Visual Mental Imagery

DARREN W. DAHL, AMITAVA CHATTOPADHYAY, and GERALD J.

In this research, the authors seek to advance the understanding of howmarketing can facilitate the new product design process. They focus onhow designers' use of a specific cognitive process, visual mental imagery,can influence the customer appeal of a design. The authors present aconceptual framework for examining how visual imagery might infiuencethe customer appeal of a design output. This is followed by two experi-ments that test the hypotheses that flow from the proposed model. Theexpehments manipulate the type of visual imagery used and the incorpo-raticn of the customer in the imagery invoked and then examine its effectson the usefulness, originality, and customer appeal of the resultingdesign. Consistent with the framework and the proposed hypotheses, thefindings show that including the customer in imagination visual imageryduring the design process has a greater effect on the usefulness of thedesign produced than including the customer in memory visual imagery.The results also show that imagery based on imagination results in moreoriginal designs than imagery based on memory. Most important, the useof bounded imagination, which results from the incorporation of the visu-al images of the customer in imagination imagery, leads to the creation ofdesigns that are more appealing to the customer. The findings are inte-grated into a discussion that clarifies the role of visual imagery in designand underscores the potential of this cognitive tool in the new product

design process.

The Use of Visual Mental Imagery in NewProduct Design

The failure rate for new products has remained steady at35% to 45% for more than 25 years (BusinessWeek 1993:Crawford 1987). Recent research in marketing (e.g.. Windand Mahajan 1997) has sought lo address this problem byhighlighting the need for new marketing-orientedapproaches. Research indicale.s that, to be successful., newproducts must offer something novel to the customer (An-gelmar 1990; Cooper and Klcinschmidt 1987). For example,the innovative Reebok "pump" design that inflates small

'Durren W. Dahl is Assislani Professor ot" Marketing. Faculty ofManagcmcnl. Universiiy of Maniiobu (e-mail: [email protected]),Amitava ChaUopadhyiiy is Ms'.iciitte Professor of Murketmg. University ofBritish Columbia (e-mail: chatKi^ inierchaiige.ubcxa), Gerald J. Otirn isProle.s.wr of Marketing. Hong Kong University of Science and Technology,und A.dvisOT> Council Prokssor ol ConsumeT Beliuvior. Depanmenl ol"Marketing. University of British Columbia (e-mail: n)[email protected]). The authors thank Joe Alba. Peier Darke. Gita Johar. JaideepSengupta. Charles Weinberg. the three JMH reviewers, and ihe ediior ibriheir helpful comments and suggestions. The financial suppon fromR.G.C.. Hong Kong, and ihe Social Sciences and Humanities ResearchCounclt of Canada is gratefully acknowledged. To interact with your col-leagues on speciHc articles in this issue, see "Feedback" on thti JMR Website at www.ama.org/pubs/jnir.

pockets of air in the heel of basketball sneakers has led tomore than $200 million in revenues (Dumaine 199!). Theproduct's originality helped it .stand out from the crowd andcreated a competitive advantage (Kotler and Rath 1984).Originality, however, is not the only criterion for new prod-uct success, as is evidenced by the recent, highly publicizedfailure of the Newton personal digital assistant (PDA) fromApple Computer Inc. (Elstrom 1996). Newton was the firstPDA to recognize handwritten characters and achieved ahigh level of recognition accuracy for single characters. Al-though Apple regarded this recognition ability as superior(Weiman and Moran 1992), it proved inadequate in the eyesof the customer (Johnson 1993; Orr 1993). Even assuming a97% level of recognition accuracy at the character level, aseven-digit telephone number would be processed incorrect-ly approximately 20% of the time, with longer entries(names with addresses and phone numbers) incorrect mostof the time.' Therefore, to appeal to the customer and besuccessful, a new product must not only be original, as theNewton undoubtedly was, hut also useful, that is, success-

^The trade press reporled word accuracy levels of 70% to 80% (On1993).

Jinirnul of Morkelinf' Rt'secirchVol. XXXVI (February 1999), 18-28 18

Visual Mental Imagery 19

fully meet some important need(s) of the customer (Baxter1995; Pugh 1996), something the Newton failed to do.

An important step in the new product developmentprocess is product design (Urban and Hauser 1993). The de-velopment ot a product design concept commonly is viewedas one of the most important phases of the product develop-ment process, because the decisions made at this stagestrongly bear on ail subsequent phases of product develop-ment (Roozenburg and Eekels 1995). When designing newproducts, the goal is to give the product physical and psy-chological attributes that will lead to success in the market-place (Oakley 1990: Urban and Hauser 1993). We examineone way in which the design process can be managed lo gen-erate designs that will appeal to the customer, ln particular,we focus on visual imagery, one of the central cognitive in-puts for the design process (Lorenz 1990; Roozenburg andEekels 1995).

We also focus on a particular design mission for a con-sumer product. As Bahrami and Dagli (1994) note, the mis-sion given to designers often varies. For example, Ulrichand Eppinger (1995) distinguish between a designer's nar-rower mission of trying to build a better pneumatic nail toolfor fastening wood, targeted at the high-end consumer mar-ket, versus a broader one of trying to find better ways to fas-ten wood together. In this research, we focus on the formertype of design situation, in which the designer is asked for aproduct improvement that will meet the needs of a particu-lar set of customers. Furthermore, we focus on functionalperformance in product design, as opposed to the product'saesthetic qualitie.s or appearance.

Product designers are encouraged to use visual mentalimagery when eoming up with their proposed designs, withdesign courses explicitly training them in visualization (cf.French 1994). Downing (1992, p. 442) notes, however, that"very little research has heen conducted that directly ex-plores the nature of the mental image and the implicationsfor its use in de.sign." We take as our starting point the no-tion that visual Imagery leads to improved de.signs and ex-pand on it in two ways. We examine the effects on designoutput of U) two types of visual mental imagery, memoryand imagination, and (2) whether the visualization containsimages of the target customer.

From a theoretical perspective, our research extends theexisting marketing and design literature on the role of im-agery. First, though prior research in marketing has ac-knowledged the distinction between memory and imagina-tion imagery (Childers and Houston 1983: Ellen and Bone1991), it has not examined the implications of this distinc-tion. By showing that these different types of imager>' canhave different impacts on the design outcome, we show thevalue of this distinction in the context of new product de-velopment. Second, our investigation of customer incorpo-ration in the visualization used during the design process isnew to the marketing and design literature. Focusing on thecustomer is central to the marketing concept. Our researchsupports and extends this basic premise by specificallystudying how visualization of the customer during the de-sign process can intluence the resulting design output.

To acquire a better understanding and confirm our intu-ition regarding the use of visual memal imagery in design,in-depth interviews were conducted with product designersfrom ten randomly selected product design firms of differ-

ent sizes who were experienced in the design of consumerproducts. In each of the ten interviews, the designer identi-fied the importance of using visual mental imagery in thedesign process. Their responses suggested that they usedmemory imagery as their main idea source, because they re-lied on existing products in the marketplace, product ideasconveyed to them by their clients, and prior experience withsimilar products. Only one of the designers indicated that heliked to look at unrelated product forms to spark new ideas.

