5047***
5. Refrigerator
Sh
2.1969
0.2800
7.8451***
Tr
2.0308
0.2685
7.5629***
Tx
1.3182
0.2663
4.9493***
6. Telephone
Sh
0.9394
0.2678
3.5073**
Tr
0.2121
0.1441
1.4725+
Tx
1.4091
0.2443
5.7688***

59
TABLE III 7 Continued
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES -
NONLINEAR UNITY SCORES
Product
Ujyr Score for
Beauty
S. E.
t-test
7. Hair dryer
Sh
0.6061
0.2470
2.4536**
Tr
0.9242
0.2226
4.1516***
Tx
2.9385
0.3375
8.7074***
8. Dresser
Sh
1.8030
0.2416
7.4624***
Tr
3.2615
0.2375
13.7356***
Tx
1.4242
0.4108
3.4666**
9. Clock
Sh
2.5846
0.2556
10.1118***
Tr
1.0303
0.2471
4.1696***
Tx
1.7121
0.2068
8.2777***
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.
While an interaction of order with the effect of unity (i. e., the Feature
A x Feature B interaction) was not indicated, order was found to interact with
factors such as Product. In part, the effect of order seems to be attributable
to subjects increasing acceptance of the use of the simplified (i. e., line
drawings without color) drawings to depict products. Such an explanation is
suggested by the fact that, in general, when the stimuli for a particular
product category were rated toward the end of the rating task they received
slightly higher ratings than when they were rated at the beginning of the

60
rating task. This was controlled for to some degree by reversing the
presentation order of stimuli in the second order.
Perceived Familiarity
Hypothesis 6 predicted that consistency with the visual organization
principle of unity was positively related to perceived familiarity. The means
for the familiarity ratings for each of the twenty-seven stimulus sets are
presented in Table III-8. The analysis of variance tables for the rating score,
linear unity score, and nonlinear unity score approaches are presented in
Tables IIX-9, HI-10, and III-ll, respectively.
The analysis of variance for the subject ratings of familiarity (Table III-
9) indicates a significant unity effect for perceived familiarity F (1,181) =
254.47, p < .0001. The results for the analysis of variance using linear unity
scores were very similar to those observed using subject ratings with the unity
effect achieving the same level of significance F (1,181) = 254.47, p < .0001.
The analysis of variance for the nonlinear score approach, which is probably
the most appropriate analysis in this case since it removes the influence of the
Unity/Prototypical cell of the design, also indicates a significant effect of unity
on perceived familiarity F (1,188) = 10.65, p < .0013.^ These results
suggest that unity can encourage feelings of perceived familiarity.
The pattern of interactions observed among Product, Transformation
type, and Order were the same as were discussed for the influence of
unity on aesthetic responses.

TABLE III 8
MEAN FAMILIARITY RATINGS AND UNITY SCORES
61
Product
(++)
(+-)
(-+)
(-)
UL
%
1. Bathroom scale
Sh
7.62
7 14**a
5.40***
5.56***
0 3i**b
0.45*
Tr
7.06
6.97
7.13
6.98
0.025
0.27
Tx
7.05
6.77
6.23*
6.10**
0.03
0.42
2. TV Remote Control
Sh
7.59
5.08**
5 44***
4 94***
0.97***
0.65
Tr
6.62
6.13*
6.51
6.89
0.53**
1 45***
Tx
7.35
5.17***
5.32***
5.86***
1 31***
1.00***
3. Flashlight
Sh
7.93
4 80***
4.16***
4.05***
1 54***
0.27
Tr
7.49
5.52***
5.79***
5.49***
0.84***
0.35
Tx
7.49
4.52***
4 27***
6.35***
2.43***
2.21***
4. Lamp
Sh
8.22
5.95***
3.98***
3.57***
0.89***
-0.18
Tr
7.21
6.10***
6.81*
6.60*
0.45**
0.67**
Tx
7.43
6.38**
6.18**
7.46
4 49***
2.02***
5. Refrigerator
Sh
7.97
3.16***
3.25***
3.84***
2.68***
0.92***
Tr
7.75
5.52***
5.28***
6.61***
1.88***
^ 74***
Tx
8.22
4.21**
4 29***
4.52***
2.16***
0.67**
6. Telephone
Sh
7.26
3.84***
4 77***
3 41***
1.10***
-0.31
Tr
7.16
6.19***
5.46***
5.52***
0.50***
0.34**
Tx
6.89
6.24**
5.73***
6.41*
0.65***
0.82**

62
TABLE III 8 continued
MEAN FAMILIARITY RATINGS AND UNITY SCORES
Product
(++)
(+-)
(-+)
(--)
UL
uN
7. Hair dryer
Sh
8.05
6.48***
4 46***
4.40***
0.83***
0.21
Tr
7.05
6.38**
6.54*
6.44**
0.33*
0.52**
Tx
7.67
3.98***
4.30***
5.69***
2.68***
2.02***
8. Dresser
Sh
7.76
6.62***
5.98***
6.29***
0.73**
0.56*
Tr
7.34
4.87***
4 97***
7.51
2.52***
2.95***
Tx
7.43
5.68***
5.63***
6.37*
1 18***
1.15***
9. Clock
Sh
7.28
4 51***
4 75***
6.43**
2.19***
2.09***
Tr
7.33
5.68***
5.78***
6.08***
0.97***
0.77**
Tx
7.05
5.44***
5.76***
6.86
1.32***
1.50***
Total0
7.45
5.53***
5.34***
5.79***
0 94***
Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
k Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.

63
TABLE III 9
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
73.7530
0.55
0.5789
Order
1
233.9050
3.48
0.0638
Version Order
2
8.6500
0.06
0.9377
Error Sub (Version Order)
181
12174.5330
Within Subiect Effects
Product
8
946.3903
20.96
0.0001
Trans
2
724.2835
64.16
0.0001
Prod Version (Residual)
14
677.6986
8.58
0.0001
Prod. Order
8
38.3970
0.85
0.5582
Trans Order
2
46.4860
4.12
0.0250
Prod Version Order (Residual)
14
189.0356
2.39
0.0100
Error Sub (Product)
1448
8172.4271
Feature A
1
1459.8545
259.23
0.0001
A Version
2
300.8827
27.71
0.0001
A Order
1
7.6979
1.37
0.2439
A Version Order
2
17.3149
1.54
0.2177
Error Sub (A)
181
1019.2983
Feature B
1
919.4761
185.88
0.0001
B Version
2
14.4650
1.46
0.2345
B Order
1
1.4952
0.30
0.5831
B Version Order
2
4.1657
0.42
0.6570
Error Sub (B)
181
895.3306
A B
1
2362.5960
254.47
0.0001
A B Version
2
69.0487
3.72
0.0261
A B Order
1
21.4075
2.31
0.1306
A B Version Order
2
22.0981
1.19
0.3066
Error Sub (A B)
181
1680.4642
Prod A B
8
549.3698
28.33
0.0001
Prod A B Version
16
434.3241
11.20
0.0001
Prod A B Order
8
109.7385
5.66
0.0001
Prod A B Version Order
16
52.1372
1.34
0.1618
Error Sub (Prod A B)
1448
3510.0931

64
TABLE III 10
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
LINEAR UNITY SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
71.9460
10.65
0.0013
Version
2
426.8649
31.61
0.0001
Order
1
54.3667
8.05
0.0050
Version Order
2
11.3613
0.84
0.4328
Error Sub (Version Order)
188
1269.5378
Within Subiect Effects
Product
8
727.7061
20.07
0.0001
Trans
2
661.0854
72.93
0.0001
Prod Version (Residual)
14
557.1115
8.78
0.0001
Prod Order
8
60.4733
1.67
0.1016
Trans Order
2
1.2788
0.14
0.2500
Prod Version Order (Residual)
14
84.6541
1.33
0.2500
Error Sub (Prod)
1504
6816.5729

65
TABLE III 11
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
NONLINEAR UNITY SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
9450.3839
254.47
0.0001
Version
2
276.1947
3.72
0.0261
Order
1
85.6300
2.31
0.1306
Version Order
2
88.3924
1.19
0.3066
Error Sub (Version Order)
181
6721.8569
Within Subiect Effects
Product
8
2197.4792
28.33
0.0001
Trans
2
395.5843
20.40
0.0001
Prod Version
14
1341.7119
9.88
0.0001
Prod Order
8
438.9539
5.66
0.0001
Trans Order
2
9.8527
0.51
0.2500
Prod Version Order (Residual)
14
198.696
1.46
0.2500
Error Sub (Prod)
1448
14040.3723
Derived Responses
The remaining three hypotheses concerned the effect of aesthetic
response on derived responses (i. e., attitude, quality, and price). It was
hypothesized that aesthetic responses would influence derived responses and
thus products that received more positive aesthetic ratings (i. e., products that
exhibit unity) would also be rated higher with regard to liking (i. e., attitude),
quality (i. e., perceived quality), and price (i. e., expected price) than would
products that received less positive aesthetic ratings. The critical tests for
examining these hypotheses are the same as the critical test for examining the
influence of unity on aesthetic response except the dependent measures reflect

66
the construct of interest (e. g., attitude, perceived quality). Thus, if the
products that received higher (more positive) aesthetic response ratings also
receive significantly higher ratings with respect to scales measuring derived
responses than this would provide at least some evidence of the broader
influence that product aesthetics may exert on consumer behavior.
The analyses and findings concerning attitude, perceived quality, and
price expectations will be discussed together since the analyses employed were
identical for all three and the results were for the most part parallel. The
means for the liking, quality, and price ratings are presented in Table III-12,
Table XII-13, and Table III-14, respectively. For the most part, the pattern of
the means that was observed for the liking and quality derived responses was
similar to the pattern observed in the case of aesthetic response. That is, both
the Unity/Prototypical and the Unity/Atypical conditions received higher
ratings than either of the two Non-unity/Atypical conditions. The pattern
observed for the price ratings was somewhat different. The means for the
price ratings across stimulus sets contained quite a number of instances where
the Unity/Typical (++) product version was the lowest rated version (this was
true for eleven of the twenty-seven stimulus sets). In twenty of the stimulus
sets the Unity/Atypical () was the highest rated product version. One
possible explanation for this is that subjects may attribute the unusual or
novel appearance of the atypical product versions to expensive "designer

TABLE III 12
MEAN LIKING RATINGS AND UNITY SCORES
67
Product
(++)
(+-)
(-+)
(-)
UL
UN
1. Bathroom scale
Sh
6.31
5.92**a
5.06**
4.92**
0.12
0.83***
Tr
4.89
5.91
5.22
6.49
0.13+b
1 71***
Tx
4.47
6.44
5.25
5.28
-0.97
0.56+
2. TV Remote Control
Sh
5.37
4.58*
5.09
5.20
0.40*
1 42***
Tr
4.33
4.73
5.44
7.09
0.62**
2 74***
Tx
5.97
4.61**
5.00**
5.74
1.10**
1.68***
3. Flashlight
Sh
7.08
4.09***
3
3 4y***
1.53***
0.78**
Tr
6.16
3.95***
4.53***
4 26***
0.96***
0.77**
Tx
6.72
2.92***
2 40***
5.48**
3.36***
3.23***
4. Lamp
Sh
6.56
5.95*
4.34***
4.53***
0.37**
0.66**
Tr
5.08
4.80
6.02
6.25
0.26*
1.61**
Tx
5.36
5.30
4.98
6.73
0.91**
2.58***
5. Refrigerator
Sh
6.97
2.06***
2.46***
3.98***
3.25***
2 14***
Tr
6.13
4.14***
4 20***
6.05
1.96***
2.55***
Tx
6.89
2.31***
2.37***
3.40***
2.72***
1.36***
6. Telephone
Sh
5.98
3.03***
4 20***
3.63***
1 19***
0.92**
Tr
6.02
5.65+
4 29***
4.40***
0.26*
0.44**
Tx
6.05
5.26**
4.97**
6.42
1 io***
1 76***

TABLE III 12 continued
MEAN LIKING RATINGS AND UNITY SCORES
68
Product
(++)
(+-)
(-+)
()
ul
uN
7. Hair dryer
Sh
7.40
5.45
3.11
3.50
1.09***
0.71**
Tr
5.22
5.37
5.14
5.74
0.24+
1 24***
Tx
6.97
3.06
2.80
5.61
3.31***
3.23***
8. Dresser
Sh
5.45
5.26
4.74
5.86
0.65**
1.85***
Tr
5.89
3.38
3.69
7.03
2.90***
3.85***
Tx
6.13
3.84
4.68
4.71
1.18**
1.56**
9. Clock
Sh
6.44
2.58
2.81
5.13
3.10***
2.75***
Tr
5.63
3.74
3.79
4.81
1.39***
1 45***
Tx
5.69
3.85
3.51
5.75
2.03***
2.38***
Total0
5.97
4.38
4.21
5.24
1.30***
1 73***
a
b
c
***
*
+
Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
Indicates unity score is significantly different from zero.
Statistical significance based on ANOVA results reported in the text.
Indicates differences significant at p <
Indicates differences significant at p <
.001 in one-tailed-tests.
.01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
Indicates differences significant at p < .1 in one-tailed-tests.

TABLE III 13
MEAN QUALITY RATINGS AND UNITY SCORES
69
Product
(++)
(+-)
(-+)
(-)
uT,
UN
1. Bathroom scale
Sh
5.28
5.03
5.25
5.39
0.16*b
0.92***
Tr
4.53
5.90
4.65
6.11
0.05
1.50***
Tx
4.35
5.81
5.02
5.65
-0.44
0.95***
2. TV Remote Control
Sh
5.24
5.26
5.34
5.37
-0.00
0.68**
Tr
4.03
5.18
5.48
6.66
-0.03
^ rjrj***
Tx
5.41
5.18
5.23
5.66
0.31+
0.77**
3. Flashlight
Sh
5.61
4.95**a
4.90*
5.02*
0.42*
0.70**
Tr
5.07
4.70+
4.69*
5.00
0.34**
0.69**
Tx
5.42
4.26**
4.16**
5.68
1 24***
1.83***
4. Lamp
Sh
5.59
5.43
5.03*
5.08+
0.08
0.53**
Tr
4.39
4.84
5.53
5.82
-0.09
1.26***
Tx
4.23
5.39
4.87
6.45
0.19
1 91***
5. Refrigerator
Sh
6.02
3.97***
4 45***
5.03**
1.33***
1 24***
Tr
5.24
4.73*
4.58*
5.66
0.79***
445***
Tx
6.36
4.74***
4.85***
5.25**
1.00***
0.85***
6. Telephone
Sh
5.55
4.71**
4.85**
4.77**
0.36*
0.37*
Tr
5.23
5.34
4.79*
4.56**
-0.17
0.06
Tx
5.55
5.33
5.13*
5.89
0.53**
1

TABLE III 13 continued
MEAN QUALITY RATINGS AND UNITY SCORES
70
Product
(++)
(+-)
(-+)
(-)
UL
%
7. Hair dryer
Sh
6.39
5.39**
4 57***
4.59***
0.50**
0.48*
Tr
4.40
5.08
5.22
5.63
-0.18
0.82**
Tx
5.82
4.00***
4.03***
5.81
]_ 77***
2.05***
8. Dresser
Sh
5.05
5.29
5.26
5.95
0.22*
1.32***
Tr
4.81
4.34*
4.32*
6.52
1.29***
2.38***
Tx
5.18
4.70*
4.75
5.64
0.63**
9. Clock
Sh
5.13
3.53***
2 7l_***
4.84
1.32***
1.57***
Tr
4.64
4.10**
4.20*
4.61
0.43*
0.69**
Tx
4.87
3.98***
4.19**
4.74
0.73***
0.92***
Totalc
5.16
4.86***
4.78***
5.46
0.48***
02***
Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
Indicates unity score is significantly different from zero.
Statistical significance based on ANOVA results reported in the text.
Indicates differences significant at p < .001 in one-tailed-tests.
Indicates differences significant at p < .01 in one-tailed-tests.
Indicates differences significant at p < .05 in one-tailed-tests.
Indicates differences significant at p < .1 in one-tailed-tests.

TABLE III 14
MEAN PRICE RATINGS AND UNITY SCORES
71
Product
(++)
(+-)
(-+)
(-)
uT,
%
1. Bathroom scale
Sh
4.42
4.57
4.86
4.93
-0.04
1.02***
Tr
3.53
5.09
3.79
5.38
0.01
1.65***
Tx
3.66
5.74
4.38
5.36
-0.55
1 H***
2. TV Remote Control
Sh
4.38
5.22
5.08
5.56
-0.18
0.91***
Tr
3.80
4.89
4.84
6.56
-0.31
2.08***
Tx
4.62
4.92
4.85
5.55
0.21
1.03***
3. Flashlight
Sh
4.49
4.89
5.21
5.46
-0.10
0.91***
Tr
4.62
4.33+a
4.68
4.57
0.06
0.47*
Tx
4.34
4.14
4.17
5.09
0.55**b
4. Lamp
Sh
4.45
4.95
5.10
5.27
-0.13
0.90***
Tr
3.53
4.52
4.95
5.28
-0.32
1.05***
Tx
3.80
4.82
4.49
5.74
0.07
1 54***
5. Refrigerator
Sh
5.58
4.81*
4.66**
5.89
0.99**
1.48***
Tr
5.07
4.84
4.93
6.25
0.72***
1.69***
Tx
5.50
4.58**
4.87*
5.10
0.57**
0.77**
6. Telephone
Sh
5.26
5.11
5.03
5.84
0.44**
1 lg***
Tr
4.85
4.97
4.25**
4.33
0.03
0.28*
5.63
5.34
5.01**
6.14
0.69**
1.30***

TABLE III 14 continued
MEAN PRICE RATINGS AND UNITY SCORES
72
Product
(++)
(+-)
(-+)
()
UL
uN
7. Hair dryer
Sh
5.17
4.58**
3.92**
4.28**
0.45**
0.75**
Tr
3.91
4.67
4.72
5.58
0.05
1.15***
Tx
5.03
4.44*
4.21**
5.87
1.08***
1.82***
8. Dresser
Sh
4.30
4.67
5.39
5.98
0.11
1.62***
Tr
4.39
4.10
4.36
6.33
1.08***
2.25***
Tx
4.65
4.42
4.80
5.32
0.31*
1 14***
9. Clock
Sh
4.61
3.09***
3.31***
4.39
1 24***
1.51***
Tr
3.90
3.33**
3.52+
3.92
0.46**
0.71**
Tx
4.09
3.20**
3.44**
3.94
0.68**
0.88***
Totalc
4.50
4.60c
4.55
5.33
0.29***
1 19***
a Indicates that the difference for a change in a feature from prototypical
to atypical was significant based on the t-statistic computed for each
product/transformation pair.
Indicates unity score is significantly different from zero.
c Statistical significance based on ANOVA results reported in the text.
*** Indicates differences significant at p < .001 in one-tailed-tests.
** Indicates differences significant at p < .01 in one-tailed-tests.
* Indicates differences significant at p < .05 in one-tailed-tests.
+ Indicates differences significant at p < .1 in one-tailed-tests.

