tactile stimulation and consumer response · tactile stimulation and consumer response jacob...

Tactile Stimulation and Consumer Response

JACOB HORNIK*

Tactile behavior is a basic communication form as well as an expression of inter-personal involvement. This articie presents three studies offering evidence for thepositive role of casual interpersonal touch on consumer behavior. iVIore specifically,it provides initial support for the view that tactile stimulation in various consumerbehavior situations enhances the positive feeling for and evaluation of both theexternal stimuli and the touching source. Further, customers touched by a requestertend to comply more than customers in no-touch conditions. Implications for con-sumer behavior theory and research are discussed.

T he scene was a university library. It could just aseasily have been the local supermarket, bank, or

restaurant. What happened took half a second, and itwas not noticed by the recipients. Remarkably, how-ever, it influenced the recipients' evaluation of the li-brary. Fisher, Rytting and Heslin (1976) investigatedthe effects of a brief "accidental" touch in a nonintimatecontext. They had male and female clerks return librarycards to some students by placing their hands directlyover the students' palms, making physical contact; otherstudents were not touched. Outside the library, studentswere given an instrument to measure their evaluationof the library. Students who were touched, especiallythe women, evaluated the library significantly more fa-vorably than those who were not touched.

The study of touch as a nonverbal communicationis emerging as an important topic in the social sciences(see Hall and Veccia [1990] and Heslin and Alper [1983]for a review). Some investigators suggest that half ormore of the variability of response in interpersonalcommunication can be attributed to nonverbal factorssuch as touch (Mehrabian 1981). Surprisingly, researchon the role of touch on consumer behavior and its im-plication for consumer research is extremely rare.

This article describes an initial exploration of thissubject in consumer research by first presenting sometheoretical considerations linking touch to consumerbehavior. Second, the article extends a recent pilot ex-periment and introduces two studies designed to ex-amine the influence of perfunctory touch in commonconsumer situations, on consumers' evaluations of thesituation, on their attitude toward the toucher, and ontheir response to a request made by a touching source.

•Jacob Hornik is on the Marketing Faculty at the Leon RecanatiGraduate School of Business Administration, Tel Aviv University,Tel Aviv, Israel and a visiting professor of marketing at the GraduateSchool of Business, University of Chicago. The author would like tothank the three JCR reviewers and Kent Monroe for their valuablesuggestions throughout the review process.

449

Third, by drawing on the research results, theoreticaland practical implications for consumer behavior com-munication and research are made. Finally, further in-vestigations into the role of touch on consumer behaviorand new avenues for consumer research are suggested.

THEORETICAL FRAMEWORKThe available information on the meaning of touch

is scarce and ambiguous. The following discussion rep-resents early ground-breaking efforts. "The inherentambiguity of the message of touch limits the precisionwith which it can be described, and to date, no com-prehensive theory of tactile communication exists"(Major 1981, p. 16). Based on a recent rigorous ex-amination of the literature on touch and behavior. Rose(1990, p. 315) claims that "although touch clearly hasimportant functions it is not clear that it plays any sortof one unique role in interpersonal communication."The ambiguity of touch in interpersonal behavior hasalso been recognized by Argyle (1988) and Johnson andEdwards (1991).

Traditionally, touch has been seen as synonymouswith affection and warmth. Montagu (1971) suggestedthat touch and love are indivisible; Mehrabian (1981)equated touch with friendship and warmth; Heslin andAlper (1983) associated touch with involvement andgaining attention, while Jones and Yarbrough (1985)made an empirical attempt to study the meaning oftouch as reported by recipients. Their results reveal 12distinct and, they claim, relatively unambiguous mean-ings such as appreciation, soliciting support, gainingattention, greeting, affection, compliance, sexual inter-est, aggression, departure, announcing a response,friendship, and hybrid meanings.