Ail of the designers indicated the importance of identify-ing and understanding the end user, though most believedthat there should be improvement in this area. They notedthai putting the cuslomer at the center of ihe design processwas still not a norm throughout the industry and that clientsoften were unwilling to put resources toward this goal.When asked about their visualization, none of the designersindicated that he or she used images of the customer as atool to enhance the design effort. However, three of the tendesigners did indicate that they visualized themselves as thecustomer when designing. Therefore, though designers rec-ognize the importance of visual imagery, no explicit distinc-tion is made between memory and imagination visual im-agery, a distinction we believe is important and that westudy. Furthermore, designers recognize the importance of acustomer focus but believed that, at that point, they did notfoeus adequately on the customer or include the eustomer inthe visualization process, something we examine in this re-search. Thus, our research assesses issues of relevance toboth marketing and design practitioners.

ln the next section, we diseuss relevant research on visu-al mental imagery in the design, marketing, and psychologyliterature, present a conceptual model of how visual mentalimagery influences the customer appeal of the product de-signed, and propose a set of hypotheses. We then present anexperiment that tests our hypotheses. The generalizabiIity ofour findings are tested in a follow-up experiment. We con-clude with a di.scussion of the results and directions for fur-ther research.

CONCEPTUAL MODEL

Visual Imagery- as a Design Tool

Visual mental imagery. Maelnnis and Price (1987. p. 473)define visual mental imagery as "a process by which visualinformation is represented in working memory." Visualmental images are functional, quasi-pictorial representa-tions, the special properties of which can influence cognitiveprocessing. Specifically, they enable the generation, inter-pretation, and manipulation of information through spatialrepresentation. Their importance has been recognized inmarketing (e.g.. Maelnnis and Price 1987). engineeringdesign (e.g., Freneh 1994). and other areas as well, includ-ing art design, industrial design, and architecture.

Types of visual mental imagery. Although reterence tovisual imagery is made in the design literature, no fonnaldistinction is made between different types of visual mentalimagery. In marketing, this distinction has been made, andsome seholars (e.g.. Childers and Houston 1983; Ellen andBone 1991) have gone a step further and referencedRichardson's (1969) taxonomy of types of imagery from thepsychological Uleralure. However, their empirical researchfocuses only on one of these types of imagery, namely,memory imagery.

20 JOURNAL OF MARKETING RESEARCH, FEBRUARY 1999

The identification of different types of visual mentalimagery has a rich research tradition i.n psychology. Animportant early distinction in imagery type (e.g., Ogden1913; Perky 1910) that also is reflected in more recentresearch (e.g., Horowitz 1983; Kosslyn 1994; Richardson1969, 1983) is the distinction between images based onimagination and those based on memory. Memory imagesrefer to events or occasions that have been personally expe-rienced or observed (Perky 1910). An example of a visualmemory image is the creation of a visual image of the eon-tents of the breakla.st a person ate yesterday. An imaginationimage differs from a memory image in that, instead ofrecalling a prior experience for the image, a new, never-before-experienced event is constructed (Perky 1910). Forexample, a breakfast visual image could be imagined inwhich the contents of the breakfast were all colored blue.Note that the imagination image also involves prior memo-ry (i.e.. breakfast and the color blue): however, it recom-bines these memory images in a new and previously unseenway.2 Differentiating between types of imagery has beenfound to be important in leaming, mood and affeet. neu-ropsychology. and problem-solving (e.g., Adeyemo 1990,1994). an area related to this research.

The distinction between the use of memory and imagina-tion imagery is important in a product design context aswell. By definition, imagination imagery is novel, because itinvolves the creation of previously unseen images. Adeye-mo (1990) has demonstrated the capacity of imagination im-agery to facilitate innovative problem solutions. Thiscapacity would stem from a broadened solution space whenimagination imagery is used. In contrast, memory imageryis eonfmed to previous experience. Although images basedin memory can provide a reference point for design (Down-ing 1992), a reliance on memory imagery might limit thefull potential of visualization.

The importance of the customer. A eustomer orientation iscentral lo marketing (cf. BlaUberg and Deighlon 1996;Deshpande, Fariey, and Webster 1993). Marketers have de-veloped many approaches for bringing eustomer needs andwants to the attention of the designer (e.g.. Griffin andHauser 1993: Urban and Von Hippei 1988; Von Htppel1978, 1986). Although this has led to the successful devel-opment of many new produets, as Baiietti and Litva (1995,p. 3) note, "despiie all best efforts, ihe design process oftenleads to the introduction of produets that do not meet cus-tomer expectations."

One possibility is that, when designing products, design-ers are not always able to incorporate information about thecu.stomer effectively. Rifkin (1994, p. 10: see also Leonardand Rayport 1997) has .suggested a move toward "empathicdesign," in which the voice of the cuslomer "comes out ofunderstanding users' needs through empathy with theirworld." Empathy might result from incorporating the cus-tomer in the content of the visual mental imagery used whendesigning a product. Incorporating the customer in designimagery refers to the actual imaging of a customer with theproduct. This form of imaging could involve images of thecuslomer as an "actor." using product prototypes, interacting

-Morris and Hampson (I983| note ihul memory and imaginaiion can beput on a conlinuum. A( imf end. there are pure memory images. At the otherend. though memory is involved, it serves as a springboard tur the imagi-nalive constructions that predominate.

with specific aspects of a design, or using previous designson which the designer is attempting to build. We proposethat including the eustomer in the visual mental imageryused during the product design process affects the designoutput. Furthermore, we propose that the impact of includ-ing the customer in the visualization depends on the type ofvisual mental imagery invoked. These propositions aredeveloped more fully in the next section.

The Design Output: Usefulness, Originality, and Appeal

Thus far, we have concentrated on two major inputs forthe design process, the type of visual imagery that is used bythe designer (memory versus imagination) and the contentof the imagery (whether the customer is incorporated).These inputs to the design process cannot, however, be dis-cussed in isolation. To establish a context for their role inproduct design, we must examine their impact on the out-puts of the design process.

The central goal of the product design process is to createproducts that appeal to the end user (Bloch 1995; Kotler andRath 1984). The question then becomes. On what factorsdoes customer appeal depend? The literature on new prod-ucl success suggests that a produci must have a competitiveadvantage to appeal to the customer. Consistent with the lit-erature, we define competitive advantage as the ability toprovide something new, original, and useful to the eustomer(Angelmar 1990; Cooper and Kleinschmidt 1987, 1990;Davidson 1976; Song and Parry 1997). In this research,which deals with concept design, we foeus on originalityand usefulness as drivers of customer appeal. We recognizethat at later stages (e.g., test marketing), other variables alsowould be important, such as economic cost. Figure 1 pre-sents our conceptual model of how imagery type and con-tent influence the usefulness, originality, and customer ap-peal of a design.

Design usefulnes.'i. Compared with memory imagery.imagination and its playful imagery manipulations are like-ly to lead to a numerically large and diverse set of imagesand to more original designs, as we predict subsequentlywith our second hypothesis (e.g.. Finke, Ward, and Smith1992; Roskos-Ewoldsen 1993). However, a free and uncon-

Figure 1CONCEPTUAL FRAMEWORK OF HOW CUSTOMER

INCORPORATION AND IMAGERY TYPE DRIVE CUSTOMER

APPEAL

*The solid lines repre ieni direci relationships. The effects of customerincorporation and imagery type on cuslomer appeal are mediated by use-fulness and originality. We use a dotted line to represent ihis relationship.


strained flight of imagination imagery also could lead to fan-ciful design ideas that have little relevance to the customer.