73
products" and thus relate novelty (atypicality) in general to the higher prices
associated with such products.
The analysis of variance results for liking, quality, and price are
reported in Table III-15 to Table IIT23. For all three of the derived responses
there was a significant effect of unity. In the case of attitude (i. e., liking
ratings) the effect of unity (i. e., Factor A x Factor B interaction) was highly
significant across all three variations of the dependent measure: rating scores
F (1,185) = 378.04, p < 0.0001; Linear unity scores F (1,185) = 378.04, p <
.0001; and pure-unity score F (1,188) = 530.00, p < .0001. That aesthetic
responses would have a strong influence on attitudes is not too surprising since
many conceptualizations of attitude maintain that attitudes contain an
affective component or that affective responses are antecedents of attitudes
(Ajzen 1989; Edwards 1990). The effect of unity on (perceived) quality was not
as strong as it was for liking but it was still significant except in the analysis
that utilized nonlinear unity scores where it was marginally significant: rating
scores F (1,179) = 83.54, p < .0001; linear unity scores F (1,179) = 83.54, p <
.0001; nonlinear unity scores F (1,188) = 3.74, p < .0545. The effect of unity
on ratings of price was significant across all the dependent measure variations:
rating scores F (1,179) = 57.67; linear unity scores F (1,179) = 57.67, p < .0001;
nonlinear unity scores F (1,188) = 398.57, p < .0001.
Once again, the patterns of interactions observed among Products,
Transformation type, and Order for these derived responses were the same as

74
TABLE III 15
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
46.0270
1.66
0.1937
Order
1
27.3070
1.96
0.1627
Version Order
2
124.4730
4.48
0.0126
Error Sub (Version Order)
185
2571.1570
Within Subiect Effects
Product
8
1223.6860
40.79
0.0001
Trans
2
166.7490
22.23
0.0001
Prod Version (Residual)
14
547.9980
10.44
0.0001
Prod Order
8
141.8833
4.73
0.0001
Trans Order
2
7.1283
0.95
0.2500
Prod Version Order
14
42.2361
0.80
0.2500
(Residual)
1480
5550.5546
Error Sub (Product)
Feature A
1
350.7149
48.87
0.0001
A Version
2
439.3120
30.61
0.001
A Order
1
21.8458
3.04
0.0827
A Version Order
2
14.4999
1.01
0.3662
Error Sub (A)
185
1327.7566
Feature B
1
136.4246
24.32
0.0001
B Version
2
154.4816
13.77
0.0001
B Order
1
36.0534
6.43
0.0121
B Version Order
2
25.8407
2.30
0.1028
Error Sub (B)
185
1037.9498
A B
1
2954.8430
378.04
0.0001
A B Version
2
95.1822
6.09
0.0027
A B Order
1
3.3147
0.42
0.5157
A B Version Order
2
8.7269
0.56
0.5732
Error Sub (A B)
185
1446.0027
Prod A B
8
1289.1620
57.08
0.0001
Prod A B Version
16
895.3401
19.82
0.0001
Prod A B Order
8
81.3623
3.60
0.0004
Prod A B Version *
16
74.3545
1.65
0.0509
Order
1480
4178.2312
Error Sub (Prod A B)

75
TABLE III 16
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
LINEAR UNITY SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subject Effects
Intercept
1
11819.3718
378.04
0.0001
Version
2
380.7288
6.09
0.0027
Order
1
13.2590
0.42
0.5157
Version Order
2
34.9076
0.56
0.5732
Error Sub (Version Order)
185
5784.0106
Within Subject Effects
Product
8
5156.6480
57.08
0.0001
Trans
2
537.0898
23.78
0.0001
Prod Version (Residual)
14
3044.2706
19.26
0.0001
Prod Order
8
325.4491
3.60
0.0004
Trans Order
2
32.1250
1.42
0.2500
Prod Version Order (Residual)
14
265.2929
1.67
0.1000
Error Sub (Prod)
1480
16712.9249
were discussed for aesthetic response. Individual t-tests for each of the
stimulus sets for all three of the derived responses were conducted to ensure
that the effect of unity was significant and in the proper direction across
stimulus sets (refer to Tables III-12, 13, and 14).
The three hypotheses concerning the influence of aesthetic response on
derived responses all seem to be supported. The products that received higher
beauty or aesthetic ratings (i. e., products exhibiting unity) also received higher
ratings for liking, quality, and price than did products that received lower
aesthetic ratings (i. e., products that did not exhibit unity). Although there
seems to be a reduction in the magnitude of the effect of the unity design
factor on the derived responses of perceived quality and expected price (as
compared with attitudes) the findings do seem to provide evidence that design

76
TABLE III 17
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
NONLINEAR UNITY DIFFERENCE SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subject Effects
Intercept
1
5141.7559
530.00
0.0001
Version
2
428.1894
22.07
0.0001
Order
1
12.1427
1.25
0.2647
Version Order
2
41.8216
2.16
0.1187
Error Sub (Version Order)
188
1823.8659
Within Subiect Effects
Product
8
352.3060
9.50
0.0001
Trans
2
143.3854
15.47
0.0001
Prod Version (Residual)
14
506.0161
7.80
0.0001
Prod Order
8
146.6168
3.95
0.0001
Trans Order
2
52.9295
5.71
0.0100
Prod Version Order (Residual)
14
100.1989
1.54
0.1000
Error Sub (Prod)
1504
6971.6544
factors may indirectly influence "non-aesthetic" product evaluations by means
of aesthetic response.

77
TABLE III 18
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
9.127
0.39
0.6782
Order
1
1.608
0.14
0.7116
Version Order
2
32.987
1.41
0.2477
Error Sub (Version Order)
179
2098.955
Within Subiect Effects
Product
8
460.3010
15.94
0.0001
Trans
2
8.4585
1.17
0.2500
Prod Version (Residual)
14
97.9144
1.94
0.0250
Prod Order
8
84.6168
2.93
0.0030
Trans Order
2
0.8922
0.12
0.2500
Prod Version Order (Residual)
14
102.4896
2.03
0.0250
Error Sub (Product)
1432
5170.5778
Feature A
1
19.6599
4.43
0.0368
A Version
2
111.1728
12.52
0.0001
A Order
1
7.9716
1.80
0.1820
A Version Order
2
0.4826
0.50
0.9471
Error Sub (A)
179
794.7515
Feature B
1
57.4549
15.22
0.0001
B Version
2
81.5176
10.80
0.0001
B Order
1
0.5777
0.15
0.6961
B Version Order
2
10.9047
1.44
0.2386
Error Sub (B)
179
675.6261
A B
1
401.5016
83.54
0.0001
A B Version
2
22.2117
2.31
0.1021
A B Order
1
0.0029
0.00
0.9804
A B Version Order
2
2.0149
0.21
0.8111
Error Sub (A B)
179
860.2664
Prod A B
8
233.0967
15.44
0.0001
Prod A B Version
16
245.1410
8.12
0.0001
Prod A B Order
8
25.6402
1.70
0.0943
Prod A B Version Order
16
25.6986
0.85
0.6271
Error Sub (Prod A B)
1432
2702.5612

78
TABLE III 19
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE
LINEAR UNITY SCORE
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
1606.0063
83.54
0.0001
Version
2
88.8466
2.31
0.1021
Order
1
0.0116
0.00
0.9804
Version Order
2
8.0595
0.21
0.8111
Error Sub (Version Order)
179
3441.0657
Within Subiect Effects
Product
8
932.3869
15.44
0.0001
Trans
2
172.8437
11.45
0.0001
Prod Version (Residual)
14
807.7205
7.64
0.0001
Prod Order
8
102.5606
1.70
0.0943
Trans Order
2
12.8424
0.85
0.2500
Prod Version Order (Residual)
14
89.9521
0.85
0.2500
Error Sub (Prod)
1432
10810.2448
TABLE III 20
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE
NONLINEAR UNITY SCORE
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
24.6230
3.79
0.0545
Version
2
243.5989
18.52
0.0001
Order
1
2.4103
0.37
0.5457
Version Order
2
3.1163
0.24
0.7893
Error Sub (Version Order)
188
1236.4512
Within Subiect Effects
Product
8
195.0827
7.91
0.0001
Trans
2
180.0369
29.21
0.0001
Prod Version (Residual)
14
270.5720
6.27
0.0001
Prod Order
8
67.3174
2.73
0.0054
Trans Order
2
5.9871
0.97
0.2500
Prod Version Order (Residual)
14
65.0598
1.51
0.1000
Error Sub (Prod)
1504
4635.4572

79
TABLE III 21
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE
RATING SCORE APPROACH
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
2
48.662
1.55
0.2149
Order
1
28.494
1.82
0.1795
Version Order
2
45.063
1.44
0.2405
Error Sub (Version Order)
179
2808.080
Within Subiect Effects
Product
8
1089.9719
38.16
0.0001
Trans
2
58.7083
8.22
0.0010
Prod Version (Residual)
14
137.2604
2.75
0.0001
Prod Order
8
182.0655
6.37
0.0001
Trans Order
2
8.0139
1.12
0.2500
Prod Version Order (Residual)
14
76.0911
1.52
0.1000
Error Sub (Product)
1432
5112.3283
Feature A
1
250.9383
56.16
0.0001
A Version
2
83.3609
9.33
0.0001
A Order
1
1.9605
0.44
0.5086
A Version Order
2
4.8154
0.54
0.5844
Error Sub (A)
179
799.8424
Feature B
1
321.3547
86.02
0.0001
B Version
2
130.0014
17.40
0.0001
B Order
1
2.3247
0.62
0.4312
B Version Order
2
11.9936
1.61
0.2037
Error Sub (B)
179
668.6915
A B
1
193.1558
57.67
0.0001
A B Version
2
26.2598
3.92
0.0216
A B Order
1
0.0024
0.00
0.9785
A B Version Order
2
4.2203
0.63
0.5337
Error Sub (A B)
179
599.4864
Prod A B
8
200.3073
15.86
0.0001
Prod A B Version
16
129.5848
5.13
0.0001
Prod A B Order
8
9.5617
0.76
0.6409
Prod A B Version Order
16
27.3044
1.08
0.3680
Error Sub (Prod A B)
1432
2260.7715

80
TABLE III 22
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE
LINEAR UNITY SCORE
Source
DF
Type III SS
F-Value
P > F
Between Subject Effects
Intercept
1
772.6232
57.67
0.0001
Version
2
105.0393
3.92
0.0216
Order
1
0.0098
0.00
0.9785
Version Order
2
16.8810
0.63
0.5337
Error Sub (Version Order)
179
2397.9457
Within Subiect Effects
Product
8
801.2291
15.86
0.0001
Trans
2
24.3032
1.92
0.2500
Prod Version (Residual)
14
494.0359
5.59
0.0010
Prod Order
8
38.2466
0.76
0.6409
Trans Order
2
1.6577
0.13
0.2500
Prod Version Order (Residual)
14
107.5598
1.22
0.2500
Error Sub (Prod)
1432
9043.0861
TABLE III 23
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE
NONLINEAR UNITY SCORE
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Intercept
1
2557.21486
398.57
0.0001
Version
2
188.1044
14.66
0.0001
Order
1
7.6367
1.19
0.2767
Version Order
2
0.9127
0.07
0.9314
Error Sub (Version Order)
188
1206.1973
Within Subiect Effects
Product
8
89.9727
4.40
0.0001
Trans
2
3.6682
0.71
0.2500
Prod Version (Residual)
14
86.8686
2.43
0.0100
Prod Order
8
39.4494
1.93
0.0519
Trans Order
2
8.2444
1.61
0.2500
Prod Version Order (Residual)
14
47.5262
1.33
0.2500
Error Sub (Prod)
1504
3842.0345

81
Summary
Overall, the results strongly support the hypothesis of a positive effect
of unity on aesthetic responses (i. e., HI). The Factor A and Factor B
interaction (i. e., unity main effect) was significant and this was shown to be
the case across product categories and across three different types of product
transformations. An effect of prototypicality (H2) on aesthetic response was
indicated, but there was not support for the moderate schema incongruity
hypotheses (H3 and H4).
The favorable aesthetic responses generated by consistency with the
unity visual organization principle was shown to carry over into subjects
attitudes, perceptions of quality, and price expectations for products (i. e., H7,
H8, and H9). The Feature A x Feature B interaction was significant for all
three derived responses. In addition, unity was shown to influence subjects
feelings of familiarity (i. e., perceived familiarity). Although not predicted, a
novelty effect on price expectations seemed to be indicated by the results (H5).
These results are summarized in Figure III-4. In Figure III-4 it can be seen
that the Unity/Atypical product () is always rated higher than the Non
unity/Atypical products [Note: MAX refers to the more highly rated of the two
Non-unity/Atypical products and MIN refers to the Non-unity/Atypical product
that received the lowest ratings].
These results clearly suggest that the unity design factor can
significantly influence both aesthetic responses and non-aesthetic product
evaluations and thus play an important role in determining consumer
behavior. The replication across the nine product categories demonstrates that
the effects observed are quite robust. Furthermore, the ability to produce

82
these effects using three very different types of (product) transformations
underscores the fact that unity can (and does) have broad application as a
design factor.

Response
83

Beauty

Familiarity

Liking
o
Quality
*
Price
Unity/Prototypicality
FIGURE III-4
MEAN AESTHETIC AND DERIVED RESPONSES

CHAPTER IV
EXPERIMENT 2
Overview
Experiment 1 investigated the effects of unity and prototypicality on
aesthetic and derived responses by presenting subjects with all four versions
of a stimulus set simultaneously and asking them to rate the versions on
semantic differential scales that measured the constructs of interest. This
experiment is a between-subjects replication of Experiment 1 in which subjects
evaluate only one design from each stimulus set. In this study, product
versions were evaluated without the context of other product versions from the
same product category. The between-subjects design of this experiment also
serves to reduce the possible demand artifacts that maybe involved in a within-
subjects experimental design that tests the influence of unity and
prototypicality on aesthetic responses, attitudes, quality perceptions, and price
expectations.
Experimental Design
A sub-set of the stimuli that were employed in Experiment 1 were used
in this experiment. One transformation was chosen from each product
category. The chosen transformation was the one with the strongest linear
84

85
unity score effect in Experiment 1. Although all nine product categories were
included, the bathroom scales stimulus set was omitted from subsequent
analyses for two reasons. First, this simplified the analyses of the results.
Second, this stimulus set was the only reversal of the unity effect observed in
Experiment 1. Each product category was represented and an effort was made
to select an equal number of each of the three types of manipulations (i. e.,
shape, texture, and trim). The nine stimulus sets that were employed were the
bathroom scale/textureT. V. Remote Control/trim, flashlight/trim, lamp/
texture, reffigerator/shape, telephone/shape, hair dryer/texture, dresser/trim,
and clock/shape.
The overall design of the experiment was a 2(Feature A) x 2(Feature B)
x 8(Products) mixed factorial design. Just as in the first experiment the
manipulation of Features A and B resulted in the manipulation of
prototypicality and unity. In this experiment, however, each cell of the crossed
Feature A and B factors [i. e., product versions: Unity/Prototypical (++), Non-
Unity/Atypical (+-), Non-Unity/Atypical (-+), Unity/Atypical (--)] was contained
in one of four questionnaire versions. That is, each questionnaire contained
all eight product categories (and the extra product); each subject received one
11
This product had proven to be problematic in Experiment 1 but was
included for the purpose of maintaining continuity between the product
presentations of the two experiments.

86
of four questionnaire versions and each questionnaire version contained one
cell (i. e., product variation, e. g., Unity/Prototypical) of the 2(Feature A)x
2(Feature B) manipulation from each of the product categories (see Appendix
D for a diagram of this design). The product versions (e. g., Unity/Prototypical)
were distributed evenly throughout the four versions of the questionnaire.
Although the three types of transformations were represented across the eight
stimulus sets, transformation type was not treated as a factor in this design.
The order in which the stimulus products were presented was reversed for half
of the subjects.
Experimental Procedure
Two-hundred and forty volunteer subjects enrolled in the Introductory
Marketing course at the University of Florida participated in this experiment.
The procedure for this experiment was the same as the procedure for
Experiment 1 except that in each of the four questionnaire versions subjects
saw and rated only one version of each product for each of the eight stimulus
sets. The dependent measures were the same measures that were used in
Experiment 1.

87
Results and Discussion
The manipulation check for the effects of a change in either Feature A
or Feature B on aesthetic responses indicated that, as expected, there was an
overall reduction in the ratings of product appearance as compared with the
prototypical product version. The average effect of a change in either of these
features was weaker than they were in Experiment 1 with the average effect
of a change in Feature A being -.425 and Feature B -.281. The effect of a
change in Feature A or Feature B from the category prototype for all eight
stimulus sets are shown in Table IV-1. In four of the eight cases the change
TABLE IV 1
MANIPULATION CHECK
Product
Feature A
Feature B
1. T.V. Remote Tr
- 0.001
0.343
2. Flashlight Tr
0.096
- 0.059
3. Lamp Tx
- 0.862
- 0.414
4. Refrigerator Sh
- 1.352
- 0.827
5. Telephone Sh
- 0.687
- 0.327
6. Hair dryer Tx
- 0.488
- 0.798
7. Dresser Tr
- 0.171
0.019
8. Clock Sh
0.064
- 0.182
Mean effect of change across
products
-0.425
- 0.281

88
of either Feature A or Feature B resulted in a lower beauty rating. The effects
of a change in the product features were mixed for the remaining four stimulus
sets. Of these four mixed-effects stimulus sets only one (T. V. remote trim)
is consistent with the manipulation check results of Experiment 1. The fact
that in three instances feature manipulations that had resulted in a reduction
of beauty ratings in Experiment 1 resulted in slightly increased ratings in
Experiment 2 would seem to indicate that the strength of the manipulations
is, not surprisingly, affected by the context of other product variations (i. e.,
whether or not the products being rated can be compared to other product
variations or brands).
The Influence of Unity and Prototypicality on Aesthetic and Derived Responses
As was the case in Experiment 1, the critical tests concerning the
influence of unity and prototypicality on aesthetic and derived responses
involves the Feature A x Feature B interaction (i. e., a main effect of unity).
A significant Feature A x Feature B interaction across products would indicate
that changes of product features that are made in a way that increases unity
[i. e., Unity/Atypical (--)] lead to more positive aesthetic (and derived)
responses. Such a result is inconsistent with the prototypicality explanation
of aesthetic response which predicts that product versions that are more
similar to the category prototype will receive more positive ratings.