Most of the literature on touch is based on the as-sumption that the individual reaction to touch is formedin childhood. Reite (1990), for example, suggested thatthe association of touch with stress reduction at an earlyage may result in a positive reaction to being touched

© 1992 by JOURNAL OF CONSUMER RESEARCH, Inc. • Vol. 19 • December 1992All rights reserved. 0093-5301/93/1903-0011$2.00

450 JOURNAL OF CONSUMER RESEARCH

for most individuals, assuming that the touch and bodyareas involved are appropriate for the setting.

The behavioral response to touch will depend on thetouching situation or context as well as on the sex ofthe source and/or the recipient. Research has shownthat people are more likely to comply with requests fromtouching than from nontouching solicitors when thesolicited response is valid (Willis and Hamm 1980). Anexamination of the literature makes it apparent thattouch has different influences on men and women.Generally, women respond more positively than mento being touched (Stier and Hall 1984). However, it isnot clear from previous research whether a casual touchby a male or female stranger also has a stronger positiveinfluence on females and, if so, under what circum-stances.

Touch and Buyer Behavior

Shopping is a frequent personal behavior. However,many consumers view shopping as a diversion fromdaily routines, a means of self-gratification, a source ofsensory stimulation, and a way of socializing. It is wellestablished in the literature that shopping is a socialexperience outside the home that includes social en-counters with salespeople. Stone (1954) even arguedthat consumers form strong personal attachments tostore employees as a substitute for social contact. Assael(1987) shows that salespeople's prosocial behavior is animportant source of inspiration for consumers' in-storebehavior.

A salesperson's influence on the customer comesfrom many sources of communication, some verbal andmany others nonverbal. It is clear, however, that asalesperson's success will depend largely on his/hercredibility, friendship, and warmth as perceived by thecustomer. It is a familiar experience that words do notalways convey the truth. It is difficult to obtain internalevidence from the words themselves to indicate whetherthe salesperson is telling the truth or not. Nonverbalsignals, on the other hand, are regarded as less controlledand are commonly perceived by recipients to be moregenuine (Rose 1990). In fact, customers depend onnonverbal cues for evidence of truth telling.

Although the mechanisms that underscore the rela-tionship between touch and recipients' responses re-main unclear, the few explanations available in the lit-erature guide the conceptualizations of the role of touchon buyers' behavior presented here. Our approach isbased on the theory of relational communication (Parks1977) and the nonverbal communication models withinthe theory. Relational communication analysis beginswith the idea that nonverbal communication containsboth an appeal or relational component and an inter-personal bond component. The theory stems from thefundamental observation that nonverbal communica-tion, such as touch, during a request encounter will re-sult in more social influence as a result of an increase

in attentional arousal and perceived interpersonal in-timacy.

Drawing from the theory of relational communica-tion and the empirical literature on the role of touchin interpersonal communication, three general (and in-terrelated) effects on consumer behavior are suggested.First, in the right context, nonreciprocal touch shouldmake consumers feel more positive about the situationsurrounding the touching events. The mechanisms un-derlying this link were advanced by some investigators(e.g., Patterson, Powell, and Lenihan 1986) who showedthat touch enhances one's general positive feelings andmood and increases attentional arousal, which subse-quently influences one's evaluation of the externalstimuli surrounding the touch situation. Therefore, inthe shopping context in which customers seek socialsupport and affection, proper touch will produce overallpositive feelings toward the store.

Second, touch can affect the way the toucher is per-ceived by recipients. Individuals associated with thehuman potential movement (encounter groups, etc.)have long maintained that touching leads to a personliking the toucher. Limited research suggests that touchplays an important role in shaping the image of thetoucher and in enhancing interpersonal involvement.It has been shown, for example, that counselors whotouch clients are more positively perceived than thosewho do not. They are also judged to be more emphatic(Pattison 1973). Others have suggested that touchinganother person during a social encounter may increasethe recipient's feelings of social attachment, intimacy,and, therefore, liking (Mehrabian 1981). More recently,Patterson et al. (1986) have argued that people tend toattribute a multitude of positive characteristics to thetoucher and are willing to exercise certain positive be-haviors toward that person. The above conceptualiza-tion can be used to underscore the role of touch oncustomers' perceptions of the touching server. Tactilestimulations will produce a feeling of social attachmentand personal involvement while also facilitating cus-tomers' perceptions of salespeople's trustworthiness andwarmth, which will enhance their liking.