Putting the eustomer into the design process imageryshould ensure that all design activity remains centered onthe customer. Previous discussions of eustomer involvementin design (e.g., Baiietti and Litva 1995; Palmiter et aL 1994)have pointed to a need for designers to "keep the eustomerin their thoughts as they design ... the product" (Palmiter etal. 1994, p. 129). Including the eustomer in imaginationimagery constrains the solution space by providing appro-priate boundaries to focus the designer's imagination. Thisuse of bounded imagination brings realism into the designprocess by providing relevant guidelines for the imaginationimaget^. The resulting design process should be able to ful-fill customer requirements better and create a product that isuseful and appealing to the end user (Baiietti and Litva1995).

We anticipate that customer incorporation in memory im-agery will have less of an effect. Previous research in designhas shown that relying on memory during the design processmay lead to suboptimal outcomes (Jansson and Smith 1991:see also Smith, Ward, and Sehumaeher 1993). Jansson andSmith (1991) asked subjects to come up with new productdesigns, giving the subjects in what they call the fixationgroup some example soiutions. They find that, relative to thecontrol group, those given example solutions tended to fix-ate on the examples and .stuck closely to them in the designsthey created. We expect that memory forms of imagery willcreate a fixation point. Including the customer in what is al-ready a more constrained solution space should have less ofan effect on the usefulness of the design output. Also, de-signs in tbe marketplace are likely to have at least some use-fulness, or they would not likely be on the market. There-fore, even without instructions to visualize the customer, theproduct designs retrieved by designers from memory arelikely to have some customer utility, thus reducing the seopefor improvement through eustomer incorporation in the vi-sualization. This, of course, would be more true for estab-lished products that have been in the marketplace for a longtime than for really new products. Here, we foeus on well-established products. We hypothesize the following:

H|: Including the customer in imagination visual itnagery dur-ing the product design process will have a greater effect onthe usefulness of the designs produced than including thecustomer in memory visual imagery.

Design originality. As we discussed previously, imagina-tion imagery is likely to lead to a more diverse set of images.This should lead to more original designs. In support of thisidea, Adeyemo (1990) shows that, in a creative problem-solving exercise, the use of imagination imagery led subjectsID develop more creative and original solutions to ihe exper-imental problem. In contrast, the dependence of memoryimagery on previous experiences constrains playful imagery'manipulations and reduces the diversity of images generatedand. therefore, the originality of design solutions (Janssonand Smith 1991). Thus, we hypothesize the following:

Hi: When de.signing a product, the use of imagination visual im-agery, compared with the use of memory vi.sual imagery,win re.sult in product designs that are more original.

Previous research does not provide a elear picture of howincorporating the eustomer in the visual imagery used in thedesign will affect the originality of the design output.Therefore, we offer no formal hypotheses regarding thispotential relationship, though we examine such effectsempirically.

Creating appealing designs. The centrai goal of the prod-uct design process is to create products that appeai to the enduser (Kotier and Rath 1984). Aceording to our conceptual-ization, at the early stages of the product developmentprocess, customer appeal depends on the usefuiness andoriginality of the product design (Cooper and Kleinschmidt1987. 1990; see also Figure I). Products that are useful tothe customer, that is, that satisfy their needs, are more ap-pealing (Pugh 1996). Products that are original will be moreliked, but only if they are useful as well. As the case of Ap-ple's Newton illustrates, products that are original but notuseful are not likely to appeai to the customer. However, aproduct thai is useful in addition to being original should beweil liked (Cooper and Kleinschmidt i987). Our conceptu-aiization aiso suggests that originality and usefulness, intum, depend on the type of imagery and whether the cus-tomer is incorporated in the visuaiization (see Figure 1, H|,and Hi). Taken together, this suggests that customer appealdepends on the interaction of imagery type and customer in-corporation and that this effeet is mediated by the usefulnessand originality of the design. More specifically,

Hv The most appealing designs will be produced when hothimagination imagery i.s u.sed and the customer i.s incorporat-ed in the imagery.

H4;,: The effects of imagery type and customer incorporationon customer appeal will be mediated by a direct etfect atusefulness.

H41,: The effects of imagery type and customer mcorporation oncustomer appeal will be mediated by an interactive effect ofusefulness and originality.

EXPERIMENT 1

Two experimentai factors (imagery type and customerincorporation in visualization) were manipulated in abetween-subjects design. There were two levels of imagerytype (memory and imagination). Because previous researchhas indicated that it is difficult for the designer to focus onthe customer (e.g.. Baiietti and Litva 1995), we used twooperational manipulations of customer incorporation. Thus,there were three levels of eustomer incorporation (no cus-tomer incorporation, customer incorporation instruction,and customer incorporation instruction plus training). A sin-gle control condition that received neither manipulation alsowas included. Thus, the design had seven cells ([2 imagerytype X 3 customer incorporation] + 1 control).

Product Designers

Designers pariicipating in the study were 140 undergrad-uate engineering students. They all had completed course-work in design fundamentals and were recruited in designengineering classes through announcements offering $10 forparticipation. Previous research in product design has usedengineering students (e.g., Jansson and Smith 1991).


Product

The product chosen for the design task was a car jack forseniors (age 60 years or older). Seniors were chosen as thetarget customers because they represented a population thatwas distinctly different than the designers. This was done torepresent the real-life situation in which designers oftendesign products for customers different than themselves. Onthe basis of informal discussions with seniors, a car jack waschosen as the product for our study, because seniors report-ed having difficulty using those jacks currently available.

Independent Variables

Manipulation of imagery type. Visual mental imagerytype was manipulated by instructing designers to rely onimages based on either imagination or memory during thedesign task. These instructions were developed throughpretesting. In the memory imagery condition, designerswere instructed as follows:

In solving design problems, many designers find thatusing past memories to form visual images (pictures inthe mind) of potential designs can help tbem produceinnovative and effective designs. Digging deep intoyour memory and visualizing car jacks thai you havepreviously seen may help in the development of designsolutions. In producing your design, please try to usethis type of visual strategy.

In the imagination imagery condition, designers wereinstructed to go beyond previously seen images, as follows:

In solving design problem.s, many designers find ihai us-ing imagination to form visual images (pictures in themind) of potential designs can help them to produce in-novative and effective design.s. Using imagery that goesbeyond car jacks that you would normLilly see can helpin tbe development of design solutions. !n producingyour design, please try to use thi.*i type of visual strategy.

Manipulation of customer incorporation. In the no cus-tomer incorporation condition, designers received noInstructions to visualize the customer. As we mentioned pre-viously, two approaches to encourage eustomer incorpora-tion during visualization were used. These were developedon the basis of pretesting, in the first approach, designerswere instructed to incorporate the eustomer in their visual-ization, as follows:

A useful strategy in design involves visualizing the po-tential customer of tbe product. Seeing in your mind anelderly person being involved with and interacting witbtbe proposed product design can assist the developmentof tbe design solution, tn producing your design, pleasetry to use tbis visual strategy.