89
Analysis
The data were analyzed using a mixed ANOVA with between-subjects
factors of questionnaire Version and Order. Products was a within-subjects
factor. The analysis was performed on product ratings since the between-
subjects design precluded the use of the unity score approaches at the
individual level. Unity scores (i. e., linear and nonlinear scores) may be
calculated using the mean product ratings at the aggregate level. However, a
nonlinear unity score calculated in this way will tend to understate the effect
of unity since the minimum case (i. e., the lowest rated Non-unity/Atypical
product variation) may not be the same for each subject. Thus, the minimum
Non-unity/Atypical mean will not necessarily be an aggregation of subjects
nonlinear unity scores. This means that the nonlinear unity scores for this
between-subjects experiment represent a more conservative test of the
influence of unity than the nonlinear unity scores in Experiment 1. The more
conservative nonlinear unity scores reported in this experiment will be
distinguished from those reported earlier by a prime () mark.
Hypothesis Testing
The means for the beauty ratings for each of the eight stimulus sets are
presented in Table IV-2. The analysis of variance for factors affecting beauty
ratings is presented in Table IV-3. The analysis shows that overall the
interaction between Features A and B was significant, F (1,1610) = 6.4238,

90
TABLE IV 2
MEANS FOR BEAUTY RATINGS
Product
Unity/
Typical
(++)
Non-Unity/
Atypical
(+-)
Non-Unity/
Atypical
(-+)
Unity/
Atypical
(-)
uT,
Un
1. T. V. Remote Tr
4.36
4.36
4.70
3.88
-0.41
-0.48
2. Flashlight Tr
4.85
4.95
4.79
5.07
0.09
0.28
3. Lamp Tx
5.00
4 14**a
4.59**
4.98
0.63**
b
0.84
4. Refrigerator Sh
4.16
2.80***
3.33*
3.41
0.72**
0.61
5. Telephone Sh
3.97
3.28
3.64
3.62
0.35
0.34
6. Hair dryer Tx
5.23
4.74*
4.43
4.23
0.15
-0.20
7. Dresser Tr
4.03
3.86
4.05
4.12
0.12
0.26
8. Clock Sh
3.03
3.10
2.85
2.71
-0.11
-0.14
Total
4.33
3.90***
4.05**
4.00**
0.19*
0.19
~~ Indicates significance from Prototypical product.
Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.
p < .025. The linear unity scores formed from the mean product ratings
indicated that six of the eight stimulus sets were in the expected direction (i.
e., positive). The "mean" nonlinear unity scores indicated that five of the eight
stimulus sets were in the expected direction. Follow-up tests for each product
indicated that individually only two products were significant [Lamp F(l,233)
= 8.85, p < .0032; Refrigerator F(l,233) = 7.74, p < .0058]. However, the fact

TABLE IV 3
ANALYSIS OF VARIANCE FOR BEAUTY DEPENDENT MEASURE
91
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
10.8722
0.66
0.5788
Order
1
37.9800
6.89
0.0092
Version Order
3
17.0346
1.03
0.3798
Error Sub (Version Order)
230
1267.2469
Within Subiect Effects
Product
7
814.4207
42.51
0.0001
Feature A
1
4.0309
1.47
0.2500
Feature B
1
26.6562
9.73
0.0100
A B
1
17.5812
6.42
0.0250
Prod Version
18
105.3644
2.13
0.0100
Prod Order
7
42.3589
2.24
0.0285
A Order
1
0.6343
0.23
0.2500
B Order
1
1.7413
0.64
0.2500
A B Order
1
2.7135
0.99
0.2500
Prod Version Order
18
51.2698
1.04
0.2500
Error Sub (Product)
1610
4406.5316
that this interaction was significant for the analysis of all the products
indicates that consistency with the unity visual organization principle is
positively related to aesthetic response.
The ANOVA for the ratings of product beauty also indicates that there
were significant effects of Order and Product as well as a significant Products
x Version interaction. These effects were addressed in the discussion of
Experiment 1 and the explanation of each of them is basically the same for this
experiment. The effect of Products reflects the differential strengths of the
Feature A and Feature B manipulations as they were adapted and applied to

92
each product category. An examination of the manipulation check for the
impact of changes in product features A and B on beauty ratings (refer to
Table IV-1) reveals that there are differences in the strength of the
manipulations across the eight products. This is also reflected in the inter
action between Products and Version (i. e., cells of the crossed feature A and
B "factors" of the design; e. g., Unity/Prototypical).
There were also significant order effects [Order F(l,230) = 6.89, p <
.0092; Product Order F(7,1610) = 2.24, p < .0285]. An examination of the
means for products for each of the presentation orders suggests that as in
Experiment 1, these effects may be attributable to subjects increasing
acceptance of the use of simple drawings to depict products.
Derived Responses
The influence of unity and prototypicality on attitude, quality
perceptions, and price expectations was also examined. The means for the
liking, quality, and price ratings are presented in Table IV-4, Table IV-5, and
Table IV-6, respectively. The patterns exhibited for the liking and quality
means are very similar to the pattern exhibited by the beauty means that were
presented in the preceding section. Interestingly, the pattern for the overall
means for price expectations were similar to those observed in Experiment 1
with the Unity/Atypical () product versions being the most highly rated and
the Unity/Prototypical (++) product versions being the lowest rated product.

93
TABLE IV 4
MEANS FOR LIKING RATINGS
Product
Unity/
Typical
(++)
Non-Unity/
Atypical
(+-)
Non-Unity/
Atypical
(-+)
Unity/
Atypical
(-)
UL
UN
1. T. V. Remote
5.38
5.56
5.81
4.88
-0.56
-0.68
2. Flashlight
5.63
5.76
5.60*
6.12
0.20
0.52
3. Lamp
5.87
4.97*a
5.16*
5.89
0.82**
b
0.92*
4. Refrigerator
5.07
3.65***
3.90**
3.89**
0.71*
0.24
5. Telephone
4.50
3.91
4.47
4.12
0.12
0.21
6. Hair dryer
6.36
5.21***
5.10
5.20
0.63**
0.10
7. Dresser
4.82
4.35
4.57
4.58
0.24
0.23
8. Clock
3.24
3.69
3.56
3.26
-0.38
-0.30
Total
5.11
4.63***
4.77**
4.74**
0.22***
0.16
a Indicates significance from Prototypical product.
b Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.
The ANOVAs for the influence of unity and prototypicality on liking, quality,
and price are presented in Table IV-7, Table IV-8, and Table IV-9, respectively.
The influence of unity (i. e., the Feature A x Feature B interaction) was
significant for liking F(l,1617) = 7.04, p < .001, approached significance for
quality F(l,1603) = 3.17, p < .10, and was not significant for price F (1,1603)
= .24 p > .25.

TABLE IV 5
MEANS FOR QUALITY RATINGS
94
Product
Unity/
Typical
C++)
Non-Unity/
Atypical
(+-)
Non-Unity/
Atypical
(-+)
Unity/
Atypical
(--)
UL
UN
1. T. V. Remote
5.26
5.28
5.50
5.00
-0.26
-0.28
2. Flashlight
5.13
5.29
5.24
5.13
-0.14
-0.11
3. Lamp
4.72
4.33*a
4.67
5.05
0.39+b
0.72
4. Refrigerator
4.76
4.45
4.09+
4.50
0.36
0.41
5. Telephone
4.72
4.31*
4.38*
5.09
0.56*
0.78
6. Hair dryer
5.85
5.47
5.38
5.15
0.08
-0.23
7. Dresser
4.30
4.03**
4.33
5.10
0.52*
1.07*
8. Clock
2.93
3.58
3.34
3.22
-0.39
-0.12
Total
4.71
4.59
4.62
4.78
0.14+
0.28
^ Indicates significance from Prototypical product.
Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.
The linear and nonlinear unity scores for liking both indicated that the
mean product ratings for six of the eight stimulus sets were in the expected
direction (i. e., positive). Five of the eight stimulus sets displayed positive
linear unity scores for quality ratings. Four of the nonlinear unity scores
formed from the mean product ratings were positive. Although the unity
scores for the price ratings indicated that five of the linear unity scores and six

TABLE IV 6
MEANS FOR PRICE RATINGS
95
Product
Unity/
Typical
(++)
Non-Unity/
Atypical
(+-)
Non-Unity/
Atypical
(-+)
Unity/
Atypical
(-)
uT.
Un
1. T. V. Remote
4.31
4.71
4.95
4.33
-0.51
-0.38
2. Flashlight
3.80
3.93
3.83
4.03
0.04
0.20
3. Lamp
3.51
3.59
3.72
4.03
0.12
0.44
4. Refrigerator
5.05
5.30
4.92
5.02
-0.08
0.10
5. Telephone
4.88
5.31
4.75
5.55
0.19
0.80
6. Hair dryer
4.57
4.67
4.29
4.64
0.13
0.35
7. Dresser
4.18
4.17
4.19
4.50
0.16
0.33
8. Clock
2.19
2.68
2.61
2.45
-0.33
-0.16
Total
4.06
4.29
4.16
4.32
-0.04
0.21
a Indicates significance from Prototypical product.
Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
Indicates that the atypical condition was rated more favorably than the
prototypical condition at the p < .05 level.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.
of the nonlinear unity scores were positive, these scores reflect the influence
of the high ratings received by the Unity/Atypical product variations. Across
the eight stimulus sets the Unity/Prototypical product variations tended to
receive price ratings that were lower than those received by the Non-
Unity/Atypical product variations. This pattern (i. e., novelty effect) is
consistent with the results reported for Experiment 1. Follow-up tests for each

TABLE IV 7
ANALYSIS OF VARIANCE FOR LIKING DEPENDENT MEASURE
96
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
34.3446
1.27
0.2844
Order
1
38.0624
4.23
0.0408
Version Order
3
95.6050
3.54
0.0154
Error Sub (Version Order)
231
2077.8006
Within Subiect Effects
Product
7
1162.2629
47.72
0.0001
Feature A
1
6.2858
1.81
0.2500
Feature B
1
30.7590
8.84
0.0100
A B
1
24.4887
7.04
0.0010
Prod Version
18
160.0874
2.56
0.2500
Prod Order
7
24.8918
1.02
0.4135
A Order
1
0.6051
0.17
0.2500
B Order
1
0.4279
0.12
0.2500
A B Order
1
14.6689
4.22
0.0500
Prod Version Order (Residual)
18
72.1990
1.15
0.2500
Error Sub (Product)
1617
5626.2035
product indicated that individually three products were significant for the
influence of unity on liking [Lamp F(l,234) = 9.26, p < .0026; Refrigerator
F(l,234) = 4.84 p< .0287; Hair dryer F(l,234) = 6.96, p < .0089, and T. V.
Remote was marginally significant F(l,234) = 3.57, p < .0602], two products
were significant for the influence of unity on quality [Telephone F( 1,231) =
5.40, p < .0210; Dresser F(l,232) = 4.36, p < .0378; and Lamp was marginally
significant F(l,232) = 3.20, p < .0747], and none of the products was significant
with respect to the influence of unity on price. Once again, significant effects

TABLE IV 8
ANALYSIS OF VARIANCE FOR QUALITY DEPENDENT MEASURE
97
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
31.9286
2.16
0.0933
Order
1
52.2380
10.61
0.0013
Version Order
3
26.4376
1.79
0.1498
Error Sub (Version Order)
229
1127.2916
Within Subiect Effects
Product
7
783.7063
37.16
0.0001
Feature A
1
1.0528
0.35
0.2500
Feature B
1
0.2176
0.07
0.2500
A B
1
9.5530
3.17
0.1000
Prod Version
18
83.3248
1.54
0.1000
Prod Order
7
21.0434
1.00
0.4310
A Order
1
2.7504
0.91
0.2500
B Order
1
7.1185
2.36
0.2500
A B Order
1
17.3136
5.75
0.0250
Prod Version Order (Residual)
18
50.2271
0.93
0.2500
Error Sub (Product)
1603
4829.7360
of Products and Order were observed as well as the significant interaction of
Products and Version.
Although the effect of unity on quality and price was not significant in
this case, these findings concerning derived responses are consistent with the
findings reported in Experiment 1. In Experiment 1 there also seemed to be
a reduction in the magnitude of the effect of the unity design factor on
perceived quality and expected price. This reduction in the effect, coupled with
the reduced influence of unity due to the removal of the context provided by

TABLE IV 9
ANALYSIS OF VARIANCE FOR PRICE DEPENDENT MEASURE
98
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
11.7739
0.77
0.5127
Order
1
11.2379
2.20
0.1393
Version Order
3
17.6058
1.15
0.3301
Error Sub (Version Order)
229
1169.4883
Within Subiect Effects
Product
7
1224.6204
73.91
0.0001
Feature A
1
1.6873
0.71
0.2500
Feature B
1
18.6191
7.87
0.0100
A B
1
0.5761
0.24
0.2500
Prod Version
18
43.3833
1.02
0.2500
Prod Order
7
64.1632
3.87
0.0003
A Order
1
0.0001
0.00
0.2500
B Order
1
0.1359
0.06
0.2500
A B Order
1
0.3289
0.14
0.2500
Prod Version Order (Residual)
18
49.1493
1.15
0.2500
Error Sub (Product)
1603
3794.1377
having all versions appear on the same page, could explain the lack of a
12
significant unity effect for quality and price. The alternative explanation,
that the absence of a comparative context increases the importance of
12
Partial omega squares (Keppel 1991, p. 223) were calculated in order to
estimate the magnitude of the unity effect. The partial omega squares
for the effect of unity on beauty ratings for Experiment 1 were .31 for
the rating score and U^ approaches and .37 for the score approach.
The partial omega square for the effect of unity on ratings of product
beauty for Experiment 2 was .006. Although these estimates indicate
that the effect of unity is quite large in Experiment 1 and very small in
Experiment 2, care should be taken in interpreting the importance of
these effects. See Keppel (1991, pp. 66-68 and 224) for discussions
concerning the importance of small effects and the difficulty of
comparing the sizes of omega squared estimates.

99
prototypicality, seems unlikely since, if that were true, both of the non
unity/atypical (+-, -+) conditions should always be rated higher than the unity
atypical (--) condition and that was not the case.
Perceived Familiarity
The relationship between unity and perceived familiarity was also
examined in this experiment. The means for the familiarity ratings for each
of the eight stimulus sets are presented in Table IV-10. The analysis of
variance for the influence of unity on perceived prototypicality is presented in
Table IV-11. In addition to the significant Product and Version (Residual)
effects that have been discussed in the previous two sections, there is a
significant interaction of Feature A and Feature B which suggests that unity
can encourage feelings of perceived familiarity, F (1,1610) = 28.76, p < .001.
The unity scores for the familiarity ratings indicated that five of the linear
scores and six of the nonlinear unity scores were in the hypothesized direction.
Follow-up tests conducted to examine the effect of unity on perceived famil
iarity for individual products indicated that the Refrigerator F(l,233 = 33.17,
p < .001), Telephone F(l,232) = 5.14,p < .0242), and Clock F(l,232 = 3.29, p <
.0709) were significant. Perceived familiarity may be particularly influenced
by product shape since all three of these stimulus sets utilized the shape
transformation. These results build on those of Experiment 1 to suggest that

TABLE IV 10
MEANS FOR FAMILIARITY RATINGS
100
Non-
Non-
Product
Unity/
Unity/
Unity/
Unity/
Typical
Atypical
Atypical
Atypical
(++)
(+-)
(-+)
(-)
UT,
Un
1. T. V. Remote
7.93
7.48
7.75
7.97
0.34
0.49
2. Flashlight
7.56
7.69
7.85
7.72
-0.13
0.03
3. Lamp
7.84
7.88
7.78
7.13
-0.35
-0.65
4. Refrigerator
8.45***
a
4 51***
5.77***
5.93***
2.05***
1.42**
5. Telephone
7.74
5.31
6.15
5.38***
0.83**
0.07**
6. Hair dryer
7.85
7.43
7.46
7.62
0.29
0.19
7. Dresser
7.54
7.21
7.27
6.82
-0.06
-0.39
8. Clock
8.09
7.57
7.70
8.07
0.45+
0.50
Total
7.87
6.88***
7.22***
7 08***
0.43***
0.21
a Indicates significance from Prototypical product.
b Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.

TABLE IV 11
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
101
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
121.7994
2.90
0.0358
Order
1
20.4949
1.46
0.2276
Version Order
3
24.4360
0.58
0.6276
Error Sub (Version Order)
230
3221.0752
Within Subiect Effects
Product
7
841.7213
39.87
0.0001
Feature A
1
25.0684
8.31
0.0100
Feature B
1
151.9619
50.39
0.0010
A B
1
86.7198
28.76
0.0010
Prod Version
18
388.4214
7.16
0.0010
Prod Order
7
8.7496
0.41
0.8938
A Order
1
4.3303
1.44
0.2500
B Order
1
7.4252
2.46
0.2500
A B Order
1
2.2846
0.76
0.2500
Prod Version Order (Residual)
18
71.5838
1.32
0.2500
Error Sub (Product)
1610
4885.5859
perceived familiarity involves more than the recall of the features shared with
the category prototype.
Summary
This experiment further examined the influence of unity and proto
typicality on aesthetic and derived responses. The experiment employed a
between-subjects design and thus the subjects evaluated the product versions
without the benefit of the three other versions (i.e., reference points) of the
stimulus set. The results of this experiment are summarized in Figure IV-1.

102
Even though rating product versions without the context of other versions
reduced the magnitude of the effects, the influence of unity on aesthetic
responses, attitudes, and perceived familiarity was still significant. The
finding that the influence of unity on quality and price is not as strong as it
is on the other responses is consistent with the findings of Experiment 1.
Clearly, context (i. e., availability to compare product versions with other
products in the same product category during the evaluation task) plays a role
in moderating the influence of factors such as unity, but the findings presented
here demonstrate that the unity effects are robust and that the influence of
unity on product evaluations may be pervasive.

Response
103
B
Beauty

Familiarity

Liking
O
Quality
*
Price
FIGURE IV-1
MEAN AESTHETIC AND DERIVED RESPONSES

CHAPTER V
EXPERIMENT 3
Overview
Experiment 2 examined the influence of unity and prototypicality on
aesthetic and derived responses when products were evaluated without the
context of other product versions from the same product category. The
experiment described in this chapter is a further investigation of the influence
of product aesthetics on derived responses. In this experiment, subjects are
provided with written product descriptions as well as drawings of the product
versions that they are asked to rate. The experiment examines whether or not
there is a moderating effect of the additional information (i. e., written product
descriptions) on the influence of unity. It is expected that as in the experiment
discussed earlier, unity will significantly influence subjects ratings of products
even though the evaluation task (i.e., rating product versions on scales
measuring either liking or quality) does not explicitly require consideration of
product appearance. In addition to demonstrating the influence of unity on
derived responses, such a finding would also provide evidence of the
importance of product aesthetics in general.
104

105
Stimulus Materials
The same sub-set of product drawings that were employed in
Experiment 2 were used in this experiment and thus the manipulations of
unity and prototypicality were the same as before. In this study, however,
each product version [i. e., "target" product, either Unity/Prototypical (++),
Unity/Atypical (--), or Non-unity/Atypical (+-)] was shown with a written
product description and was rated against a "control" product. The control
product was one of the Non-unity/Atypical product versions with a very
favorable product description (the Non-unity/Atypical condition for each
stimulus set that received higher beauty ratings was selected for use as the
control product). The product descriptions developed for the target products
were less favorable than the product descriptions for the control products.
"Strong" (i. e., favorable) product descriptions dominated "weak" (i. e., less
favorable) product descriptions on two of five or six attributes. The product
descriptions, which were developed from catalog descriptions of products, were
pilot tested (without the product drawings) in order to establish that the
"strong" descriptions were, in fact, perceived as being stronger than the "weak"
product descriptions. The strong and weak product descriptions are presented
in Appendix E.

106
Experimental Design
The design of this experiment may be thought of as being a 2(product
description: weak vs. strong) x 3(product appearance: Unity/Prototypical, Non-
unity/Atypical, Unity/Atypical) x 9(Products). The comparison of interest,
however, involves the differences between the three product appearance
conditions with the weak product description. Since it is the differences in the
ratings received by the products that were rated against (i. e., along with) the
control product with its more favorable product description that are of interest,
the study is actually a Ktarget product with weak product description) x
3(product appearance: Unity/Prototypical, Non-unity/Atypical, Unity/Atypical)
x 9 (Products) mixed design. The basic design of this experiment is shown in
Figure V 1. Product appearance was a between-subjects factor on any given
product but was balanced across each of the three questionnaire versions that
were administered. Product was a within-subjects factor since each subject did
rate one of the three possible a product pairings (control target) from each
stimulus set. The dependent measure was either a scale measuring quality
(High Quality/Low Quality) or a scale measuring general attitude
(Like/Dislike). Subjects received only one of the two scales for all of the
products that they rated. (This was to insure that there was no carry-over of
one scale to the other). The position of the target and control products on each
13
Inclusion of the bathroom scale stimulus set simplified programming
for the statistical analysis of this experiment.