In the second customer incorporation condition, designerswere given a five-mlnute training task before receivingeither the imagery or customer incorporation instructions.The training exercise used a guided imagery procedure. Thisinvolved the designers reading a short story about a youngadult putting together a television stand. As the designersread the short story, they were instructed to stop after read-ing each sentence in the story and try to picture in theirminds the events about which they had just read. After com-pleting the training exercise, designers were given the

appropriate imagery instructions, depending on the condi-tion to which they had been assigned, as well as the instruc-tions to visualize the customer in the design task.

Control group. The control group was not given any in-structions to use visual imagery, that is, they received nei-ther imagery type nor customer incorporation instructions.Given their academic training, we expected that the design-ers would use visual mental imagery spontaneously and thatthis visual imagery would be primarily memory rather thanimagination imagery. If this tumed out not to be the ease,any differences we might find between the imagination andmemory imagery conditions might be attributed to ourmemory imagery treatment hurting and the imaginationtreatment encouraging the production of more original andappealing designs.

Procedure

Designers participating as subjects were tested individu-ally. The group that received the training task (customerincorporation instruction plus training) received the trainingfirst. Designers in all conditions then received the instruc-tions for the design task. They were instructed to design acar jack for elderly adults (age 60 years or older) and pro-duce a "thumbnail sketch" of the central aspects of theirdesign Idea. They were told to disregard any constraints(e.g., economic, material, regulatory) as they developedtheir idea. On the basis of pretesting, they were given 30minutes to complete the task. Designers in the control con-dition proceeded to the design task at this point. The remain-ing designers received the imagery type (imagination ormemory) treatment. One-third of the subjects in each of theimagery type conditions, who had been assigned to the nocustomer incorporation condilion. prtjceeded to Ihe designtask. The remaining two-thirds received the customer incor-poration instructions, foilowed by the design task. Half ofthese subjects had received the prior training in customerincorporation, a.s we noted previously. Note that the trainingtask was given before the main task. This precluded the pos-sibility that, during training, .subjects would be thinking ofthe design they would later have to create, leading to themhaving more experience with designing car jacks than thosein the other conditions. After compieting the design task.designers filled out a short questionnaire, were debriefed,and were paid for their participation.

Manipulation Check Measures

The questionnaire completed by the designers startedwith an open-ended measure that asked them to describe anymental pictures they had during the design exercise. Theseresponses were used to assess the success of the customerincorporation manipulation. Next, the designers were askedto complete three seven-point scales that sought to deter-mine the nature and form of their visuaiization. Designerswere asked how hard ihey were trying lo use visual imageryin producing their design ("not very hard" [ 1 ] to "ver)' hard"[7]), to what degree they used mental pictures of previously.seen car jacks ("not very much" 11] to "very much" [7]), andto what degree they used mental pictures that went beyonda "normal" ear jack ('"not very much" |1] to "very much"[7]). Finally, they completed an open-ended question thatasked what they thought the purpose of the study was.


Dependent Measures

Target customer fudges. When designing new products,the goal is to appeal to a potential user. Therefore, responsesto the dependent measures were obtained from a sample ofelderly judges who were recruited from local senior citizencenters. To qualify, judges were required to have a valid dri-ver's license and have driven a vehicle in the past year. Aseparate sample of judges evaluated all of the 140 designsfor each of the dependent variables of interest (originality:n = 14 judges, useful: n = 16 judges, appeal: n - 12 judges).Pretesting indicated that this was a reasonable task for anindividual judge to complete; it took between one and a halfand two hours, including breaks. Separate judges were usedfor each dependent variable to eliminate shared methodvariance (see Olney, Holbrook, and Batra i99M.

Twenty booklets were prepared, each with a different ran-dom order of the 140 designs. Judges were randomly assignedone of the booidets. Judges worked individually and wentthrough the booklet al their own speed, rating each designalong the three items that pertained to the construct they wereassigned. They were paid $20 for their participation.

Target customer judgements. Originality, usefuiness, andappeal each were measured by a three-item, seven-point se-mantic differential scale drawn from previous research (An-drews and Smith 1996: Besemer and O'Quin 1986: Kardes1986; Mano and Oliver 1993). The three items composingthe originality scale were "unique" (7) to "ordinary" (1)."original" (7) to "commonplace" (1), and "fresh" (7) to"routine" (I). Items for the usefulness .scale were "useful"(7) to "useless" (1). "effective" (7) Lo "ineffective" (1), and"worthwhile" (7) to "worthless" (I). The eustomer appealscale was composed of "appealing" (7) to "unappealing"(I), "likeahle" (7) lo "not likeable" (U. and "desirable" {1)to "undesirable" (1).

RESULTS

Preliminaiy Analyses

Manipulaticm checks. The type of imagery used duringthe design task was examined with scale measures of mem-ory and imagination image usage. As we noted previously,because memory would be involved in both the memory andimagination conditions, we did not expect any significantdifferences between these conditions on the memory imageusage scale. A two-way ANOVA with imagery type and eus-tomer incorporation as the two independent lactors showedthat there were no significant effects (Fs(l.l34)< 1).- Atwo-way ANOVA for the imagination image usage scale,with imagery type and customer incorporation as the inde-pendent factors, revealed only a main effect for imagerytype (F(l,134) = 8.i7,/>< .01). Designers in the imaginationcondition used imaginative images to a greater extent(means: imagination = 5.1 i. memory - 4.30). For thesemeasures, the controi condition was not significantly differ-ent than the memory condition (/>s > .10).

Open-ended responses, used lo assess the success of thecustomer incorporation manipulation, were coded by a

iFor Ihe ANOVA.-; and plunneti conira.' is. 10 obiain a better estimate ofwjihin-cell error, we use data from the conirol condition in addititin to thedata iram the six experimental conditions (Keppel 1991). Using the errorICTm trom jusi \hc sis cxperimenlal coiidiiions di>es noi chaiigc our icsull.'vsubstanlively.

research assistant to identify the number of distinctly differ-ent images involving the customer that were used by thedesigner. As we expected, a two-way ANOVA indicated asignificant main effect for the customer incorporationmanipulation (F(2,I3O) = 30.37. p < .001). Contrasts acrossthe three levels of eustomer incorporation revealed signifi-cant differences (/JS < .01) between each of the three condi-tions (means: no eustomer incorporation = .59. eustomerincorporation instruction = 1.50. customer incorporationinstruction plus training = 2.69). The control condition didnot differ significantly from the condition that received noinstructions to incorporate the customer {p > .10).

It was also considered important to establish that the de-signers across all conditions made an equal effort on the vi-sualization task. The two-way ANOVA on the scale thatassessed how hard the designers tried to use imagery indi-cated that there were no significant differences across theexperimental conditions (all Fs < 1). In addition, the controlcondilion did nol differ from the manipulated conditions inefforts to visualize (p> .\0)-

Finally. designers' responses to the question related to theperceived purpose of the study were examined. No design-ers were aware of the experimental hypotheses, and nonesuspected that the study had anything to do with differencesin the nature of visualization during the design process.

Reliability assessment of judges' ratings. Coefficient al-pha was used to assess reliability. Following Olney. Hol-brook, and Batra (1991). reliability was assessed for indi-vidual items across judges, scale indices within judges,within-judge scale indices across all judges, and the aggre-gate scale index composed of the mean of all judges on eachitem.

The alphas for single-item interjudge reliability on eachof the nine scale items ranged between .70 and .77. Scaleindices for each of the 42 individual judges ranged from .75to i .00. The index formed by summing the three-scale itemsforeaeh construct within each judge and then calculating theinterjudge reliability on that sum produced reliabilities of.75 for original, .78 for useful, and .71 for appeal. Finally.the aggregate index composed of the average of the scaleindices across all the judges was .98, 1.00. and .99 for theoriginal, useful, and appeai items, respectively.

To assess the validity of the three construct measures, themean scores across the judges tor each of the nine items thatmake up the three constructs were subjected to principaicomponent analysis, foliowed by varimax rotation. As weexpected, a three-factor solution was obtained (variance ex-piained = 98%). The three items measuring each constructloaded on a .separate factor (all loadings > .91). Follow-upconfirmatory factor anaiysis revealed that the three-factormodel provided an acceptabie fit (Bentler-Bonett nonned fitindex [NFl] = .964. comparative tit index [CFIj ^ .973, X"(25) - 99.85. p < .OOi), because the fit indices exceeded .90,the benchmark for acceptable fit (Bentler and Bonett 1980).In addition, neither the single-factor model (NFI = .421. x-(27) = 1629.78. p < .001). nor a two-factor modei, in whichthe items for the usefulness and customer appeal constructswere combined, fit the data weii (NFI = .641, x- f-6) =1010.26. p < .001). For the hypotheses tested subsequently.the mean score across the three ilems measuring each con-struct was used as the operationai measure.


Memory versus control. As we noted previousiy, those inthe controi condition u.sed the same type of imagery as thosein the memory condition (see the manipulation cheek re-sults). This should result in the designs produced being sim-ilar to those produced in the memory condition in terms oforiginality, usefulness, and customer appeal. A series ofplanned contrasts compared those in the control conditionwith those in the memory imagery, no customer incorpora-tion condition. The results showed that the designs produceddid not differ in terms of originality (F(l,138) < I. means:control = 3.96, memory = 3.82). usefuiness (F(i,I38) < 1,means: control = 3.10. memory = 3.12), and customer ap-peal (F(I.I38) < 1, means: control = 3.32, memory = 3.27).Thus, the findings with respect to the main hypotheses aredriven by the superiority of providing explicit instructions touse imagination imagery, not because instructions to usememory imagery placed an artificial constraint on what de-signers intuitively might do.

Test of Hypotheses

In this section, tests of the four hypotheses are presented.Mean scores for each dependent variable, along with stan-dard deviations and cell sizes, appear in Table 1.

H/: Effect of customer incorporation and imagery type cmusefulness. Consistent with H,, the ANOVA results for theusefulness variable revealed only a significant interactioneffeet (F(2.134) = 3.03. p < .05. co .032). In addition, aplanned contrast indicated that the designs produced in theimagination imagery, customer incorporation conditions(mean = 3.14) were rated as more useful than the designsproduced in the imagination imagery, no customer incorpo-

ration condition (F(l,138) = 6.33, p < .01, mean = 2.64). Acontrast between the designs in the memory imagery, cus-tomer incorporation condition (mean = 3.02) and the mem-ory imagery, no customer incorporation condition (mean =3.12) was not significant (F(i,138) < 1), Thus, the data areconsistent with H[: customer incorporation enhanced theusefulness of the designs more when imagination imagerywas used than when memory imagery was u.sed. includingthe customer in memory visualization did not significantlychange the outcome.

It is also noteworthy that, in the imagination imagery,customer incorporation conditions, subjects who receivedinstruction and training produced designs that were judgedthe most useful (planned contrast: F( 1,138) = 3.60,;? < .05;means: imagination customer incorporation instruc-tion = 2.92, imagination customer incorporation instructionplus training = 3.35).

M2: Effect of imagery type on originality. Hi states thatdesigners in the imagination imagery condition will producemore original product designs than those in the memory im-agery condition. Consistent with H2, the ANOVA results fororiginality reveal only a significant main effect for imagerytype iF(l.l34) = 23.35. p < .001, oP = .159: means: memo-ry imagery = 3.81. imagination imagery = 4.40).

Hj: Effect of customer incorporation and imagery type onappeal. The ANOVA results for the appeal variable revealedonly an interaction effect (F(2.134) = 3.22. p < .05, co- =.035). In support of H3. a planned contrast showed that thecustomer appeai of the designs produced by those in theimagination imagery, customer incorporation conditions(mean - 3.74) was greater than the customer appeal of the

Table 1EXPERIMENT 1: TABLE OF MEANS

DepeiideniVariable

Usefulness

Originality

Customer Appeai

Imagery TvpeCondition

Control

Memory

Imagination

Control

Memory

Imagination

Ccmirol

Memory

imagination

Conirol

3-10

t .61H

3,96{.69]

3.32{-69]

Customer Incorporation Condition

No customerincorporation

3.12(•8212.64

3.82I-46J4.49i.741

3.27{•6713.03

!I . I61

Customerincorporation

3.02 (2-99. 3.04)"{,68] {.73, .64]3.14 (2,92.3.35)1,771 1.64. ,841

3,81 (3.80, 3.82}{.80] {.89. .71]4.36 (4.48.4.23){.52] {-50. -52]

3.22 (3-12.3.31){.72| {.76. .69]3.74 (3.82. 3.651{.861 {.80. .93]

Note: Cull si^es are n = 20.JThe standard deviation for each reported mean is contained in brackets.

e ins ibt itie cusiomev vncovporaiion insirucsion and cusiomer incorporaiioit instruction pius Vraitiing are conlaineci in parentheses.

Visuai Mentai Imagery 25

designs for aii the other experimentai eonditions combined(F(i,138)= li.63,p<.001;mean = 3.i9).4

H4a and H4^: The effect of usefulness and originality onappeal. The interaction effect produced hy the manipula-tions of imagery type and customer incorporation on cus-tomer appeai (H3) was hypothesized to be mediated by amain effect of usefulness (H4a) and an interaction betweenoriginaiity and usefuiness (H^y^).'^ According to Baron andKenny (i986), three conditions must be met to infer media-tion: First, the independent variables must have a significanteffeet on the mediators and the ultimate dependent variabie.Second, when the effects of the independent variables on themediator variables are covaried out, the effects of the inde-pendent variables on the dependent variable shouid eitherdiminish or disappear. Third, the eovariates used in theanalysis should be significant.

As we have shown, the independent variables have sig-nificant effects on both the mediators (H| and i42) and oncustomer appeal {Hi,). To satisfy the second and third crite-ria for mediation (Baron and Kenny i986). we needed toconduct an analysis of covariance with imagery type, cus-tomer incorporation, their interaction, and the hypothesizedmediator variables as covariates. When an interaction termis hypothesized as a mediator, a.s in this case. Aiken andWest (1990) recommend that (1) we incorporate all thedirect effects: in our case, this means including the directeffect of originality beyond the hypothesized effect.s and (2)given that all of the effects must be included, the mediatorvariables should be mean centered to reduce multicollinear-ity. Following their recommendations, the raw usefuinessand originaiity scores were mean centered, and the .ANCO-VA model estimated inciuded the direct effect of usefuiness.the direct effect of originality, and their interactive effect ascovariates.