Control Stimulus
107
Target Stimulus
"Weak" Product Description
"Strong" Product Description
Unity/Prototypical
Non-Unity/Atypical
(++)
(-+)
(Control)
Unity/Atypical
Non-Unity/Atypical
(-)
(-+)
(Control)
Non-Uni ty/Atypical
Non-Unity/Atypical
(+-)
(-+)
(Control)
Design: KTarget product) x 3(Unity/Prototypical, Unity/Atypical,
Non-unity/Atypical) x 9(Products) Mixed Design.
FIGURE V 1
EXPERIMENT 3

108
page (left or right side) was reversed for half of the subjects and was balanced
across questionnaire versions. The order in which the stimulus products (i. e.,
sets of target and control versions) were presented was also reversed for half
of the subjects and this was confounded with the product position
manipulation. Thus, the experiment was a l(target product with weak product
description) x 3(Product appearance: Unity/Prototypical, Non-unity/Atypical,
Unity/Atypical) x 9(Product) x 2(position/order) design that was conducted for
two dependent measures (quality or liking).
Experimental Procedure
Two-hundred and fifty seven subjects participated in this experiment.
The procedure that was followed was very similar to the procedures employed
in Experiments 1 and 2 except that in this study subjects were presented with
two product versions (control and target) on a page and each version was
accompanied with a product description. All conditions of this design were run
concurrently.
Results and Discussion
It was predicted that the unity aspect of product appearance would
influence subjects ratings of products in the same way that it had in earlier
experiments despite the addition of written product information. If this were
the case, one would expect that the particular product drawings (i. e.,

109
appearance conditions: Unity/Prototypical, Non-unity/Atypical, Unity/Atypical)
that were coupled with the target products (i. e., the weak product
descriptions) would influence the ratings of the target products.^
Analysis
The design was analyzed using a mixed ANOVA to assess the effects of
Unity, Product, Order/Position, and their possible interactions. Questionnaire
Version and Order/Position were between-subjects factors. Subjects ratings
of the control products were averaged and used as a covariate in the analysis
in order to reduce the error due to individual rating differences across subjects.
Hypothesis Testing
The critical test for examining whether or not the unity aspect of
product appearance influences subjects attitudes toward products and
perceptions of product quality despite the presence of additional information
involves the differences between the three target conditions. These conditions
differ only with respect to the appearance of the product version
(Unity/Prototypical, Non-unity/Atypical, Unity/Atypical) that is coupled with
the weak product description. If the results are consistent with the findings
14 Coupling the appearance conditions with the weak product
descriptions avoids ceiling effects which might otherwise occur if the
appearance conditions had been coupled with the strong product
descriptions.

110
of the previous two experiments, that is, products that exhibit unity are more
highly rated than products that do not exhibit unity, then they would support
the conclusion that the effect of unity on product evaluations is quite robust.
The means for the liking ratings are presented in Table V 1 and the
ANOVA for the effect of unity on subjects attitudes toward the products is
presented in Table V 2. The analysis indicates that there is a significant
effect of Cell, F (2,960) = 4.52, p < .025, which refers to the three target
product versions (Unity/Prototypical, Non-unity/Atypical, or Unity/Atypical)
seen by the subjects in each condition (i. e., questionnaire version) saw. The
effect of unity is also present in the effect of Version, F(2,120) = 3.43, p <
.0357), since the product version (i. e., Unity/Prototypical, Non-unity/Atypical,
and Unity/Atypical) for each product category are contained in the three
questionnaire versions. Contrasts between the different cells (i. e., product
versions) of this design indicated that the difference between the mean for the
Unity/Atypical products (M = 6.06) and the mean for the Non-unity/Atypical
products (M = 5.67) was significant, F(l,120) = 4.79, p < .0306).^ Although
the nonlinear unity scores calculated from the mean product ratings indicated
that seven of the nine were in the expected (i. e., positive) direction, follow-up
tests on the individual products indicated that the contrast between the
15
The difference between the means for the Unity/Prototypical products
(M = 5.90) and the Unity/Atypical products (M = 6.06) was not
significant.

TABLE V 1
LIKING
111
Product
Unity/
Prototypical
Non-Unity/
Atypical
Unity/
Atypical
UN
1. Bathroom scale
5.52
5.86
6.40
0.54
2. T. V. Remote
6.30
6.02
6.40*a
0.38
3. Flashlight
5.57
5.49
5.98
0.49
4. Lamp
6.64
6.05
6.19
0.14
5. Refrigerator
5.37
4 36***a
6.55
2
6. Telephone
5.88
4.74**
5.40**
0.66
7. Hair dryer
5.79
7.00
5.93
-1.07
8. Dresser
6.14
5.29
6.00
0.71
9. Clock
5.93
6.23
5.71
-0.52
Total
5.90
5.67**
6.06
0.39*
a Indicates significance from Prototypical product.
D Indicates significance of the differences between Unity/Atypical and
Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.

TABLE V 2
ANALYSIS OF VARIANCE LIKING
112
Source
DF
Type III SS
F-Value
P > F
Between Subjects Effects
Intercept
1
181.0534
24.72
0.0001
Covariate
1
129.7023
17.71
0.0001
Version
2
50.1938
3.43
0.0357
Order and Position
1
5.0247
0.69
0.4092
Version Order
2
18.5099
1.26
0.2864
Error
120
878.9758
Cell
2
18.6838
4.52
0.0250
Prod Version
14
156.2446
5.39
0.0100
Error (Product)
960
1984.9216
Cell Order
2
0.1300
0.03
0.2500
Prod Version Order
14
27.1226
0.94
0.2500
Error (Product)
960
1984.9216
Unity/Atypical and Non-unity/Atypical product versions was significant only
for the refrigerator product F (1,121 = 28.41, p < .0001). These results indicate
that product appearance, and particularly unity, does influence subjects
attitudes toward products even when written attribute information (i. e.,
technical information) is available concerning the products.
The means for the quality ratings are presented in Table V-3 and the
ANOVA for the effect of unity on subjects perceptions of quality is presented
in Table V-4. Although the pattern exhibited by the means is consistent with
the prediction that the unity aspect of product appearance influences perceived
quality, the effect of unity was not significant across products F (2,984 = .22,
p > .25), nor was the contrast between the mean for Unity/Atypical products

113
TABLE V 3
QUALITY
Product
Unity/
typical
Non-Unity/
Atypical
Unity/
Atypical
Un
1. Bathroom scale
5.24
5.58
6.68
1.10
2. T. V. Remote
6.44
6.00
6.21
0.21
3. Flashlight
5.30
5.46
5.98
0.52
4. Lamp
6.91
5.19
5.83*a
0.64
5. Refrigerator
5.83
5 70***a
5.97
0.27**b
6. Telephone
6.05
5.44*
5.65***
0.21
7. Hair dryer
6.11
6.77*
5.44**
-1.33
8. Dresser
6.29
5.67
6.09
0.42
9. Clock
5.49
6.37
6.00
-0.37
Total
5.96
5.79
5.98
0.19
a Indicates significance from Prototypical product.
b Indicates significance of the differences between Unity/Atypical
and Non-unity/Atypical conditions.
*** Indicates differences significant at p < .001.
** Indicates differences significant at p < .01.
* Indicates differences significant at p < .05.
+ Indicates differences significant at p < .10.

114
TABLE V 4
ANALYSIS OF VARIANCE QUALITY
Source
DF
Type III SS
F-Value
P > F
Between Subjects Effects
Intercept
1
2.5064
0.68
0.4101
Covariate
1
474.7499
129.42
0.0001
Version
2
21.0677
2.87
0.0604
Order and Position
1
4.7474
1.29
0.2575
Version Order
2
1.1033
0.15
0.8605
Error
123
451.1998
Cell
2
0.6545
0.22
0.2500
Prod Version
14
200.7009
9.85
0.0010
Error (Product)
984
1431.8646
Cell Order
2
0.3782
0.13
0.2500
Prod Version Order
14
47.7123
2.34
0.1000
Error (Product)
984
1431.8646
(M = 5.98) and the mean for Non-unity/Atypical products (M = 5.79) significant
F (1,123) = .24, p > .25). Seven of the nine nonlinear unity scores were in the
expected direction. Follow-up tests on the individual products indicated that
the contrast between the Unity/Atypical and Non-unity/Atypical product
versions was only significant for the refrigerator product F (1,123 = 6.84, p <
.01).
Summary
It was predicted that the unity aspect of product appearance would
influence subjects attitudes towards products and their perceptions of product
quality despite the availability of written attribute information. The results

115
presented here indicate that while this certainly does seem to be the case for
attitudes toward products, the additional information seems to moderate the
influence of product aesthetics (i. e., unity) on perceptions of product quality.
As was pointed out in Chapter III, attitudes seem to be more closely linked (or
more susceptible) to aesthetic responses than other derived responses such as
quality. It may be that the more general nature of attitudes results in their
being more inclusive and thus allows (consciously or unconsciously) aesthetic
information to be weighted more heavily than it is in the case of derived
responses that are more focused on specific product attributes.

CHAPTER VI
GENERAL DISCUSSION
The results of these three studies provide strong evidence of the
influence of unity on aesthetic responses. Unity was also shown to
significantly influence attitudes toward products, perceived familiarity, and in
some situations perceptions of product quality. The context of the evaluation
situation (i. e., whether or not other product variations or written product
descriptions were presented for comparison along with the product being
evaluated) was shown to moderate the influence of unity (i. e., aesthetic
response) on perceived quality. In addition to these insights concerning the
influence of unity on aesthetic and derived responses, these studies help to
clarify the role of prototypicality in influencing aesthetic and derived
responses. Although prototypicality did seem to have a positive effect on
aesthetic responses and attitudes toward products, it did not seem to have a
positive effect on perceived quality or price expectations. In fact,
prototypicality seemed to have a negative effect on price expectations.
In Experiment 1 a products consistency with the unity visual
organization principle was shown to significantly influence subjects aesthetic
responses toward products. The favorable aesthetic response generated by a
116

117
products consistency with the unity visual organization principle carried over
into subjects attitudes toward products, perceptions of product quality, and
product price expectations. In addition, unity was shown to influence subjects
perceptions of product familiarity. The effect of unity was quite robust as
these results were replicated across nine product categories using three
different types of product transformation. The results of Experiment 1 provide
a clear indication that the unity visual organization principle is an important
design factor that can significantly influence aesthetic responses and non-
aesthetic product evaluations.
Experiment 2 found that unity significantly influenced aesthetic
responses even when products were evaluated without the context of other
product versions from the same product category. The favorable aesthetic
response generated by consistency with the unity visual organization principle
was also shown to carry over to attitudes toward products and perceived
familiarity. Although the overall pattern of the means for the quality ratings
of the products was positive for the majority of the stimulus sets tested, it was
not significant. This experiment, which provided a between-subjects test of the
influence of the unity and prototypicality factors, indicates that just as in
Experiment 1 unity provides a better explanation of aesthetic response than
does prototypicality. This between-subjects replication further demonstrates
that the effects of unity are quite robust.

118
The "comparative context" (i. e., different product versions rated
together) of Experiment 1 was similar to shopping situations where competitive
product offerings in the same category are displayed together (e. g., electronics
stores). This is also the case for products displayed in catalogs that feature
goods made by more than one manufacturer. The majority of products seem
to be displayed along with competitive product offerings and therefore are
evaluated in comparative contexts. In this sense Experiment 1 would seem to
be ecologically valid. However, most advertising and some products are
evaluated under conditions similar to the "non-comparative" context employed
in Experiment 2. The evaluative context in Experiment 2 was "non
comparative" in that product versions were evaluated singularly, without the
benefit of other similar product versions as reference points. This is the case
for most advertising. It is also the case for products such as automobiles,
boats, furniture, etc. that often involve exclusive distribution networks. The
evaluation contexts for these products is usually one where the consumer is not
able to directly (i. e., side by side) compare competitive product offerings. With
respect to these purchase situations Experiment 2 would seem to be
ecologically valid.
The third experiment, which focused on derived responses, found that
the unity aspect of product appearance influenced subjects attitudes toward
products despite the availability of information contained in product
descriptions. This study involved a context that was similar to many everyday

119
shopping situations in that the products being evaluated were presented along
with descriptive information. Even though subjects were allowed to evaluate
the products in whatever manner that they wished, the aesthetic information
(i. e., product appearance) clearly played a role in influencing their attitudes
toward the products. The descriptive information did, however, seem to
supersede aesthetic information with respect to perceived quality.
Across the three experiments the effects of unity were quite robust
although they were more pronounced in the experiment that was conducted
within-subjects than they were in the between-subjects experiments. This is
not surprising since within-subjects designs afford more statistical power than
do between-subjects designs. The fact that these effects were indicated despite
the use of rather primitive line drawings suggests that the effects may be
stronger in the real-world since a "greater range of perceptual salience can be
achieved" through the use of color, materials, and three dimensions
(Hutchinson and Alba 1991, p. 342). It is important to note that the effects of
unity on derived responses that were observed in Experiments 1 and 2 could
have been influenced by the procedure that was employed (i.e., having subjects
rate the products on all five of the dependent measures). The fact that
the patterns exhibited by each of the dependent variables were not the same
(e. g., price) coupled with the fact that the effect of unity was observed in the
16
The direction of two of the rating scales was reversed in order to
reduce response bias on the part of subjects.

120
third experiment in which subjects rated products on only one of the dependent
measures would seem to suggest that the results are not merely an artifact of
the procedure employed.
The findings concerning the effect of evaluation context on the influence
of product aesthetics has some interesting managerial implications which are
especially relevant to the design and marketing of new products. Experiment
3 seems to indicate that the context in which product evaluations occur (e. g.,
presence of competitive products, descriptive product information) can
influence how heavily consumers weight aesthetic information (i. e., product
appearance) in their product evaluations. This suggests that there may be a
trade-off between the influence of descriptive information and product
appearance in the promotion or merchandising of a product. If, for example,
the addition of certain descriptive product information causes consumers to
focus less on a very well designed products appearance, then the net effect of
adding the information may be a reduction in the total "impact" of the
promotional effort. Depending on the goals (i. e., desired impact; e. g.,
increased sales, awareness, quality perceptions, etc.) of the promotion this
could seriously affect the results of the campaign.
Experiments 1 and 2 suggest that context (i. e., the opportunity to
compare a product version with other products in the same product category
during the evaluation phase of the buying process) plays a role in mediating
the influence of the aesthetic aspects of products. This suggests that managers

121
should pay particular attention to product design for products that will be
evaluated in close proximity (either physically or in photographs) to
competitors products since the influence of aesthetic response is likely to play
a role in affecting key product perceptions (e. g., beauty, quality). This could
also be true for instances where consumers collect product brochures in order
to "directly" compare products (e.g., automobile brochures). The findings of
these studies also seem to suggest that in cases where a product is known to
have a poor appearance an effort should be made to display it apart from other
products that have more appealing designs. Such action would reduce the
17
salience of the aesthetic aspects of the product.
One interesting finding that would seem to have implications for the
designing of new products concerns the relationship between novelty and price.
The novel (i. e., atypical) products in Experiments 1 and 2 were associated
with higher prices regardless of the level of unity that they exhibited. The
association of higher priced products with the prototypicality of a products
appearance and the association of quality with the unity aspect of a products
design would seem to suggest that product appearance (i. e., design) can be
used to position a product with respect to (seemingly) "non-aesthetic"
dimensions (i. e., price, quality). The implication is that design could be used
17
This course of action is offered with great reservation. The best
long-term course of action would be to redesign the product.
However, in cases where a basically sound product has failed to
achieve commercial success due to an initially poor design such a
course of action may be warranted.

122
by a company to create an entire "line" of products that target different
consumer segments at different price points, levels of quality, etc., while
utilizing the same internal components. While this would be an extreme use
(i. e., exploitation) of product design, the more likely and desirable use is to
employ design to better communicate the nature (i. e., positioning and proper
use) of a product to consumers.
An important issue to be investigated further is the relationship between
unity and prototypicality. The findings of Experiments 1 and 2 indicated that
consistency with the unity visual organization principle is positively related to
perceived familiarity. One possible explanation for this is that there may be
a natural tendency in people toward organization principles such as unity and
symmetry. Whether the tendency of the viewer to look for organization and
prefer it to chaos is innate or learned is still the subject of debate. The debate
between biological and cultural determinism in aesthetics is but a small part
of a larger ongoing debate on the relative roles of "nature" and "nurture" in
human behavior. In all likelihood, the tendency to perceive visual elements in
certain ways or as integrated wholes involves both biological and cultural
influences. The natural tendency toward organization with respect to
principles such as unity and symmetry may stem (at least in part) from our
earliest encounters with ourselves and our environments (Johnson 1987). Our
bodies (i. e., forms) and those of the animals and plants that we encounter
exhibit the "regularities" that we have come to know as unity, symmetry,

123
balance, proportion, etc. Experience with these regularities would most
certainly influence subsequent perception and over time become internalized.
In this way people may develop a level of familiarity with and preference for
certain relations (i. e, regularities; e. g., unity, symmetry). Goldstone, Medin,
and Gentner (1991) have investigated how relational similarities (i. e.,
descriptions of connections between two or more objects or attributes; e. g.,
same color) affect similarity judgments. Their work shows that relations and
attributes are psychologically distinct and that relations can significantly affect
similarity judgments. Thus, relations (i. e., visual organization principles)
which are learned (consciously or nonconsciously) would seem to have the
potential to influence judgments of a products prototypicality.
There is also a need for further investigation into the systematic nature
of aesthetic response. This work as well as others (e. g., Veryzer 1993)
indicates that the systematic nature of aesthetic response in the visual domain
stems from the perceptual tendency toward organization. This tendency has
been studied by the Gestalt psychologists and aestheticians and is the basis for
general rules of perception such as the Gestalt laws (e. g., proximity,
similarity) and design principles (e. g., unity, contrast, proportion). While the
source of the visual organization principles (i. e., design principles, Gestalt
laws) that operate to organize perception is still open to debate, there is
evidence that the perception principles are present very early in life and that
preferences related to these principles develop over time and may be modified

124
or influenced by cultural forces (Bornstein, Ferdinandsen, and Gross 1981;
Segall 1976). Although the acquired principles are undoubtably modified from
time to time, aesthetic response preference patterns are fairly stable within
individuals over time (Huber and Holbrook 1981). Similarities and differences
observed between individuals may be due to similarities or differences in the
physical, socio-economic, or cultural environment in which people live.
Although on the surface it would seem that aesthetic response operates
on a conscious level involving rational evaluation, it has been suggested that
the intervening cognitive response system may operate below the threshold of
consciousness at a subconscious or pre-conscious level (Holbrook and
Hirschman 1982). Aesthetic response seems to operate on at least two levels -
conscious awareness and nonconscious awareness (Holbrook and Hirschman
1982). The conscious level involves attending to the object and registering
feelings or appreciation of the object. This level is the conscious registering of
the unconscious input and is therefore primarily a function of the nonconscious
awareness level (Zajonc 1980). The nonconscious level of awareness involves
perceiving the object and determining its consistency with rules (e. g., design
principles) which have been acquired. Regardless of whether these rules are
innate or learned nonconsciously, they often seem to be applied without
conscious awareness. Thus, while differences in the appearance of products
may be readily perceived (i. e., conscious awareness), the underlying process

125
by which the differences are transformed into an aesthetic response often
seems to occur nonconsciously.
The relationship between design principles and aesthetic response seems
to develop, at least in part, through learning. While this learning process may
be a formal one as in the case of educational programs for art and architecture,
it often seems to be the result of an informal and nonconscious process. Thus,
the phenomenon of aesthetic response may involve the nonconscious
development of design principle internal processing algorithms (design
principle IPAs) as well as the nonconscious application of these design
principle algorithms. Objects (products) that are consistent with a persons
design principle IPAs would be expected to produce more positive affect than
objects that violate a persons relevant design principle IPAs (Veryzer 1993).
Another area that merits further investigation is the tendency of people
to underweight the influence of product aesthetics on their evaluations of
products. Berkowitz (1987) found evidence of the misattribution of product
preferences by the perceivers of products. In a study that examined peoples
preferences for two food products, ratings on product attributes revealed that
subjects did not attribute the reasons for their choices to visual appeal or more
pleasing shape even though shape was the only actual difference between the
two types of products. The fact that consumers attributed their preferences to
other aspects such as "freshness" and "taste" even though shape was the only

126
difference between the two products raises questions about consumers level of
awareness concerning the influence of aesthetics on their product evaluations.
It is hoped that in addition to improving our understanding of the
influence of aesthetics on consumer behavior the research that has been
presented here will serve as a foundation for a theory of consumer aesthetics.
This work takes an important step toward increasing our knowledge of the
relationship between product aesthetics and consumer behavior. The goal of
this research has been to investigate product aesthetics in a manner that
yields the concrete principles governing peoples responses to product designs.
The theory and propositions concerning unity, prototypicality, and aesthetic
response that were presented here lay a foundation for theorizing about the
aesthetic aspects of product design. The results from the three studies indicate
that there are, indeed, factors that systematically and significantly influence
consumers aesthetic responses and product perceptions. These findings
provide empirical evidence that design is not simply a superficial, frivolous
concern, but rather that it is an important variable that can have a significant
impact on consumers responses to products. This should encourage
researchers and managers alike to pay more attention to design issues. It
should also help to more firmly establish product design as a legitimate
marketing interest.