Consistent with the criteria proposed by Baron and Ken-ny (1986). the analy.sis of covariance revealed that the pre-viously significant interaction effect (F(2,I14) = 3.06. p <.05) of imagery type and customer incorporation on cus-lomer appeal disappeared (F(2.1in = 2.32. p > .10). Fur-thermore, usefulness as a covariate was significant(F( 1.1 i 1) = 57.76, /? < .001), as was the interactive effect be-tween usefuiness and originaiity, albeit at -dp < iO ievel(F(l.i l l) = 2.93). Thus, all three criteria required to infermediation are met. providing support for hoth H^.^ and H i .

EXPERIMENT 2

Experiment 2 attempts to test the generalizahiiity of thefindings of the first study. In the first study, we had chosena target segment very different than the designers, namelyseniors. In this study, we wanted to examine whetherdecreasing the difference between the designer and the .seg-ment ibr which they were designing attenuated the effect ofthe eustomer incorporation manipuiation. We seiected a tar-get segment thai was similar to the designers in an important

''When performed separately, the contrasts between this condition andthe other conditions were all significantly ditTereni at the .05 ievel.Moreover, the other conditions did not differ significatilly from oneanother,

sWe te.st for mediation using a sequential approach most commonly u.sedin cxperimenial research, nameiy. the one suggested by Baron and Kenny(19861. An alternative approach would be simultaneous paih analy.sis (e.g.,LISREL).

way, but not exaetiy the same as the designers, because ourinterviews had shown that designers sometimes use imagesof themseives when designing products. Using a target seg-ment identical to the designers would not ailow for us toseparate the effects of including the seif from those ofinciuding the eustomer in the visualization. Therefore, wechose a sample of young women (age 20 to 24 years),because ail the designers in the sample used in this studywere young men. The designers were therefore simiiar to thetarget segment in terms of an important demographic char-acteristic, age, but differed on another, gender, in this study.we aho used a different product from thai used in the firslstudy, a decision driven by a concem for generaiizability.

The study u.sed a 2 x 2 between-subjects factoriai designwith imagery type (memory and imagination) and customerincorporation (no customer incorporation and customer in-corporation) as the two independent factors. Seventy-threedesigners participated in the study. Designers were seniorundergraduate engineering students who had completed de-sign coursework.

Method

As we discussed previously, this study differed from thefirst study in two important ways. First, the product chosenfor the design tasiv was an umbreila. Second, the target cus-tomers were young women. The dependent measures andmanipulation checks used in the second study were similarto those used in the first study, A target sample of judges,composed of young women, was used lo evaluate the pro-duced designs. Simiiar to the first study, a separate sampleof judges was used to evaluate originality (n = i9). useful-ness (n = 19). and customer appeal n = 20).

Results

Preliminary analyses. Manipulation checks indicated suc-cess of both the imagery type (F( 1,69) - 3-17, /? < .05) andthe cuslomer incorporation (F(i,69} = 23.87, p < .001)manipulations. Reliability analysis for the ratings of the cus-tomer judges indicated sati.sfactory reliability foreaeh of themeasured constructs (all alphas > .70). Confirmatory factoranalysis showed that a three-factor model provided a goodfit ibr the data (NFI ^ .918, CFi = .934. X" (25) = i 14.23,p < .001). Consequently, as in the first study, nnean scoresacross the three items were used as the dependent measuresin the analyses. Mean scores for each dependent variable,along with standard deviations and cell sizes, are presentedin Table 2.

Test of hypothe.ses. Consistent with Hi-an ANOVA for theusefulness variable reveaied only an interaction effect(F( 1.69) = 5.75. /) < .05. (O- = .060). A planned contrast indi-cated that the designs produced in the imagination imagery,customer incorporation condition were rated as more useful(mean = 3.5 i) than the designs produced in the imaginationimagery, no customer incorporation condition (F(l,7i) =9.28. p < .01, mean - 2.85). A contrast between the memo-ry imagery, customer incorporation condition (mean = 3.15)and the memory imagery, no customer incorporation condi-tion was not significant (F(l,71) < 1, mean = 3.23). Thu.s.the data again support H|.

In support of Hi. the ANOVA results for originalityrevealed oniy a significant main effect for imagery' type

26 JOURNAL OF MARKETiNG RESEARCH, FEBRUARY 1999

Table 2EXPERIMENT 2: TABLE OF MEANS

Dependent Variable Imagery Type Condilion Cuslomer Incorporation Condition

Usefulness

Originalily

Cu.slomer Appeal

Memory

Memory

Imagination

Memory

Imagination

No customerincorporation

3.23{.64]^'2.85(.761

3.96{-99}4.91{.731

3.42{.6713.17{.79]

Cu.stomerincorporation

3.15{.77j3.51{.42(

3.93(1,27]4.25

11-411

3.30{.72]3.69{.41]

Note, Ceil SV/.CA for Ihe memory conditions and the imagination customer incorporation condition were n - 18; for the imagination condition ii was n -^Tht; standard deviation for each reported mean is contained in brackets.

(F(l,69) = 5.90, p < .05. (o = .062; means: memory ^ 3.94.imagination = 4.59).

The ANOVA results for cuslomer appeal reveaied oniy aninteraction effect (F(i,69) = A.\%, p < .05, aP = .042). Insupport of H3, the planned contrast between the imaginationimagery, cuslomer incorporation condilion (mean = 3.69)and all the other experimental conditions combined was sig-nificant (F( 1.71) = 4.71, /J < .05. mean = 3.30).^

Following the criteria proposed by Baron and Kenny(i986) and Aiken and West (1990). originality and usefui-ness were tested as mediators of the effeet of imagery typeand imagery content on customer appeal. In support of bothH4y and H4h, an analysis of eovariaiice showed that the pre-viously significant interaction effect (F(1.69) =• 4.!8, p <-05) of imagery type and customer incorporation on cus-tomer appeai (H3) disappeared (F(l,66) < i). In addition,usefulness as a covariate was significant (F{i.66) = 164.40.p < .001), as wa.s the interactive effect between usefulnessand originaiity. albeit at a;j < iO level (F(1.66) = 2.8i).

DISCUSSION

The resuits of both studies are consistent and lend supportto the intuition that the lype and content of visual mentalimagery play important roles in Ihe design proces.s for newproducts. The findings support our conceptual frameworkfor how visual mentai imagery can influence reievanta.spects of a design.

Consistent with the proposed hypotheses, the results indi-cate thai including the customer in imagination visual im-agery during the design process has a greater effect on the

''When performed separately, the contrasts between the imagination, cus-uimer iiicorporaiion condition (mean = 3.69). ihe menrory. customer incor-poration condition (mean = 3-31). and the imagination, no customer incor-poration condition (mean = 3.17) were significantly different ai the ,05level. The contrast between the imagination, customer incorporation condi-tion and the memory, no cuslomer incorporation condition (mean = 3,42)was marginally different {p < .10), The other conditions did noi differ sig-riitii:amly from one another.

usefulness of the designs produced than including the cus-tomer in memory visual imagery. Interestingly, the use ofimagination imagery, when it is not focused on the cus-tomer, leads to designs that are les.s usefui than those in anyof the other conditions. The fmdings also show that imagerybased on imagination results in more original designs thanimagery based on memory. Designs produced when visualimages of the customer are incorporated in imagination im-agery also are shown to be more appealing to the customer.The effects on customer appeai are mediated by both theperceived usefulness of the design and its originality.

The effect sizes for the ANOVAs. as refiected by the aPstatistic obtained across both studies, ranged between .032and .159, According to Cohen (1977). a "iarge"' effect pro-duces an OJ- value of .15 or greater. A "medium" effect is.06. and a "smail" effeet is .Oi. Thus, but tor the main effectof originaiily in the first .study, which was large (.i59), aiithe other effect sizes wouid be considered moderate. There-fore, we believe that the variables we studied are not oniyimportant. buV also explain a reasonable portion of the vari-ance In the data.

Our re-seareh has a number of iimitations. Because of ex-perimental approach, we had lo mai>;e choices about the vari-ables to include and were not able to study all the variableson which our resuits might be eontingent. In the product de-sign context, we might expect thai the eharacteristics of thedesigner (e.g.. ievel of experience, ability to visualize. work-Ing individually or in a team), the designer-mari<:el interface(e.g., the similarity between the designer and the targei seg-ment), the product being designed (e.g.. product for an enduser I i.e.. a can openerl versus a component [i.e., a bali bear-ing]), and the nature of the design task (e.g., time provided,scope of design mission) might intluence the reiationshipsobserved in this research. The research reported here usedstudent designers working individualiy on a design conceptfor a consumer product with a short time limit. In our twostudies, the similarity between the designer and the targetmarket varied, as did the product that was designed, but theresulls were the same. This provides some support for the

Visual Mental imagery 27

robustness of our conceptual framework and findings acrossdifferent product design situations.

Our interview.s with designers suggest that they oftenwork individually, design products for a range of target seg-ments, and are given a variety of design missions, includingthe development of design concepts similar to those used inthis research. They aiso work with serious lime constraints,often with less than a day allocated to concept-generationtasks. Therefore, our results should have relevance for atleast some of the situations that designers face. However,more research, varying the factors previously identified, isnecessary to establish the generaiizability of our findings toa broader range of product design contexts.

A further opportunity for research is the identification ofdependent measures that may infiuence customer appeal,other than originality and usefulness. In our research, we fo-cused on concept design, an early stage in the new productdevelopment process. Perhaps al a laier stage in produel de-velopment, other factors {e.g., reliability, economic cost)would be important in generating customer appeal.

This research adds to the marketing and design literaturein several ways. Foremost, we deveiop a theoretical frame-work that identifies two factors, type and content of im-agery, that are important in the eontext of designing newproducts. The distinction between types of imagery has beenmade previousiy (e.g.. Chiiders and Houston i983), but thisis the first attempt lo study different types of imagery em-piricaliy and show the vaiue of the distinction in a market-ing context. The importance of the customer aiso has beenrecognized in the marketing literature (e.g.. Blattberg andDeighton 1996), but. as the innumerable anecdotes aboutmarketing failures indicate, it is difficult to implement a cus-tomer foeus- Our results suggest that the feedback the de-signers obtain from visualization enhances the customerappeal of the products designed. Thus, our research providesa mechanism for keeping the customer front and center dur-ing the design process. Finally, our resuits suggest two ad-ditional important roles for marketers in the design process.First, the marketer must specify the target segment precise-ly, so that the righ! kind of customer can be visualized. Sec-ond, the marketer shouid provide those insights aboutcustomers and their behavior that facilitate the designer'sability to visualize the customer.

RFFERFNCES

A.deyemo. S.A. (19901, "Thinking Imagery and Pnihlem-Solving,"Psychological Studies. 35 (3). 179-90. '

(1994), "IndlviLtual Differetices in Thinking and Problem.Solving." Personality and Individual Differences. 17 (I ).I 17-24 .

Aiken. Leona S. and Stepheti G. West (1990). Muhiple Res-ression:Testing and Inierpreting Interactions. Newbury Park. CA: SagePublications.

Atidrew-*;, Jnnlee and Daniel C. Smith (1996). "In Search of theMarktJting Imagination: Factors Affecting the Creativity of Mar-keting Programs for Mature Products." Journal of MarketingResearch, 33(2). 174-87.

Angeltnar. Reinhard (1990). "Product Innovation: A Tool for Com-petitive Advantage." European Journal of Operational Re-search, 47 (2). 182-S9.

Bahrami. AH and Cihan H. Dagli (1994). "Design Science." in In-telligent Sy.stems in De.sign and Manufacturing. Cihan H. Dagliand Andrew Kusiak. eds. New York: ASME Press. 7-25.

Baiietti. Antonio J. and Paul F. Litva (1995). "Integrating CustomerRequirements into Produci Designs," Journal of Produci Inno-vation Management, 12 (I), 3-15-

Baron. Reuben M. and David A. Kenny (I9S6). "The Modera-tor-Mediaior Variable Distinction in Social Psychological Re-search: Conceptual. Strategic, and Statistical Considerations,"Journal of Personality and Social Psychology. 51 (6). 1173-82.

Baxter. Mike (1995). Product Design. New York; Chapman andHall.

Bentler, P.M. and D-G. Bonett (1980). "Significance Tests andGoodness of Fit in the Anaiysis of Covariance Structures." Psy-chological Bulletin. 8S (3). 588-606.

Besemer. Susan and Karen O'Quin (1986). "Analyzing CreativeProducts: Refinemen! and Test of a Judging Instrument." TheJournal of Creaiire Behavior. 20 (2), 115-26.

Blattberg. Robert C. and John Deighton (1996). "Manage Market-ing by the Customer Equity," Harvard Business Review. 74 (Ju-ly/August). 136^W.

Bloch. Peter H. (1995). "Seeking the Ideal Form: Product Designand Consumer Response," Journal of Marketing. 59 (July).16-29.

BusinessVJeek \ 1993). "Flops." (August 16), 76-S2.Childers, Terry L. and Michael J, Houston (1983), "Imagery Para-

digms for Consumer Research: Alternative Perspectives fromCognitive Psychology." in Advances in Consumer Research,Vol. to, Richard P, Bagoz?.i and Alice M. Tybout, eds. Provo,UT: Association for Consumer Research. 59-64.

Cohen. Jacob (1977), Statistical Power Analysis for the BehavioralSciences. New York: Academic Press,

Cooper, Roben G, and E.J. Kleinschmidt (1987). "New Products:What Separates Winners from Losers'" Journal of Product In-novation Manugemenl. 64 (3t. 169-84.

and — (1990). "New Produci Success Factors: AComparison of 'KiMs' Versus Successes and Failures," R & DManagement, 20 (I). 47-63.

Crawford. C, Merle (1987). "New Product Failure Rates: AReprise." Research Management. 30 (4). 20-24.