APPENDIX A
PRODUCTS EXPLORED FOR USE AS STIMULI

APPENDIX A
PRODUCTS EXPLORED FOR USE AS STIMULI
The products employed in the experiment were drawn from a larger set
of stimuli that had been explored with respect to the proposed manipulations.
Suitable stimuli can be created for virtually any product by manipulating two
features of the product (e. g., shapes: circle/square; texture: smooth/rough;
trim: present/absent; angle: 90-degree/60-degree). A partial list of products
that have been explored with respect to the proposed manipulations includes
the following:
Alarm clocks
Ice chests
Silverware
Calculators
Irons
Staplers
Cameras
Kitchen timers
Stereos
Clock radios
Lamps
Sunglasses
Coffee pots
Make-up compacts
Telephones
Cologne bottles
Mirrors
Thermoses
Cordless phones
Pencil sharpener
Toasters
Cups
Pots/pans
Toothbrushes
Dressers
Refrigerator
Television remote controls
Electric fans
Scissors
Televisions
Flashlights
Sewing machines
Vacuum cleaners
Glue guns
Shirts
Watches
128

APPENDIX B
STIMULUS SETS

Please indicate vour reaction to the appearance of the
products by marking (circling) the scales below each.
proposed
Bathroom scales
Dislike
12 3 4
5 ei s
0
Like
Ugly 1
2 3 4
5 6 ~ i
9
Beautiful
Familiar 1
2 3 4
5 6'8
9
Unfamiliar
High Quality 1
2 3 4
5 6^8
9
Low Quality
Low Price 1
2 3 4
5 6 S
9
High Pnce
Dislike
Ugly
Familiar
High Quality
Low Pnce
3 4 5 6 7 8 0 Like
3 4 5 6 7 8 9 Beautiful
3 4 5 6 7 8 0 Unfamiliar
3 4 5 6 7 8 9 Low Quality
3 4 5 6 7 8 9 High Pnce
Dislike 1 2 3 4 5
Ugly 12 3 4 5
Familiar 12 3 4 5
High Qualiry 1 2 3 4 5
Low pnce 12 3-5
6 £ 9 Like
7 8 0 Beautiful
7 8 9 Unfamiliar
- 89 Low Quality
^ 0 High Pnce
Dislike
Ugh
Familiar
High Quajjr\
Low Pnce
6 1 s q Like
8 9 Beautiful
6 8 9 Unfamiliar
8 0 Low Quailr\
* > 9 High Pnce
130

Please
products
the
proposed
Bathroom scalp.:
131

Please indicate vour
products bv marking
C
reaction to the
(circling) the scales
appearance of the
below each.
proposed
fiathroorn_s£ales
132

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
TV Remote Controls
/ taacHi if
C2QQ ¡I
/ aaa //
ocjca f
m
fags)
/ sso j
/ aa a //
/ aa a if
C3E3C3 f
iaipcy /
/ /
V
/
Dislike 5 6 8 9 Like
L civ 1 2 5 4*6789 Beautiful
Famiiiar 12 3 456789 Unfamiliar
High Quality 12 3 456789 Low Qualitv
Low Pnce 12 3 456789 High Pnce
Dislike 1 2 3 4 5 6 7 8 4 Like
Ugly 1 234 r 6789 Beautiful
Familiar 12 3 4 5 6 7 8 9 Unfamiliar
High Quality 1 2 3 4 5 6 7 8 9 Low Quality
Low Pnce 1 2 3 4 5 6 7 8 9 High Pnce
SBSi|
/ C2C2Q i
/ aut a //
/ C3QQ //
/ mr-vtf S
/ /ao' //
/ L3LJC3 h
/ Lsaa j
/ C3CJL3 //
/ OOZ3 //
j jqS 1
Dislike 12 5 4J6789 Like
Ugly 1 2 3 4 5 6 7 8 9 Beautiful
Fanuhar 1 2 3 4 5 6 8 9 Unfamiliar
High Quality 12 3 456_S9 Low Qualm
Low Pnce 123456789 High Price
Dislike 1 234567s Like
Ugly 1 2 3 -1 5 6 7 8 Beautiful
Farruhar 2 3 4 ; 6 7 8 9 Unfamiliar
High Quality : 234 5 *780 Low Quality
Low Price 1 2 3 4 5 6 7 8 9 High Pnce
133

Please indicate vour reaction to the appearance of the proposed
products by marking (circling) the scales below each.
TV Remote Controls
Dislike
Ugly
Familiar
High Qualiry
Low Pnce
1 3 : 6 8 9 Like
1 3 d 6 8 9 Beautiful
1 2 j- -i r 6 8 9 U ni amiliar
1 - 6 7 8 9 Low Qualirv
Dislike
Ugly
Familiar
High Quality
12 3 4
6 7 8 9
High Pnce
Low Pnce
i 2 3 4 5 6 7 8 9 Like
1 3 4 5 6 7 8 9 Bcaumui
1 2 456789 Unfamiliar
*2 3 4 5 6 7 8 9 Low Quality
i 27456789 High Pnce
crac.?
B3&
>333
Disiixe
2 3 4 f6
~ s
0
Like
Dislike
2
3 4 5
r 7 8
9
Lute
U gly
2 3 4 5 o
7 8
9
Beautiful
Ugly
2
3 4 5
i 8
9
Beautiful
Familiar !
2 3 4 5 6
- 8
9
Unfamiliar
Familiar
2
3 4 5
ft 7 8
0
Unfamiliar
High Quality
2 3 4 5 6
8
9
Low Quality
High Quality
2
3 4 5
6 7 8
9
Low Quabr.
Low Price
2 3 4 5 6
7 8
9
High Price
Low Price
2
3 4 5
6 7 8
9
High Pnce
134

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
TV Remote
Controls
//aaa/A
nr-ir-t r, h\
/iMssl/l
jbF
Sill
f~~ y
Dislike 123456'89 Like
Ugly 1 23456789 Beautiful
Familiar 12 3 4567S9 Unfamiliar
High Quality 123456789 Low Quaiitv
Low Pnce 123456789 High Pnce
Dislike 1 2 3 4567S9 Like
Ugly 1 2 3 4 5 6 7 8 9 Beautiful
Familiar 1 23456789 Unfamiliar
High Quality 123456789 Low Quality
Low Pnce 1 2 3 4 5 6 7 8 9 High Pnce
7 1
!§§§!)
0s§//¡
¡38g¡)
iSasaUI
F j/
y
Dislike 12 3- 56789 Like
Ugly 12 3 -56789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 12 3 456-89 Low Quality
Low Pnce 12 3 456789 High Pnce
Dislike 1 2345~7sa Like
Ugly ; 23456789 Beautiful
Familiar 1 23456789 Unfamiliar
High Quality 1 2 3 4 5 6 7 S 9 Low Quailt>
Low Price 1 23456789 High Pnce
135

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Flashlights
Dislike
2 3
- 5
6 7
8
9
Like
Dislike
2
3
4
5
6
7
8
9
Like
Ugly
*> ^
- 5
6 7
8
9
Beautiful
Ugly
2
3
4
5
6
7
8
9
Beautiful
Familiar
2 3
4 5
6 7
8
9
Unfamiliar
Familiar ]
2
3
4
5
6
7
8
9
Unfamiliar
High Quality
- 5
6 7
8
9
Low Quality
High Quality
p
3
4
5
6
7
8
9
Low Quality
Low Price
2 3
4 -
6 7
S
9
High Price
Low Price
2
3
4
5
6
7
8
9
High Pnce
Dislike
2 '
- ;
6
7
8
9
Like
Dislike
2
3
4
5
&
7 8
Q
Like
Ugly
2 3
4 5
6
7
8
9
Beautiful
Ugly
2
3
4
5
6
7 8
9
Beautiful
Familiar ]
2 3
-
5
7
8
9
Unfamiliar
Familiar ]
n
3
4
5
6
" 8
9
Unfamiliar
High Quality
2 3
- 5
6
7
8
9
Low Quality
High Quality
2
3
4
5
6
7 8
9
Low Quality
Low Pnce
2 3
4 5
6
7
8
9
High Price
Low Price
2
3
4
5
6
7 8
9
High Price

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Flashlights
Dislike 2 3 -1 5 o 7 8 9 Like
Ugly 1 2 .' 456789 Beautiful
Familiar 1 23-5 6789 Unfamiliar
High Quality 1 23456789 Low Quality
Low Pnce 1 23456789 High Pnce
Dislike 1 2 5 4 5 6 7 5 Like
Ugly 1 2 3 4 5 6 7 8 3 Beautiful
Familiar 12 3 45678 Unfamiliar
High Quality 1 2 3 J 5 6 7 8 a Low Quality
Low Pnce 1 2 3 4 5 6 7 8 3 High Pnce
Dislike
2 3 4 ?67
s
9
Like
Dislike
6
7
y
~ Like
Ugly
2 3 4 5 6 7
8
9
Beautiful
Ugly
: ? -t 5
6
7
J
0 Beautiful
Familial
2 3 4 5 6 -
8
9
Unfamiliar
Familiar
2 3 4 5
6
7
S
0 Unfamiliar
High Quality
2 3 4 5 6 7
8
9
Low Quality
High Quality
2 3 4 5
6
7
8
0 Low Quaiir\
Low Pnce
2 3 4 5 6 7
8
9
High Price
Low Price
2 3 4 5
6
7
8
Q High Pnce
137

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Flashlights
Dislike 1
6 7
8
Q
Like
Ugly 1
2 3 -i 5
6 7
8
9
Beautiful
Familiar i
2 3 4 5
6 7
8
9
Unfamiliar
High Quality 1
2 3 4 5
6 7
8
9
Low Quality
Low Pnce i
2 3 4 f
6 7
8
9
High Pnce
Dislike !23456'8 Like
Ugiy 1 2 3 4 5 6 8 0 Beautiful
Familiar 12345678a Unfamiliar
High Quality 12 3 456-89 Low Quaiir.
Low Pnce 12 3 4 5 6 7 8 9 High Pnce
Disiike 1
2 3-56
7 8
0
Like
Dislike
2 3-5
ft
0
Like
Ugly
2 5 5 6
7 8
9
Beautiful
Ugly
2 3 4 5
6
- Jj
9
Beautiful
Famihar
2 3-56
7 8
9
Unfamiliar
Familiar
2 3 4 5
6
7 8
0
Unfamiii
High Quality
2 3-56
7
9
Low Quality
High Quality
2 3 4 5
6
6
Q
Low Quaj
Low Pnce 1
2 3-56
~ s
9
High Pnce
Low Price
2 3-5
6
7 8
9
High Pnc
138

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Lamps
Dislike 1
3
4
5
6 7
8
Q
Like
Dislike
2 3
4 5
6
7 S
9
Ugly 1
*7
3
4
5
6 7
8
9
Beautiful
Ugly
2 3
4 5
6
7 8
9
Familiar 1
2
3
4
5
6 7
8
9
Unfamiliar
Familiar
2 3
4 5
6
7 8
9
High Qualirv 1
3
4
5
6 7
8
9
Low Quality
High Qualirv
2 3
a 5
t
7 8
9
Low Pnce 1
3
5
6 7
8
9
High Pnce
Low Price
4 5
6
7 8
9

Dislike
2
3 4
5
6 7
8
9
Like
Dislike
;
345
5
7 8
Ugly
2
3 4
5
6 7
8
9
Beautiful
Ugly
2
3 4 5
6
7 8
rammer
2
3 4
5
6 7
8
9
Unfamiliar
Familiar
;
3 4 5
6
7 S
High Quality
2
3 -
5
7
8
9
Low Quality
High Quality
2
3 4 5
6
7 8
Low Pnce
2
3 4
5
6 7
8
9
High Price
Low Price
;
3 4 5
6
7 8
Like
Beautiful
Unfamiliar
Low Qualify
High Price
Like
Beautiful
Unfamiliar
Low Qualirv
High Pnce
139

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Lamps
Dislike
2 3-1
5 6
7 8
9
Like
Dislike
Ugly
: 4
5 6
7 8
9
Beautiful
Ugly
Familiar !
: 3 4
5 6
7 8
9
Unfamiliar
Familiar
High Quality
5 6
~ s
9
Low Quality
High Quality
Low Price
2 3 4
5 6
7 8
9
High Pnce
Low Pnce
2 3 4 5 6 7 8 9 Like
23456789 Beautiful
23456789 Unfamiliar
2 3 -156789 Low Quality
2 3 456789 High Pnce
Dislike
: 3 4
5 6 7
8
9
Like
Dislike
2
4 5
6
7
8
9
Like
Ugly
: 3 4
5 6 7
8
9
Beautiful
Ugly 1
-
3
4 5
6
*7
8
9
Beautiful
Familiar
: 3 4
5 6 7
8
9
Unfamiliar
Familiar
2
3
4 5
6
8
Q
Unfamiliar
High Quality
2 3 4
5 6 7
8
9
Low Quality
High Quality 1
2
3
4 5
6
7
8
9
Low Quality
Low Pnce
2 3 4
5 6 7
8
9
High Price
Low Price 1
2
3
4 5
6
7
8
9
High Price
140

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Lamps
Dislike 123-i5s789 Like
Ugly 123456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 123456789 Low Quality
Low Price 123456789 High Price
Dislike ]
12 3 4 5
6 8
9
Like
Ugly 1
2 3 4 5
0 7 8
9
Beautiful
Familiar 1
2 3 4 5
s 7 S
9
Unfamiliar
High Qxiah tv 1
2 3 4 5
6 7 8
9
Low Quality
Low Pnce 1
2 3 4 5
6 7 8
9
High Pnce
Dislike
i 2
3
*
5
6 7
S
9
Like
Ugly 1
2
3
4
5
6 7
8
9
Beautiful
Familiar 1
2
3
4
5
6 7
8
9
Unfamiliar
High Quality 1
-
3
4
5
6 7
8
9
Low Quality
Low Price 1
2
3
4
5
6 7
s
9
High Price
Dislike !
3 4
5
5 ~
S 9
Like
Ugly 1
3 4

6 ~
5 9
Beautiful
Familiar 1
i
3 4
5
6
8 9
Unfamiliar
High Quality 1
-
3 4
5
6
? 9
Low Quality
Low Price 1
2
3 4
5
6
8 9
High Price
141

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Refrigerators
Dislike
:
3 4
;
6 "
8
9
Like
Dislike
2
3
5

S
0
Like
Ugly
"i
3 4
5
6 ?
s
Q
Beautiful
Ugly
2
3
4
5
6 "
8
9
Beautiful
Familiar
*
3 4
5
6 "
8
9
Unfamiliar
Familiar 1
2
3
4
5
b 7
8
Q
Unfamiliar
High Quality
? -
5
6 7
8
9
Low Quality
High Quality
2
3
4
5
b 7
8
9
Low Quality
Low Price
;
3 4
7
S
9
High Price
Low Pnce
-
3
4
5
o 7
8
9
High Price
Dislike
2 3 4 3
6 "
s
9
Like
Dislike
p
4 5
D
7 8
Like
Ugly
2 3 4 5
6 7
8
9
Beautiful
Ugly
2 p
4 5
9
7 8
Beautiful
Familiar
2 3 4 >

8
9
Unfamiliar
Familiar 1
2 3
4 5
9
j
9
Unfamiliar
High Quality
2 3 4 5

S
9
Low Quality
High Quality
p
4 5

7 8
9
Low Qualirv
Low Pnce
2 3 4 5
6 7
8
9
High Price
Low Pnce
2 3
4 5
7 8
0
High Price
142

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
...Refrigerators.
Dislike
? -1
5 6
7
8
9
Like
Dislike
:
3
4
5
6
7
8
9
Like
Ugly
3 4
5 6
7
8
9
Beautiful
Ugly
2
3
4
5
6
7
8
9
Beautiful
Familiar
3 4
5 6
7
8
9
Unfamiliar
Familiar 1
2
3
4
5
6
7
8
9
Unfamiliar
High Quality
*>
3 4
5 6
7
8
9
Low Quality
High Qualirv
3
4
5
6
7
8
9
Low Quality
Low Price
2
3 4
5 6
7
8
9
High Pnce
Low Price
o
3
4
5
6
7
8
9
High Pnce
Dislike
2
3 -
5
6
7
8
9
Like
Dislike
"i
3 4
5
6
-
8
9
Like
Ugly
2
3 9
5
6
7
8
9
Beautiful
Ugly
3 4
5
6
7
9
Beautiful
Familiar 1
i
3 4
5
A
7
8
9
Unfamiliar
Familiar ]
o
3 4
S
6
7
8
9
Unfamiliar
High Quality
*>
3 4
5
6
7
8
9
Low Quality
High Quality
->
3 4

6
7
8
9
Low Quality
Low Price
i
3 4
5
6
7
8
9
High Price
Low Pnce
2
3 4
5
6
7
8
9
High Price
143

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Refrigerators
Dislike
2
3
4
;

7 8
9
Like
Dislike
2
3
4
5
6
7 -3
9
Like
Ugly
2
3
-
5
6
7 8
9
Beautiful
Ugly
-)
3
4
5
6
7 8
9
Beautiful
Famiiiar
*)
3
4
5
6
7 8
9
Unfamiliar
Famiiiar
2
3
4
5
6
7 8
9
Unfamiliar
High Quality
9
3
4
5
6
7 8
9
Low Quality
High Quality
2
3
4
5
6
7 8
9
Low Quality
Low Pnce
2
3
4
5
6
7 8
9
High Price
Low Price
2
3
4
5
6
7 8
9
High Price
Disiike
3 4
5
6
7
8
9
Like
Disiike
2
3
A
5
6
8
9
Like
Ugly
*
3 4
5
6
7
8
9
Beautiful
Ugly
3
4
5
6
7 8
9
Beautiful
Famiiiar 1
-
3 4
5
6
7
8
9
Unfamiliar
Familiar
2
3
4
5
6
7 S
o
Unfamiliar
High Quality
:
3 4
5
6
7
8
9
Low Quality
High Quality
-i
3
4
5
6
7 8
0
Low Quality
Low Price
2
3 4
5
6
7
8
9
High Price
Low Price
2
3
4
5
6
7 8
9
High Pnce
144

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Telephones
Dislike 1 234567*9 Like
L si)' 1 3^56789 Beautiful
Familiar 133456789 Unfamiliar
High Quality 123456789 Low Quality
Low Pnce 123456789 High Pnce
Disilke 1 2 3 -i 5 6 7 S y Like
uS!y 12 3 456789 Bcauui'Ji
Familiar 1 2 3 4 5 6 7 8 9 Un/amiiiar
High Quality 12 3 4 5 6 7 8 9 Low Quality
Low Pnce 12 3 456789 High Pnce
Dislike
3
a 5
6 7
8 9
Like
Ugly 1
*>
3
4 5
6 7
8 9
Beautiful
Familiar 1
-
3
4 5
5 7
5 9
Unfamiliar
High Quaiitv l
3
4 5

5 9
Low Quality
Low Pnce 1
2
3
4 5
6 7
8 9
High Price
Dislike
2 3 4 5
6 7
8
4 Luce
Ugly
2 3 4 5
6 7
8
9 Beauuiu.
Familiar
2 5 4 5
6 7
s
0 Lnfarmuar
High Quality 1
2 3 4 5
6 7
§
9 Low Quail?.
Low Pnce 1
2 3 4 5
6 7
S
0 High Pnce
145

Pleass indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
-Telephones
Dislike
3
4
5
6
7
8
Q
Like
Dislike
i
4 5
6
7
8
9
Like
Lgiv
2
3
4
5
6
7
S
9
Beautiful
Ugly
) ^
4 5
6
7
8
9
Beautiful
Familiar
3
4
5
6
7
8
9
Unfamiliar
Familiar
2 3
4 f
6
7
8
9
Unfamiliar
High Quality
3
4
5
6
7
8
9
Low Quality
High Quality
2 3
4 5
6
7
8
9
Low Quality
Low Pnce
2
3
4
5
6
7
8
9
High Price
Low Pnce
; ?
4 5
6
7
8
9
High Pnce
Dislike
;
3 4
5
6 7
8
9
Like
Disiike
I 3 4 5
7
8
9
Like
Ugiy
2
3 4
5
6 7
8
9
Beautiful
Ugly 1
7
8
9
Beautiful
Famiuar
2
3 4
5
6 7
s
9
Unfamiliar
Familiar
3 3 4 5 6
7
8
9
Unfamiliar
High Quality
2
2 2
5
6 7
8
9
Low Quality
High Quality 1
2 3-56
7
8
9
Low Quality
Low Pnce
3 4
5
6 7
S
9
High Price
Low Price
2 3,56
7
8
9
High Price
146

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Telephones
Dislike 1 2 3 4 5 t> 1
Ugly 1 2 3 4 5 6 7
Familiar 1 2 3 4 5 6 7
High Quality 1 2 3 4 5 t> 7
Low Price 1 2 3 4 5 6 7
8
9
Like
Dislike
:
3
4
5
6
2 i
9
Like
8
9
Beautiful
Ugly
2
3
4
5
6
7 8
9
Bcauufui
8
9
Unfamiliar
Familiar
n
3
4
5
6
7 8
9
Unfamiliar
8
9
Low Quality
High Quality
2
3
4
5
6
7 g
9
Low Qualify
8
9
High Price
Low Price
9
3
4
5
6
7 8
9
High Price
Dislike
3
4 5
6 7
8
9
Like
Dislike
-i
3
4
5
6
" S
9
Like
Ugly
;
3
4 5
6 7
8
9
Beautiful
Ugly
n
3
4
5
6
g
9
Beautiful
Familiar 1
3
4 5
0 7
8
9
Unfamiliar
Familiar 1
0
3
4
5
6
" 8
9
Unfamiliar
High Quality
2
3
4 5
6 7
8
9
Low Quality
High Quality
p
3
4
5
6
_ 8
9
Low Quality
Low Price
:
3
4 5
6 7
8
9
High Price
Low Pnce
p
3
4
5
6
7 8
9
High Price
147

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Hair-drvers
Dislike
5 6
7 8
9
Like
Dislike
*1 -S
L'giy
: : -
5 6
7 S
0
Beautiful
Ugiv
*1 -
Familiar
: 3 -
5
7 8
9
Unfamiliar
Familiar ]
: 3
High Quaiirv
2 3a
5 6
7 8
9
Low Quality
High Quality
2 3
Low Pnce
2 3 4
5 6
7 8
9
High Price
Low Pnce
: ?
4
%
u
- 5 6 x 0 Like
- 5 6 $ 9 Bcauiuui
9 5 6 8 9 Unfamiliar
4 5 6 ** 8 9 Low Quality
- 5 6 7 8 9 High Pnce
Dislike
^ ^ 4
5
6
7
S
9
Like
Dislike
-i
3 4
5
6
' 3
9
Like
Ugly
2 3-i
5
6
7
8
9
Beautiful
Ugly
-i
3 4
5
6
7 8
9
Beautiful
Familiar
2 3 4
5
6
7
8
9
Unfamiliar
Familiar 1
->
3 4
5
6
7 8
9
Unfamiliar
High Quality
: 3 4
5

7
8
9
Low Quality
High Quality
2
3 4
5
6
7 8
9
Low Quality
Low Price
2 3 4
5
6
7
8
9
High Price
Low Price
3 4
5
6
7 8
9
High Pnce
148

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Hair-drvers
Dislike
Ugly
Familiar
High Quality
Low Pnce
3
4
5
6
7
8
9
Like
Dislike
i
3
4
5
6
7
8
9
Like
*>
3
4
5
6
7
8
9
Beautiful
Ugly
3
4
5
6
7
8
9
Beautiful
n
3
4
5
6
7
8
9
Unfamiliar
Familiar 1
2
3
4
5
6
7
8
9
Unfamiliar
"i
3
4
5
6
7
8
9
Low Quality
High Quality
3
4
5
6
7
8
9
Low Quality
3
4
5
6
7
8
9
High Pnce
Low Price
-
3
4
S
6
7
8
9
High Pnce
Dislike
2
3 4
5
6
7
8
9
Like
Dislike
3
4
5
6
7
8
9
Like
Ugly
2
3 4
5
6
7
8
9
Beautiful
Ugly
-
3
4
5
6
7
8
9
Beautiful
Familiar
-
3 4
5
6
7
8
9
Unfamiliar
Familiar 1
-
3
5
6
7
8
9
Unfamiliar
High Quality
2
3 4
5
6
7
8
9
Low Quality
High Quality
3
4
5
6
7
8
9
Low Quality
Low Pnce
2
3 4
5
6
7
8
9
High Price
Low Price
3
4
5
6
7
8
9
High Price
149

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Hair-drvers
Dislike
2
3 4
5
6
7
8
9
Like
Dislike
Ugly
3 4
5
6
7
8
9
Beautiful
Ugly
Familiar 1
-
3 4
5
6
7
8
9
Unfamiliar
Familiar
High Quality
3 4
5
6
7
8
9
Low Quality
High Quality
Low Price
o
3 4
5
6
7
8
9
High Pnce
Low Pnce
1 2 3
6 7 8 9 Like
6 7 8 9 Beautiful
2 3
6 7 8 9 Unfamiliar
1
6 7 8 9 Low Quality
3-156
9 High Pnce
Disiike
6 7
s
Q
Like
Dislike
3 3-5 6"
8
-
LiKe
Ugly
2 3 4 5
6 7
8
9
Beautiful
Ugly
-3-56"
8
0
Beautiful
Familiar
2 3 4 5
6 7
8
9
Unfamiliar
Familiar
3 3 5 6
8
9
Unfamiliar
High Quality
2 3 4 5
6 7
8
9
Low Quality
High Quality
2 3 56-
8
Q
Low Quality
Low Pnce
2 3 4 5
6 7
8
9
High Price
Low Pnce
2 3 5 6 7
8
9
High Pnce
150

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Dressers (chest of drawers)
Dislike
? 4
5
6
7
8
9
Like
Dislike
2
3
4
5

7
8
9
Like
Ugly
3 4
5
6
7
s
9
Beautiful
Ugly
n
3
4
5
6
7
8
9
Beautiful
Familiar
-
3 4
5
6
7
8
9
Unfamiliar
Familiar 1
-
3
4
5
6
7
8
9
Unfamiliar
High Quail rv
3 4
5
6
7
8
9
Low Quality
High Quality
3
4
5
6
7
8
9
Low Quality
Low Price
:
3 -i
<
6
7
8
9
High Price
Low Price
n
3
4
5
6
7
8
9
High Price
Dislike
9 5
6
7 8
9
Like
Dislike
2
3
4
5
7
J
9
Like
Ugly
3 9 5
6
7 8
9
Beautiful
Ugly
2
3
4
5
6
7
8
9
Beautiful
Familiar
-
3 9 5

7 8
9
Unfamiliar
Familiar
3
4
5
6
7
S
9
Unfamiliar
High Quality
-
3 4 5
6
7
8
9
Low Quality
High Quality
2
3
4
5
6
7
8
9
Low Quality
Low Price
3 4 5
6
7
8
9
High Price
Low Pnce
-i
3
4
5
6
7
8
9
High Price
151

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Dressers (chest of drawers)
Dislike
2
3
-1
5
6
7 8
9
Like
Dislike
:
3
4
5
6
7
8
9
Like
U glv
i
3
4
3
6
7 8
q
Beautiful
Ugly
3
4
5
6
7
8
9
Beautiful
Familiar
2
3
4
5
6
7 8
9
Unfamiliar
Familiar
3
4
5
6
7
8
9
Unfamiliar
High Quality
p
3
4
5
6
7 8
9
Low Quality
High Quality
3
4
5
6
7
8
9
Low Quality
Low Price
2
3
4
5
6
7 8
9
High Price
Low Price
2
3
4
5
6
7
8
9
High Pnce
Dislike
-i
3
4
5
5 7
8
9
Like
Dislike
2
3
4
5
6
7
8
9
Like
Ugly
p
3
5
6 7
8
9
Beautiful
Ugly
2
3
4
5
6
7
8
9
Beautiful
Familiar
2
3
4
5
6 7
8
9
Unfamiliar
Familiar
-
3
4
5
6
7
8
9
Unfamiliar
High Quality
P
3
4
5
6 7
8
9
Low Quality
High Quality
p
3
:
6
p
8
9
Low Qualm
Low Price
3
4
f
6 7
8
9
High Price
Low Price
2
3
4
5
6
7
8
9
High Price
152

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Dressers (chest of drawers')
Dislike 1 23456789 Like
Ugly 123456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 12345 6789 Low Quality
Low Pnce 123456789 High Price
Dislike 1 2 3 4 5 6 7 3 9 Like
t2 gly 1 23456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 123456789 Low Quality
Low Price 123456789 High Pnce
Dislike
2
3
4
5
6 "
8
9
Like
Dislike
>
3
4
5
6
7
8
9
Like
Ugly
2
3
4
5
6 7
8
9
Beautiful
Ugly
2
3
4
5
6
7
5
q
Beautiful
Familiar
-i
3
4
5
6 7
8
9
Unfamiliar
Familiar
"
3
4
5
6
7
$
9
Unfamiliar
High Quality
*>
3
4
5
6 7
8
9
Low Quality
High Quality
2
3
4
5
6
7
8
9
Low Quaiirv
Low Pnce
2
3
4
5
6 7
8
9
High Price
Low Price
2
3
4
5
6
7
8
9
High Price
153

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Clocks
Dislike
- -
5
6 '
6
9
Like
Dislike
;
3 4
<
6
7
8
9
Like
Ugly
5
6 7
S
9
Beauuful
Ugly
*i
3 4
5
6
7
8
9
Beauuful
Familiar
2
? 4
5
6
8
9
Unfamiliar
Familiar
I
3 4
5
6
7
8
9
Uni amiiiar
High Quality
3 4
5
6 7
6
9
Low Quaiiry
High Qualirv
:
3 4
5
6
7
8
9
Low Quality
Low Pnce
2
3 4
5
6 7
8
9
High Pnce
Low Pnce
2
3 4
5
6
7
8
9
High Pnce
Dislike
: ?
4 5

s
9
Like
Dislike
z
3 4
j
6 7
8
Q
Like
Ugly
-
d 5
6 7
8
9
Beautiful
Ugly
:
3 4
5
6 7
8
O
Beautiful
Familiar
: 3
4 5
6 "
2
0
Uni amiiiar
F amiiiar 1
2
3 4
5
6 7
8
0
Unfamiliar
High Quaiiry
- 3
4 5
6 7
S
9
Low Quaiirv
High Quality
2
3 4
5
6 7
8
9
Low Quaiiry
Low Pnce
2 3
4 5
6 7
8
9
High Price
Low Pnce
2
3 4
;
6 7
8
0
High Pnce
154

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scaies below each.
Clocks
Dislike
Ugly
Familiar
High Quality
Low Pnce
1 2 3 J
5 6
5 6
5 6
5 6
5 6
7 8
7 8
7 8
7 8
7 8
0 Like
9 Beautiful
9 Unfamiliar
9 Low Qu&liry
9 High Pnce
Dislike 1
Ugly 1
Familiar 1
High Quality 1
Low Pnce ]
2 3 4 5 6 8 9 Like
23456~S9 Beautiful
23456780 Unfamiliar
23456*89 Low Quality
2 3 4 5 6 7 8 0 High Pnce
Dislike
2 3
4 5
6
*7
8
9
Like
Dislike
"i
3
4
5
6
s
9
Like
Ugly
2 3
4 5
6
7
8
9
Beautiful
Ugly
-
3
4
5
6
8
9
Beau fui
Famiiiar
2 :*
4 5
6
7
8
9
Unfamiliar
Familiar ]
;
3
4
5
6 '
8
9
Unfamiliar
High Quality
2 3
4 5
6
7
8
9
Low Quality
High Quality
;
3
4
5
6 "
s
9
Low Quality
Low Pnce
2 3
4 5
6
7
8
9
High Pnce
Low Pnce
-)
3
4
5
6 *
8
9
High Pnce
155

Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Clocks
Dislike 123456789 Like
Ugly 123456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 123456789 bow Quality
Low Pnce 123456789 High Pnce
Dislike ]
i 2
3
-
5
6
7
8
9
Like
Ugly 1
3
*
5

7
8
9
Beautiful
Familiar 1
3
4

7
S
9
Unfamiliar
High Quality 1
2
3
4
c
J
6
7
8
9
Low Quality
Low Pnce 1
2
3
4
5
6
7
8
9
High Price
Dislike 1 23456789 Line
Ugly 1 23456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 1 23456789 Low Quality
Low Pnce 1*23456789 High Pnce
Dislike
1 2
3
-1
f
6 7
g
9
Like
Ugly 1
2
3
4
5
6 7
8
9
Beautiful
Familiar 1
*
3
*
5
6 7
8
9
Unfamiliar
High Quality 1
-7
3
4
5
6 7
8
9
Low Quality
Low Pnce 1
3
4
5
6 7
8
9
High Pnce
156

APPENDIX C
DIAGRAM OF EXPERIMENT 1

158
TABLE C-l
DIAGRAM OF EXPERIMENT 1
Bathroom
Scale
TV
Remote
Flash-
Light
Lamp
Refrig
erator
Tele
phone
Hair
dryer
Dresser
Clock
Version 1
Sh
Tx
Sh
Tr
Tx
Tr
Sh
Tx
Tr
Version 2
Tr
Sh
Tr
Tx
Sh
Tx
Tr
Sh
Tx
Version 3
Tx
Tr
Tx
Sh
Tr
Sh
Tx
Tr
Sh
Sh = Shape Transformation
Tr = Trim Transformation
Tx = Texture Transformation
Note: The order in which the stimulus sets were presented was reversed for half of the subjects.

APPENDIX D
DIAGRAM OF EXPERIMENT 2

160
TABLE D-l
DIAGRAM OF EXPERIMENT 2
TV
Remote
Flash
light
Lamp
Refrig
erator
Tele
phone
Hair
dryer
Dresser
Clock
Questionnaire
Version 1
(++)
(+-)
(-+)
(-)
(++)
(+-)
(-+)
(--)
Questionnaire
Version 2
(-+)
(++)
()
(+-)
(-+)
(++)
(-)
(+-)
Questionnaire
Version 3
(+-)
(-)
(++)
(-+)
(-+)
()
(++)
(-+)
Questionnaire
Version 4
(--)
(-+)
(+-)
(++)
(--)
(-+)
(+-)
(++)

APPENDIX E
PRODUCT DESCRIPTIONS USED IN EXPERIMENT 3

APPENDIX E
PRODUCT DESCRIPTIONS FOR EXPERIMENT 3
"Stroner"
"Weak"
Bathroom Scales:
Bathroom Scales:
*5 Year Warranty
*3 Year Warranty
^Maximum Weight 330 lbs.
*Maximum Weight 320 lbs.
*Easy to Read Dial
*Easy to Read Dial
*5 lbs.
*5 lbs.
TV Remote Control
TV Remote Control:
*Controls 8 devices (TV, VCR,
*Controls 6 devices (TV, VCR,
Cable, CD, Satellite, and 3
Cable, CD, and 2 auxiliaries)
auxiliaries)
*Performs all functions of the
*Performs all functions of the
original remote control
original remote control
*On screen Programming Keys
*On screen Programming Keys
*Uses 4 AAA Batteries
*Uses 4 AAA Batteries
Flashlight:
Flashlight:
* Adjustable Beam (from spot to
*Dual Reflector System
flood)
*3 Year Warranty
*Lifetime Warranty
*Uses 2-D Batteries
*Uses 2-D Batteries
*Waterproof
* Waterproof
Lamp:
Lamp:
*2 way switch (high/low)
*Rotary on/off switch
*Uses up to 100 watt bulb
*Uses up to 100 watt bulb
*Flexible Gooseneck
*Flexible Gooseneck
*Weighted (desktop base)
*Weighted (desktop base)
Refrigerator:
Refrigerator:
*21.6 cu. ft.
*20 cu. ft.
*Freezer has Removable shelf
*Frostless
*Frostless
*Energy Efficient
*Energy Efficient
*6 Storage compartments
*6 Storage compartments
*2 Door Shelves
*2 Door Shelves
162

163
APPENDIX E continued
"Strone"
"Weak"
Telephone:
Telephone:
*10 Number Memory
*6 Number Memory
*Receiver and Ringer volume
*Receiver and Ringer volume
control
control
*Mute and Redial Buttons
*Mute and Redial
*3 lbs.
*3 lbs.
Hair dryer:
Hair dryer:
*1,600 watts
*1,500 watts
*2 speeds and 3 heat settings
*2 speeds and 2 heat settings
*Dual Voltage
*Dual Voltage
*Super-quiet operation
*Super-quiet operation
*2 lbs.
*2 lbs.
Dresser:
Dresser:
*2 Spacious Drawers
*2 Spacious Drawers
*Drawers have steel slides with
*Metal on wood Drawer guide
ball bearings for smooth
*Durable chip resistant finish
operation
*Hardwood Back
*Durable chip-resistant finish
*Hardwood Back
Clock:
Clock:
*Illuminated Dial (2 brightness
*Illuminated Dial
settings)
*Snooze Button
*Snooze Button
*Adjustable soft/loud Alarm
* Adjustable soft/loud Alarm
*9 v Battery Backup
*9 v Battery Backup