Davidson. J,H. (1976). "Why Most New Consutner Brands Fail."Harvard Business Review. 54 (March/.April). I 17-22-

Deshpandti- Rdhit. John U. Farley, and Frederick E- Webster(1993). "Corporiite Culture. Customer Orientation, and inno\a-Uveness in Japanese Firms: A Quadrad Analysis." Journal ofMarketing. 52 (January). 23-36

Downtng. Francis (1992). "Image Banks: Dialogues Between thePast and tbe Future," Environmeni and Behavior. 24 (4),441 -70.

Dumame. Brian (1991). "Design that Sells and Sells and...." F^r-m«c, (March II). 86-94.

Ellen, Pam S. and Paula F Bone (1991). "Measuring Communica-tion-Evoked Imagery Proce.ssing." in Advances in ConsumerResearch. Voi. 18. Rebecca H. Holman and Michael R.Solomon, edS- Provo. UT: Association for Consumer Research.806-12,

Elstrom. Peter (1996). "PD,A May .Always Mean Pretty Darn Aver-age." BusinessWeek. (June 24). 110.

Finke, Ronald. Thomas B. Ward, and Steven M. Smith I 1992).Creative Cogniiion: Theory. Research, and Applications. Cam-bridge. MA: MIT Press.

French, Michael (1994). Invention and Evolution: Design ir Na-ture and Engineering. New York: Cambridge University Press-

Griffin, Abbie and John R, Hauser (1993). "The Voice of the Con->iumcv" Marketing Science. 12(1). 1-25.

Horowitz, Mardi J- (I9S3). Image Eormatton and Psychotherapy.New York: Ja.son Aronson.

Jansson, David G. and Steven Smith (1991), "Destgn Fixation."Design Studies. 12 ( I ) . 3 - 1 1 .

Johnson. Bradley (1993). "Nuts About Newton? Nah!" .4t/A,i,'f. 64(44). 3. 53.

28 JOURNAL OF MARKETiNG RESEARCH, FEBRUARY 1999

Kardes, Frank R. (1986), "Effects of Initial Product Judgements onSubsequent Memory-Based Judgements," Journal of ConsumerResearch. 13 (June). 1-11.

Keppel, Geoffrey (1991), Design and Analysis: A ResearchersHartdhook, 3d ed. Engiewood Cliffs. NJ: Premice Hall.

Kosslyn, Stephen Michael (1994). Image and Brain. Cambridge.MA: MIT Press.

Kotler, Philip and G. Alexander Rath (1984). "Design: A Powerfulbut Neglected Strategic Tool." Journal of Business Strategy, 5(Fall), if^-21.

Leonard, Dorotby and Jeffrey F. Rayport (1997), "Spark Innova-tion Through Empathic Design," Harvard Business Review. 75(November/December). 102-13.

Lorenz, Christopher (1990), The Design Dimension: The NewCompetitive Weapon for Product Strategy and Clobal Market-ing. Catnbridge, MA: Basil Blackwell.

Maelnnis. Deborah J. and Linda L. Price (1987), "The Role of Im-agery tn Information Processing: Review and Extensions," .lour-nai of Consumer Research. 13 (March), 473-91,

Mano. Haim and Richard C. Oliver (1993), "Assessmg the Dimen-sionality and Structure of the Consumption Expeiience: Evalua-tion, Feeling, and S-dlisfaction," Journal of Consumer Re.search,20 (December), 451-66.

Morris. Peter E. and Peter J. Hampson (1983), Imagery and Con-sciousness. New York: Academic Press.

Oakley. Mark (1990), Design Management: A Handbook of Issuesand Methods. Cambridge. MA: Blaekwell Reference.

Ogden, R.M. (1913), "Experimental Criteria for DitfeientialingMemory and Imagination in Projected Visual Images." Psycho-logical Review. 20 (5), 378-4l0!

Olney, Thomas J., Morris B. Holbrook, and Rajeev Batra (1991),"Consumer Responses to Advertising: The Effects of Ad Con-tent, Emotion, and Attitude Toward the Ad on Viewing Time,"Journal of Consumer Research. 17 (March), 440-53.

OiT. Joel N, (1993), "Almost Is Not Enough," Computer-Aided En-gineering. 12 (9), 4 2 ^ 3 .

Palmiter, Susiin. Gene Lynch, Scott Lewis, and Mark Stempski(1994). "Breaking .Away from the Conventional Usability Lab:The Customer-Centered Design Group at Tektronix. Inc.." Be-haviour and Information Technology, 1 3 ( 1 , 2 ) . 12S-3 I.

Perky, C.W. (1910). "An Experimental Study of Imagination,"American Journal of Psychology, 21 (3), 422-52.

Pugh. Stuart (1996). Creating Innovative Products: Using TotalDesign- Don Mills. ON: Addison-Wesley Publishing Company.

Richardson, Alan (1%9). Mental Imagery. New York: Springer(1983), "Imagery: Definition and Types." in Imagery: Cur-

rent Theory. Research and Application, Aness Sheik, ed. NewYork: John Wiley & Sons, 3-42.

Rifkin. Glenn (1994), "Product Development: Empathic DesignHelps Understand Users Better," Harvard Business Review, 72(2). 10-11.

Roozenburg. N.F.M. and J. Eekels (1995), Product Design: Fun-damentals and Methods. Toronto, ON: John Wiley & Sons.

Roikos-Ewoldsen, Beverly ()993). "Discovering Emergent Prop-erties of Images.'" in Imagery. Creativity, and Discovery: A Cog-nitive Perspective, Beverly Roskos-Ewoldsen. Magaret Jean In-tons-Peterson, and Rita E. Anderson, ed.s. Amsterdam. Nether-lands: North-Holland Elsevier 223-54.

Smith. Steven M., Thomas B. Ward, and Jay S. Schumacher(1993). "Constraining Effects of Examples in a Creative Gener-ation Task." Memory and Cognition. 21 (6), 837—45,

Song. X. Michael and Mark E. Parry (1997), "The Determinants ofJapanese New Product Suecesses," Journal of Marketing Re-search. 34 (February), 64-76.

Ulrich. Karl T- and Sieven D. Epptnger (1995), Produci Designand Development. New York: McGraw-Hill.

Urban. Glen L- and John R. Hauser (1993). Design and Marketingof New Products. 2d ed. Engiewood Cliffs. NJ: Prentice Hall.

— and Eric Von Hippei (1988). "Lead User Analyses for theDevelopment of New Industrial Products." Management Sci-ence. 34(5). 569-82.

Von Hippei Eric (1978)- "Sueeessful industTial Producis fromCustomer Idea.s," Journal of Marketing. 42 (11, 39^9.

(1%6), "Lead Users: A Source of Novel Produci Con-cepts." Management Science, 32 (7), 791-805.

Weiman. Liza and Tom Moran (1992), "Newton: A Step Towardthe Future." MacWorld. 9(8). I 29-3 I.

Wind, JetTy and Vijay Mahajan (1997), "Issues and Opportunitiesin New Product Development: An Introduction to the Special Is-sue," Journal of Marketing Re.search, 34 (Febmary), 1-12,

Documents

Visual Mental Imagery