REFERENCE LIST
Ajzen, leek (1989), "Attitude Structure and Behavior," in Attitude Structure
and Function, eds. Anthony R. Pratkanis, Steven J. Breckler, and
Anthony G. Greenwald, Hillsdale, NJ: Lawrence Erlbaum Associates,
Publishers, pp. 241-274.
Auld, Frank (1981), "A Theory Deriving Preference from Conflict," in Advances
in Intrinsic Motivation and Aesthetics, ed. Hy I. Day, New York: Plenum
Press, pp. 203-224.
Bell, Stephen S., Morris B. Holbrook, and Michael R. Solomon (1991),
"Combining Esthetic and Social Value to Explain Preferences for
Product Styles with the Incorporation of Personality and Ensemble
Effects," The Journal of Social Behavior and Personality. Vol. 6, No. 6,
pp. 243-273.
Berkowitz, Marvin (1987), "The Influence of Shape on Product Preferences,"
Advances in Consumer Research. Vol. 14, p. 559.
Berlyne, D. E. (1970), "Novelty, Complexity, and Hedonic Value," Perception
and Psychophysics. Vol. 8, pp. 279-286.
Berlyne, D. E. (1971), Aesthetics and Psychology. New York: Appleton-
Century-Crofts.
Berlyne, D.E. (1974), Studies in the New Experimental Aesthetics. New York:
John Wiley and Sons.
Bornstein, Marc H., Kay Ferdinandsen, and Charles G. Gross (1981),
"Perception of Symmetry in Infancy," Developmental Psychology.
Vol. 17, No. 1, pp. 82-86.
Carruthers, Margaret (1970), "Color-Form Dominance and Memory for Color,"
M. A. thesis, Psychology Dept., University of Glasgow.
Ching, Francis D. K. (1979), Architecture: Form, Space, and Order. New York:
Van Nostrand Reinhold Company.
164

165
Cohen, J (1977), Statistical Power Analysis for the Behavioral Sciences. (Rev.
ed.), New York: Academic Press, 1977.
Dickson, Peter R. and James L. Ginter (1987), "Market Segmentation, Product
Differentiation, and Marketing Strategy," Journal of Marketing. (April),
pp. 1-10.
Edwards, Kari (1990), "The Interplay of Affect and Cognition in Attitude
Formation and Change," Journal of Personality and Social Psychology,
Vol. 59, pp. 202-216.
Eisenman, R. (1968), "Novelty Ratings of Simple and Complex Shapes,"
Journal of General Psychology. 78, pp. 275-278.
Garvin, David A. (1984), "What Does "Product Quality" Really Mean?" Sloan
Management Review, (Fall), pp. 25-43.
Glass, Lewis A. and Keith J. Holyoak (1986), Cognition, Second ed., New York,
N.Y.: Random House.
Goldstone, Robert L., Douglas L. Medin, and Dedre Gentner (1991), "Relational
Similarity and the Nonindependence of Features in Similarity
Judgements," Cognitive Psychology, Vol. 23, pp. 222-262.
Gorski, D. (1987), The Cognitive Condition of Design, ed. Brian Zaff,
Symposium Proceedings (March) 1987.
Gruenwald, George (1985), New Product Development: What Really Works,
Chicago, IL: Crain Books.
Hartmann, George W. (1935), Gestalt Psychology: A Survey of Facts and
Principles, New York: The Ronald Press Company.
Holbrook, Morris B. (1981), "Introduction: The Esthetic Imperative in
Consumer Research," in Symbolic Consumer Behavior, eds. Elizabeth C.
Hirschman and Morris B. Holbrook, Ann Arbor, MI: Association for
Consumer Research, pp. 36-37.
Holbrook, Morris and Elizabeth Hirschman (1982), "The Experiential Aspects
of Consumption: Consumer Fantasies, Feelings, and Fun," Journal of
Consumer Research, Vol. 9 (2), pp. 132-140.

166
Holbrook, Morris B. and Robert B. Zirlin (1985), "Artistic Creation, Artworks,
and Aesthetic Appreciation: Some Philosophical Contributions to Non
profit Marketing," Advances in Nonprofit Marketing, Vol. 1, pp. 1-54.
Holt, S. (1985), "Design, the Ninth Principle of Excellence: The Product Half
of the Business Equation," Innovation, (Fall), pp. 2-4.
Huber, Joel, and Morris B. Holbrook (1981), "The Use of Real Versus Artificial
Stimuli in Research on Visual Esthetic Judgements," in Symbolic
Consumer Behavior, eds. Elizabeth C. Hirschman and Morris B.
Holbrook, Ann Arbor, MI: Association for Consumer Research, pp. 60-68.
Hutcheson, F. (1725), An Inquiry into the Original of Our Ideas of Beauty and
Virtue, London: Darby.
Hutchinson, J. Wesley and Joseph W. Alba (1991), "Ignoring Irrelevant
Information: Situational Determinants of Consumer Learning," Journal
of Consumer Research, Vol. 18, (December), pp. 325-345.
Johnson, Mark (1987), The Body in the Mind, Chicago: The University of
Chicago Press.
Katz, David (1950), Gestalt Psychology, New York: The Ronald Press
Company.
Keppel, Geoggrey (1991), Design and Analysis: A Researchers Handbook, 3rd
ed., Englewood Cliffs, NJ: Prentice-Hall, Inc.
Koffka, K. (1935), Principles of Gestalt Psychology, New York: Harcourt, Brace
and Company.
Kohler, Wolfgang (1929), Gestalt Psychology: An Introduction to New Concepts
in Modern Psychology, New York: Liveright Publishing Corporation
(1947).
Kotler, Philip and G. Alexander Rath (1984), "Design: A Powerful But
Neglected Strategic Tool," The Journal of Business Strategy, Vol. 5
(Fall), pp. 16-21.
Kunst-Wilson, W. R. and R. B. Zajonc (1980), "Affective Discrimination of
Stimuli That Cannot Be Recognized," Science, Vol. 207, pp. 557-558.

167
Landy, David and Harold Sigall (1974), "Beauty Is Talent: Task Evaluation as
a Function of the Performers Physical Attractiveness," Journal of
Personality and Social Psychology, Vol. 29, pp. 299-304.
Langlois, Judith H. and Lori A. Roggman (1990), "Attractive Faces Are Only
Average," Psychological Science. Vol. 1, No. 2, March, pp. 115-121.
Lauer, David A. (1979), Design Basics. New York: Holt, Rinehart, and
Winston.
Lennon, Sharron J. (1990), "Effects of Clothing Attractiveness on Perceptions,"
Home Economics Research Journal. Vol. 18, pp. 303-310.
Lewalski, Z. (1988), Products Esthetics: An Interpretation for Designers.
Carson City, New York: Design and Development Engineering Press.
Lewicki, Pawel (1986), Nonconscious Social Information Processing. New York:
Academic Press.
Loken, Barbara and James Ward (1990), "Alternative Approaches to
Understanding the Determinants of Typicality," Journal of Consumer
Research. Vol. 17, pp. 111-126.
Lorenz, Christopher (1986), The Design Dimension. New York, NY: Basil
Blackwell Inc.
Mandler, George (1982), "The Structure of Value: Accounting for Taste," in
Affect and Cognition: The 17th Annual Carnegie Symposium, eds.
Margaret S. Clark and Susan T. Fiske, Hillsdale, NJ: Lawrence
Erlbaum Associates, pp. 3-36.
Meyers-Levy, Joan and Alice M. Tybout (1989), "Schema Congruity as a Basis
for Product Evaluation," Journal for Consumer Research. Vol. 16 (June),
pp. 39-54.
Monroe, Kent B. (1973), "Buyers Subjective Perceptions of Price," Journal of
Marketing Research. Vol. 8, February, pp. 70-80.
Mothersill, Mary (1989), "Aesthetic Laws, Principles and Properties: A
Response to Eddy Zemach," The Journal of Aesthetics and Art Criticism,
Vol. 47, No. 1 (Winter), pp. 77-82.

168
Nedungadi, Prakash and J. Wesley Hutchinson (1985), "The Prototypicality of
Brands: Relationships with Brand Awareness, Preference and Usage,"
in Advances in Consumer Research, eds. Elizabeth C. Hirschman and
Morris Holbrook, Provo UT: Association for Consumer Research, Vol.
12, pp. 498-503.
Nisbett, Richard E. and Timothy DeCamp Wilson (1977), "Telling More Than
We Can Know: Verbal Reports on Mental Processes," Psychological
Review, Vol. 84 (May), pp. 231-259.
Nunnally, Jum C. (1981), "Explorations of Exploration," in Advances in
Intrinsic Motivation and Aesthetics, ed., Hy I. Day, New York: Plenum
Press, pp. 87-130.
Nussbaum, Bruce (1991), "Winners: The Best Product Designs of the Year," in
Business Week, June 17, 1991, pp. 62-80.
Oakley, Mark (1990), "Design and Design Management," in Design
Management: A Handbook of Issues and Methods, ed. Mark Oakley,
Cambridge, MA: Basil Blackwell Inc., pp. 3-14.
Olson, Jerry C. (1981), "What is an Esthetic Response?" Symbolic Consumer
Behavior, eds. Elizabeth C. Hirschman and Morris B. Holbrook, Ann
Arbor, MI: Association for Consumer Research, pp. 71-74.
Patzer, G. L. (1985), The Physical Attractiveness Phenomena, New York:
Plenum Press.
Pickford, R. W. (1972), Psychology and Visual Aesthetics, London: Hutchinson
Educational Ltd., (Distributed in the U. S. by Crane, Russak and Co.
Inc., New York, N. Y.).
Pleydell-Pearce, A.G. (1970), "Objectivity and Value in the Judgements of
Aesthetics," The British Journal of Aesthetics, Vol. 10, No. 1, pp. 25-38.
Pye, David (1978), The Nature and Aesthetics of Design, New York: Van
Nostrand Reinhold Company.
Reed, J. D. (1990), "The New Shape of Sound," in Time, August 27, 1990, p. 58.
Rosch, E. (1978), "Principles of Categorization," in E. Rosch and B. B. Lloyd
(Eds.), Cognition and Categorization, Hillsdale, NJ: Erlbaum, pp. 27-47.

169
Segall, Marshall H. (1976), "Visual Art: Some Perspectives from Cross-Cultural
Psychology," in Beyond Aesthetics: Investigations into the Nature of
Visual Art, ed. Don Brothwell, Thames and Hudson: London, 1976.
Sproles, George B. (1981), "Analyzing Fashion Life Cycles Principles and
Perspectives," Journal of Marketing. Vol. 45 (Fall), pp. 116-124.
Urban, Glen L. and John R. Hauser (1993), Design and Marketing of New
Products. Englewood Cliffs, NJ: Prentice-Hall, Inc.
Veryzer, Robert W. (1993), "Aesthetic Response and the Influence of Design
Principles on Product Preferences," in Advances in Consumer Research.
eds., Leigh McAlister and Michael L. Rothschild, Provo, UT: Association
of Consumer Research, Vol. 20, pp. 224-228.
Wallendorf, Melanie (1980), "The Formation of Aesthetic Criteria Through
Social Structures and Social Institutions," in Advances in Consumer
Research, ed., Jerry C. Olson, Ann Arbor, MI: Association for Consumer
Research, Vol. 7, pp. 3-6.
Whitney, Danial E. (1988), "Manufacturing by Design," Harvard Business
Review, (July-August), pp.83-90.
Wilkie, William L. (1990), Consumer Behavior, 2nd ed., New York: John Wiley
and Sons.
Wohlwill, Joachim F. (1981), A Conceptual Analysis of Exploratory Behavior:
The "Specific-Diverse" Distinction Revisited, in Advances in Intrinsic
Motivation and Aesthetics, ed. Hy I. Day, New York: Plenum Press, pp.
341-364.
Winer, B. J., Donald R. Brown, and Kenneth M. Michels (1991), Statistical
Principles in Experimental Design, New York, NY: McGraw-Hill, Inc.
Woll, Stanley and Arthur Graesser (1982), "Memory Discrimination for
Information typical and Atypical of Person Schemata," Social Cognition,
Vol. 1, No. 42, pp. 287-310.
Zaff, Brian S. (1987), The Cognitive Condition of Design, ed. Brian S. Zaff,
Symposium Proceedings (March), 1987.
Zajonc, R. B. (1980), "Feeling and Thinking: Preferences Need No Inferences,"
American Psychologist, 35, pp. 151-175.

170
Zeithaml, Valarie A. (1988), "Consumer Perceptions of Price, Quality, and
Value: A Means-End Model and Synthesis of Evidence," Journal of
Marketing, Vol. 52, pp. 2-22.

BIOGRAPHICAL SKETCH
Robert Whitman Veryzer, Jr. is the third of four sons born to Robert and
Marion Veryzer. He attended Birmingham Seaholm High School in
Birmingham, Michigan. From 1979 to 1983 Robert attended Olivet College
where he was the recipient of the Olivet College Presidential Scholarship and
was graduated with honors. After working for a year for a small
manufacturing company, he left to pursue a Master of Business Administration
degree in marketing at Michigan State University. In addition to his studies,
Robert worked for the State of Michigan as a marketing assistant and later for
General Motors Corporation-Oldsmobile Product Planning Department as a
product research analyst. After completing his M. B. A. degree, he spent one
semester conducting an independent research project that involved examining
the role of product design in business. Robert then went to work for General
Foods Corporation as an assistant product manager. In the Fall of 1989, he
entered the Ph.D. program in marketing at the University of Florida. Robert
has accepted an offer to join the faculty of Rensselaer Polytechnic Institute.
171

I certify that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of I^optor^f I^hilo^qghy.
J/Wesley Hutchinson, Chairman
associate professor of Marketing
I certify that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of Doctor of Philosophy
Richard J. Lutz
Professor of Marketing
I certify that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of Doctor of Philosophy.
Associate Professor of Marketing
I certify that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of Doctor of Philosophy.
David G. Mick
Assistant Professor of Marketing
I certifjr that I have read this study and that in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate, in
scope and quality, as a dissertation for the degree of Doctor of Philosophy.
^Jonathan Hamilton
Associate Professor of Economics
This dissertation was submitted to the Graduate Faculty of the
Department of Marketing in the College of Business Administration and to the
Graduate School and was accepted as partial fulfillment of the requirements
for the degree of Doctor of Philosophy.
August, 1993
Dean, Graduate School



23
the physical composition of the product (e. g., texture, color, flavor, etc.), are
very important in signaling perceived quality to the consumer. In a study by
Berkowitz (1987), consumers seemed to make unconscious inferences
concerning freshness, taste, and quality based on the shapes of the products.
Berkowitz examined consumer reaction to a food product frozen corn on the
cob of two shapes (full ears with squared-off ends and full ears with
untrimmed ends), in order to determine: (1) whether the shape of the product
would influence preference; and (2) whether preference levels would vary with
involvement and experience with the product category. The experimental
design involved paired comparison tests at laboratory kitchens in enclosed
malls and sequential monadic tests in subsequent home placements. Test
panelists included 286 female homemakers of which 184 currently purchased
the frozen variety of corn and 102 bought only the fresh variety. The findings
showed a marked preference for the untrimmed shape. Preference ratios
comparing preference scores for the rounded, untrimmed shape to those for the
squared-off shape were:
- Laboratory test 1.1: 1 frozen users; 2.0: 1 fresh only
- Home placement 1.8: 1 frozen users; 2.2: 1 fresh only
The results were statistically significant at the .01 level. Nearly four out of
five consumers said the reason for their choice in the home test was better
taste, about half said the untrimmed was a more natural product, and half
reported better texture. Visual appeal or a more pleasing shape, per se, were


121
should pay particular attention to product design for products that will be
evaluated in close proximity (either physically or in photographs) to
competitors products since the influence of aesthetic response is likely to play
a role in affecting key product perceptions (e. g., beauty, quality). This could
also be true for instances where consumers collect product brochures in order
to "directly" compare products (e.g., automobile brochures). The findings of
these studies also seem to suggest that in cases where a product is known to
have a poor appearance an effort should be made to display it apart from other
products that have more appealing designs. Such action would reduce the
17
salience of the aesthetic aspects of the product.
One interesting finding that would seem to have implications for the
designing of new products concerns the relationship between novelty and price.
The novel (i. e., atypical) products in Experiments 1 and 2 were associated
with higher prices regardless of the level of unity that they exhibited. The
association of higher priced products with the prototypicality of a products
appearance and the association of quality with the unity aspect of a products
design would seem to suggest that product appearance (i. e., design) can be
used to position a product with respect to (seemingly) "non-aesthetic"
dimensions (i. e., price, quality). The implication is that design could be used
17
This course of action is offered with great reservation. The best
long-term course of action would be to redesign the product.
However, in cases where a basically sound product has failed to
achieve commercial success due to an initially poor design such a
course of action may be warranted.


56
TABLE III 6
ANALYSIS OF MANIPULATION STRENGTH DIFFERENCES -
LINEAR UNITY SCORES
Product
UtiScore for
Tleauty
S. E.
t-test
1. Bathroom scale
Sh
0.8335
0.2032
0.8201
Tr
0.1212
0.1449
1.6726*
Tx
-0.8923
0.4442
-4.0172
2. TV Remote Control
Sh
0.5076
0.3853
2.6345**
Tr
0.4384
0.3913
2.2410*
Tx
0.8409
0.4711
3.5696**
3. Flashlight
Sh
1.2032
0.3527
6.8233***
Tr
0.6969
0.3547
3.9298***
Tx
2.9394
0.4888
12.0267***
4. Lamp
Sh
0.0000
0.2412
0.0000
Tr
0.1385
0.2787
0.9938
Tx
0.6770
0.4214
3.2124**
5. Refrigerator
Sh
2.7955
0.4302
12.9961***
Tr
1.2770
0.4308
5.9281***
Tx
2.3359
0.4761
9.8119***
6. Telephone
Sh
0.9000
0.4309
4.1773***
Tr
0.1485
0.2062
1.4401
Tx
0.9015
0.4160
4.3341***


TABLE IV 11
ANALYSIS OF VARIANCE FOR FAMILIARITY DEPENDENT MEASURE
101
Source
DF
Type III SS
F-Value
P > F
Between Subiect Effects
Version
3
121.7994
2.90
0.0358
Order
1
20.4949
1.46
0.2276
Version Order
3
24.4360
0.58
0.6276
Error Sub (Version Order)
230
3221.0752
Within Subiect Effects
Product
7
841.7213
39.87
0.0001
Feature A
1
25.0684
8.31
0.0100
Feature B
1
151.9619
50.39
0.0010
A B
1
86.7198
28.76
0.0010
Prod Version
18
388.4214
7.16
0.0010
Prod Order
7
8.7496
0.41
0.8938
A Order
1
4.3303
1.44
0.2500
B Order
1
7.4252
2.46
0.2500
A B Order
1
2.2846
0.76
0.2500
Prod Version Order (Residual)
18
71.5838
1.32
0.2500
Error Sub (Product)
1610
4885.5859
perceived familiarity involves more than the recall of the features shared with
the category prototype.
Summary
This experiment further examined the influence of unity and proto
typicality on aesthetic and derived responses. The experiment employed a
between-subjects design and thus the subjects evaluated the product versions
without the benefit of the three other versions (i.e., reference points) of the
stimulus set. The results of this experiment are summarized in Figure IV-1.


27
perceptions, then one would expect these design factors to influence the ratings
of products in accordance with (the level of) their presence (or absence) in the
products. Aesthetic responses would be expected to be more positive for objects
(products) exhibiting high unity than they would for objects (products) that
were not consistent with this visual organization principle. In general,
prototypicality would be expected to lead to more favorable aesthetic responses
than would atypicality; however, there is some question about the nature of
this relationship with regard to the level of prototypicality (i. e., prototypical
or moderately atypical) that maximizes positive aesthetic response. The unity
and typicality factors would also be expected to influence non-aesthetic
evaluations or derived responses such as attitude, perceived quality, and price
expectations.
In the chapters that follow a series of three experiments that examine
the hypotheses that have been presented here are discussed. Experiments 1
and 2 examine the influence of unity and prototypicality on aesthetic responses
(i. e., HI, H2, H3, H4, and H5) and derived responses (i. e., H7, H8, and H9).
The influence of unity on perceived familiarity is also examined (i. e., H6).
Experiment 3 focuses on the influence of product aesthetics on the attitude and
perceived quality derived responses (i. e., H7 and H8). In this experiment,
additional information is presented during the evaluation task in order to
examine whether or not the presence of the information moderates the
influence of aesthetic response on the derived responses.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Telephones
Dislike 1 2 3 4 5 t> 1
Ugly 1 2 3 4 5 6 7
Familiar 1 2 3 4 5 6 7
High Quality 1 2 3 4 5 t> 7
Low Price 1 2 3 4 5 6 7
8
9
Like
Dislike
:
3
4
5
6
2 i
9
Like
8
9
Beautiful
Ugly
2
3
4
5
6
7 8
9
Bcauufui
8
9
Unfamiliar
Familiar
n
3
4
5
6
7 8
9
Unfamiliar
8
9
Low Quality
High Quality
2
3
4
5
6
7 g
9
Low Qualify
8
9
High Price
Low Price
9
3
4
5
6
7 8
9
High Price
Dislike
3
4 5
6 7
8
9
Like
Dislike
-i
3
4
5
6
" S
9
Like
Ugly
;
3
4 5
6 7
8
9
Beautiful
Ugly
n
3
4
5
6
g
9
Beautiful
Familiar 1
3
4 5
0 7
8
9
Unfamiliar
Familiar 1
0
3
4
5
6
" 8
9
Unfamiliar
High Quality
2
3
4 5
6 7
8
9
Low Quality
High Quality
p
3
4
5
6
_ 8
9
Low Quality
Low Price
:
3
4 5
6 7
8
9
High Price
Low Pnce
p
3
4
5
6
7 8
9
High Price
147


44
TABLE III 1 continued
MANIPULATION CHECK FOR THE EFFECT OF A CHANGE
IN FEATURE A OR B ON BEAUTY RATINGS
Product
Feature A
Feature B
7. Hair dryer
Sh
-1.59
-3.72
Tr
0.33
0.33
Tx
-3.58
-3.77
8. Dresser
Sh
-0.13
-0.54
Tr
-1.88
-1.75
Tx
-2.07
-1.20
9. Clock
Sh
-3.33
-3.00
Tr
-1.43
-1.44
Tx
-1.60
-1.72
Mean effect of change
across products
-1.24
-1.34
manipulation (i. e., change in Features A and B via the transformation type)
as it is adapted for each of the stimulus sets.
The Influence of Unity and Prototypicality on Aesthetic and Derived Responses
The critical test concerns what happens to aesthetic responses in the
case where both Feature A and Feature B are altered so that they are less
prototypical but more unified. The prediction consistent with the
"prototypicality" hypothesis is that in such a case the aesthetic response
ratings should go down since the resulting stimulus is quite atypical. The


167
Landy, David and Harold Sigall (1974), "Beauty Is Talent: Task Evaluation as
a Function of the Performers Physical Attractiveness," Journal of
Personality and Social Psychology, Vol. 29, pp. 299-304.
Langlois, Judith H. and Lori A. Roggman (1990), "Attractive Faces Are Only
Average," Psychological Science. Vol. 1, No. 2, March, pp. 115-121.
Lauer, David A. (1979), Design Basics. New York: Holt, Rinehart, and
Winston.
Lennon, Sharron J. (1990), "Effects of Clothing Attractiveness on Perceptions,"
Home Economics Research Journal. Vol. 18, pp. 303-310.
Lewalski, Z. (1988), Products Esthetics: An Interpretation for Designers.
Carson City, New York: Design and Development Engineering Press.
Lewicki, Pawel (1986), Nonconscious Social Information Processing. New York:
Academic Press.
Loken, Barbara and James Ward (1990), "Alternative Approaches to
Understanding the Determinants of Typicality," Journal of Consumer
Research. Vol. 17, pp. 111-126.
Lorenz, Christopher (1986), The Design Dimension. New York, NY: Basil
Blackwell Inc.
Mandler, George (1982), "The Structure of Value: Accounting for Taste," in
Affect and Cognition: The 17th Annual Carnegie Symposium, eds.
Margaret S. Clark and Susan T. Fiske, Hillsdale, NJ: Lawrence
Erlbaum Associates, pp. 3-36.
Meyers-Levy, Joan and Alice M. Tybout (1989), "Schema Congruity as a Basis
for Product Evaluation," Journal for Consumer Research. Vol. 16 (June),
pp. 39-54.
Monroe, Kent B. (1973), "Buyers Subjective Perceptions of Price," Journal of
Marketing Research. Vol. 8, February, pp. 70-80.
Mothersill, Mary (1989), "Aesthetic Laws, Principles and Properties: A
Response to Eddy Zemach," The Journal of Aesthetics and Art Criticism,
Vol. 47, No. 1 (Winter), pp. 77-82.


32
all of the products that were used as stimuli was established through an
examination of the products available on the market and was confirmed in
pilot tests.1 This product form is also unified in that there is consistency or
compatibility among the shapes of the base (Feature A) and the handset
(Feature B). In this (prototypical) case the number of shared features with the
category schema is high (positive) and unity or a visual connection or
consistency among product features is also high (positive). In the upper right-
hand cell of Figure III-l, the base (Feature A) of the product has been altered
using a shape transformation so that it is no longer the same as that of the
category prototype. This change has at the same time decreased the unity
between the handset feature and the base feature. Thus, this version of the
telephone product is atypical (+-) with respect to one of the two features being
manipulated and is no longer unified. The same is true of the product in the
lower-left cell of Figure III-l. In this case, the shape of the handset was
altered in order to effect the manipulations of prototypicality and unity that
would produce an atypical (-+) stimulus on one product feature and an "un
unified" appearance. The base and handset features of the product in the
lower-right corner of Figure III-l have both been altered. In this case the
product no longer shares the shape of either Feature A or Feature B with the
1 The aggregate mean ratings for the familiarity of the eleven stimulus
sets that were pilot tested were as follows: Unity/Prototypical (M =
6.46), Non-unity/Atypical (M = 4.86), Non-unity/Atypical (M = 4.96), and
Unity/Atypical (4.76). Product variations were rated on 9-point semantic
differential scales with 9 being the most familiar.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Clocks
Dislike 123456789 Like
Ugly 123456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 123456789 bow Quality
Low Pnce 123456789 High Pnce
Dislike ]
i 2
3
-
5
6
7
8
9
Like
Ugly 1
3
*
5

7
8
9
Beautiful
Familiar 1
3
4

7
S
9
Unfamiliar
High Quality 1
2
3
4
c
J
6
7
8
9
Low Quality
Low Pnce 1
2
3
4
5
6
7
8
9
High Price
Dislike 1 23456789 Line
Ugly 1 23456789 Beautiful
Familiar 123456789 Unfamiliar
High Quality 1 23456789 Low Quality
Low Pnce 1*23456789 High Pnce
Dislike
1 2
3
-1
f
6 7
g
9
Like
Ugly 1
2
3
4
5
6 7
8
9
Beautiful
Familiar 1
*
3
*
5
6 7
8
9
Unfamiliar
High Quality 1
-7
3
4
5
6 7
8
9
Low Quality
Low Pnce 1
3
4
5
6 7
8
9
High Pnce
156


86
of four questionnaire versions and each questionnaire version contained one
cell (i. e., product variation, e. g., Unity/Prototypical) of the 2(Feature A)x
2(Feature B) manipulation from each of the product categories (see Appendix
D for a diagram of this design). The product versions (e. g., Unity/Prototypical)
were distributed evenly throughout the four versions of the questionnaire.
Although the three types of transformations were represented across the eight
stimulus sets, transformation type was not treated as a factor in this design.
The order in which the stimulus products were presented was reversed for half
of the subjects.
Experimental Procedure
Two-hundred and forty volunteer subjects enrolled in the Introductory
Marketing course at the University of Florida participated in this experiment.
The procedure for this experiment was the same as the procedure for
Experiment 1 except that in each of the four questionnaire versions subjects
saw and rated only one version of each product for each of the eight stimulus
sets. The dependent measures were the same measures that were used in
Experiment 1.


119
shopping situations in that the products being evaluated were presented along
with descriptive information. Even though subjects were allowed to evaluate
the products in whatever manner that they wished, the aesthetic information
(i. e., product appearance) clearly played a role in influencing their attitudes
toward the products. The descriptive information did, however, seem to
supersede aesthetic information with respect to perceived quality.
Across the three experiments the effects of unity were quite robust
although they were more pronounced in the experiment that was conducted
within-subjects than they were in the between-subjects experiments. This is
not surprising since within-subjects designs afford more statistical power than
do between-subjects designs. The fact that these effects were indicated despite
the use of rather primitive line drawings suggests that the effects may be
stronger in the real-world since a "greater range of perceptual salience can be
achieved" through the use of color, materials, and three dimensions
(Hutchinson and Alba 1991, p. 342). It is important to note that the effects of
unity on derived responses that were observed in Experiments 1 and 2 could
have been influenced by the procedure that was employed (i.e., having subjects
rate the products on all five of the dependent measures). The fact that
the patterns exhibited by each of the dependent variables were not the same
(e. g., price) coupled with the fact that the effect of unity was observed in the
16
The direction of two of the rating scales was reversed in order to
reduce response bias on the part of subjects.


15
cality theories were not developed to specifically address aesthetic issues, the
absence of a theory of aesthetic response has led consumer researchers as well
as others to rely on prototypicality as a default theory for explaining aesthetic
response.
A number of explanations have been suggested for the relationship
between prototypicality and preference. One explanation for the relationship
between prototypicality and preference/attitude suggests that more prototypical
items are more familiar and therefore better liked. Familiarity refers to either
an items meaningfulness (i. e., perceived knowledge about an item) or the
frequency of exposure to the item (Loken and Ward 1990). Another
explanation suggests that more prototypical category members are preferred
because they have more valued attributes. This explanation does not hold that
prototypically per se leads to product preference, but rather maintains that as
product categories evolve one or a few products tend to become market-share
leaders because they have attributes widely desired by consumers who buy the
product. Competitive brands are designed to appeal to the same segment(s)
of consumers so they are similar in many ways to market leaders (Loken and
Ward 1990). It has also been suggested that the link between prototypicality
and preference may in part be due to the information theory notion of
redundancy in that prototypes appear to be just those members of a category
that most reflect the "redundancy structure" of the category as a whole (Rosch
1978, p. 37). The preceding discussion suggests the following hypothesis:


40
Products, and Transformation were all within-subjects factors. Particular
transformations of the nine product categories (i. e., stimulus sets) were
contained in the questionnaire Versions. Questionnaire Version and Order
were between-subjects factors. The order in which the stimulus products
(stimulus sets) were presented was reversed for half of the subjects.
Experimental Procedure
One-hundred and ninety-seven volunteer subjects enrolled in the
Introductory Marketing course at the University of Florida participated in this
O
experiment. Subjects received extra credit for their participation. The
subjects were run in groups of 10-20 participants. The stimulus materials
were contained in a booklet. The introductory page informed the subjects that
they would be shown drawings of products and asked to evaluate the appeal
of the product ideas based on their appearance. The subjects were told that
the purpose of the study was to obtain consumers reactions to products that
companies were considering for introduction. It was also explained to the
subjects that the products they were going to be evaluating were in the early
stage of the product development process and for this reason the drawings
were in a very rough (unfinished) form. Subjects were then told that each of
the versions of a product performed equally well and that they were to rate all
O
Power estimates using pilot data and the procedure suggested by Cohen
(1977, pp. 364-379) for factorial designs indicated that this sample size
would be sufficient for estimating effect sizes at the p < .05 level.


168
Nedungadi, Prakash and J. Wesley Hutchinson (1985), "The Prototypicality of
Brands: Relationships with Brand Awareness, Preference and Usage,"
in Advances in Consumer Research, eds. Elizabeth C. Hirschman and
Morris Holbrook, Provo UT: Association for Consumer Research, Vol.
12, pp. 498-503.
Nisbett, Richard E. and Timothy DeCamp Wilson (1977), "Telling More Than
We Can Know: Verbal Reports on Mental Processes," Psychological
Review, Vol. 84 (May), pp. 231-259.
Nunnally, Jum C. (1981), "Explorations of Exploration," in Advances in
Intrinsic Motivation and Aesthetics, ed., Hy I. Day, New York: Plenum
Press, pp. 87-130.
Nussbaum, Bruce (1991), "Winners: The Best Product Designs of the Year," in
Business Week, June 17, 1991, pp. 62-80.
Oakley, Mark (1990), "Design and Design Management," in Design
Management: A Handbook of Issues and Methods, ed. Mark Oakley,
Cambridge, MA: Basil Blackwell Inc., pp. 3-14.
Olson, Jerry C. (1981), "What is an Esthetic Response?" Symbolic Consumer
Behavior, eds. Elizabeth C. Hirschman and Morris B. Holbrook, Ann
Arbor, MI: Association for Consumer Research, pp. 71-74.
Patzer, G. L. (1985), The Physical Attractiveness Phenomena, New York:
Plenum Press.
Pickford, R. W. (1972), Psychology and Visual Aesthetics, London: Hutchinson
Educational Ltd., (Distributed in the U. S. by Crane, Russak and Co.
Inc., New York, N. Y.).
Pleydell-Pearce, A.G. (1970), "Objectivity and Value in the Judgements of
Aesthetics," The British Journal of Aesthetics, Vol. 10, No. 1, pp. 25-38.
Pye, David (1978), The Nature and Aesthetics of Design, New York: Van
Nostrand Reinhold Company.
Reed, J. D. (1990), "The New Shape of Sound," in Time, August 27, 1990, p. 58.
Rosch, E. (1978), "Principles of Categorization," in E. Rosch and B. B. Lloyd
(Eds.), Cognition and Categorization, Hillsdale, NJ: Erlbaum, pp. 27-47.


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scaies below each.
Clocks
Dislike
Ugly
Familiar
High Quality
Low Pnce
1 2 3 J
5 6
5 6
5 6
5 6
5 6
7 8
7 8
7 8
7 8
7 8
0 Like
9 Beautiful
9 Unfamiliar
9 Low Qu&liry
9 High Pnce
Dislike 1
Ugly 1
Familiar 1
High Quality 1
Low Pnce ]
2 3 4 5 6 8 9 Like
23456~S9 Beautiful
23456780 Unfamiliar
23456*89 Low Quality
2 3 4 5 6 7 8 0 High Pnce
Dislike
2 3
4 5
6
*7
8
9
Like
Dislike
"i
3
4
5
6
s
9
Like
Ugly
2 3
4 5
6
7
8
9
Beautiful
Ugly
-
3
4
5
6
8
9
Beau fui
Famiiiar
2 :*
4 5
6
7
8
9
Unfamiliar
Familiar ]
;
3
4
5
6 '
8
9
Unfamiliar
High Quality
2 3
4 5
6
7
8
9
Low Quality
High Quality
;
3
4
5
6 "
s
9
Low Quality
Low Pnce
2 3
4 5
6
7
8
9
High Pnce
Low Pnce
-)
3
4
5
6 *
8
9
High Pnce
155


APPENDIX A
PRODUCTS EXPLORED FOR USE AS STIMULI
The products employed in the experiment were drawn from a larger set
of stimuli that had been explored with respect to the proposed manipulations.
Suitable stimuli can be created for virtually any product by manipulating two
features of the product (e. g., shapes: circle/square; texture: smooth/rough;
trim: present/absent; angle: 90-degree/60-degree). A partial list of products
that have been explored with respect to the proposed manipulations includes
the following:
Alarm clocks
Ice chests
Silverware
Calculators
Irons
Staplers
Cameras
Kitchen timers
Stereos
Clock radios
Lamps
Sunglasses
Coffee pots
Make-up compacts
Telephones
Cologne bottles
Mirrors
Thermoses
Cordless phones
Pencil sharpener
Toasters
Cups
Pots/pans
Toothbrushes
Dressers
Refrigerator
Television remote controls
Electric fans
Scissors
Televisions
Flashlights
Sewing machines
Vacuum cleaners
Glue guns
Shirts
Watches
128


Feature A
Base
31
(++) Prototypical Handset, Prototypical Base
(+-) Prototypical Handset, Atypical Base
(-+) Atypical Handset, Prototypical Base
(--) Atypical Handset, Atypical Base
FIGURE III-1
EXAMPLE OF UNITY AND PROTOTYPICALITY MANIPULATIONS


Please indicate your reaction to the appearance of the proposed
products by marking (circling) the scales below each.
Hair-drvers
Dislike
5 6
7 8
9
Like
Dislike
*1 -S
L'giy
: : -
5 6
7 S
0
Beautiful
Ugiv
*1 -
Familiar
: 3 -
5
7 8
9
Unfamiliar
Familiar ]
: 3
High Quaiirv
2 3a
5 6
7 8
9
Low Quality
High Quality
2 3
Low Pnce
2 3 4
5 6
7 8
9
High Price
Low Pnce
: ?
4
%
u
- 5 6 x 0 Like
- 5 6 $ 9 Bcauiuui
9 5 6 8 9 Unfamiliar
4 5 6 ** 8 9 Low Quality
- 5 6 7 8 9 High Pnce
Dislike
^ ^ 4
5
6
7
S
9
Like
Dislike
-i
3 4
5
6
' 3
9
Like
Ugly
2 3-i
5
6
7
8
9
Beautiful
Ugly
-i
3 4
5
6
7 8
9
Beautiful
Familiar
2 3 4
5
6
7
8
9
Unfamiliar
Familiar 1
->
3 4
5
6
7 8
9
Unfamiliar
High Quality
: 3 4
5

7
8
9
Low Quality
High Quality
2
3 4
5
6
7 8
9
Low Quality
Low Price
2 3 4
5
6
7
8
9
High Price
Low Price
3 4
5
6
7 8
9
High Pnce
148


106
Experimental Design
The design of this experiment may be thought of as being a 2(product
description: weak vs. strong) x 3(product appearance: Unity/Prototypical, Non-
unity/Atypical, Unity/Atypical) x 9(Products). The comparison of interest,
however, involves the differences between the three product appearance
conditions with the weak product description. Since it is the differences in the
ratings received by the products that were rated against (i. e., along with) the
control product with its more favorable product description that are of interest,
the study is actually a Ktarget product with weak product description) x
3(product appearance: Unity/Prototypical, Non-unity/Atypical, Unity/Atypical)
x 9 (Products) mixed design. The basic design of this experiment is shown in
Figure V 1. Product appearance was a between-subjects factor on any given
product but was balanced across each of the three questionnaire versions that
were administered. Product was a within-subjects factor since each subject did
rate one of the three possible a product pairings (control target) from each
stimulus set. The dependent measure was either a scale measuring quality
(High Quality/Low Quality) or a scale measuring general attitude
(Like/Dislike). Subjects received only one of the two scales for all of the
products that they rated. (This was to insure that there was no carry-over of
one scale to the other). The position of the target and control products on each
13
Inclusion of the bathroom scale stimulus set simplified programming
for the statistical analysis of this experiment.