Finally, touch may play an important role in per-suasion. The limited literature on this topic would sug-gest that consumers will comply more readily with arequest made in conjunction with an interpersonaltouch. Kleinke (1977) found that subjects returned adime left in a phone booth significantly more often tothe "telephone caller" who touched them than to thecaller who did not. Similarly, Hornik and Ellis (1988)showed that touched individuals were more likely toagree to participate in mall intercept interviews. On thebasis of these studies Rose (1990) maintains that therecipient tends to assume that the toucher is in genuineneed and that the toucher likes and trusts him. Theperception of either a great need or a positive feelingtends to increase compliance.

TACTILE STIMULATION AND CONSUMER RESPONSE 451

Perhaps the most basic meaning of touch in the buy-ing setting is that an interpersonal bond is being offeredor established by the server. Like other verbal and non-verbal communication, touch seems to strengthen othermessages. The assertive component of touch might alsoinfluence the recipient. In short, we propose the follow-ing mechanism: touch exerts influence on customers'behavior by enhancing liking of the salesperson and bycreating higher emotional involvement with the shop-ping situation. In addition, a touched customer mightbe more inclined to comply or help the server becausetouch increases feelings of̂ intimacy and closeness tothe touching source.

SummaryThe preceding conceptualization and some of the

empirical evidence suggest several interesting relation-ships between tactile stimulation and buyer behavior.Specifically, it is posited that in-store touch will enhancecustomers' positive attitudes toward the store and theserver and encourage shopping behavior.

The studies presented here were designed to assessthe effect of touch on consumers' responses in legitimatedaily and common purchasing situations: first, on cus-tomers' evaluations of a store (the external stimuli);second, on the customers' evaluations of the server (thetouching source); and third, on the willingness of therecipient to comply with a request. Actually, the twolater investigations were prompted by the results of thefirst pilot study, which alluded to the potential effectsof touch on consumers' behavior.

Most research on the influence of nonverbal com-munication like touch tends to be laboratory studies incontrived situations. More recently, there has been agrowing concern for the fact that laboratory studies andthe results they produce may have little to say abouthow touch actually influences behavior in natural set-tings. This problem has led researchers to suggest thatrigorously designed experiments be carried out in morerealistic field settings. (For a comprehensive review onthis issue, see Scherer and Ekman [1982].) Therefore,all three studies reported here were conducted in thefield in genuine buying or consumption situations. Inaddition, an attempt was made, wherever possible, tocompare the influence of the source on male and femalecustomers using cross-sex and same-sex interactions.

METHODS

OverviewThe overall design and most of the basic methods

employed were purposely retained in the three studies.Specifically, all experimental groups were touched andcompared to (no-touch) control groups. Second, allthree studies were conducted in realistic field settingswhere customers were touched during genuine pur-chasing situations. These situations were chosen because

they facilitated proper experimental manipulations andcontrol over the actions of the experimenters. They alsoprovided good methods for assessing qualitative andquantitative behavioral responses. Third, because of thecomplexity of the study procedures and the experimen-tal manipulations, particular emphasis was placed onexperimenter training and control. To ensure that allexperimenters used the same touching method, main-tained a normal body posture, and followed the specifiedsampling procedures and recording methods, each ex-perimenter completed at least one rigorous trainingsession as well as a pretest procedure.

Specifically, before conducting the experiment, ex-perimenters practiced with role-playing subjects.Professional observers recorded their behavior duringthe practice and pretest sessions to ensure that it wouldbe possible to manipulate the experimental treatmentsin a consistent manner. These treatments includedtouching the subject on the arm or shoulder and main-taining a normal approach throughout the interaction.The alternative behavior was a control no-touch con-dition: experimenters were instructed and trained tomaintain normal behavior without touching. Experi-menters in both the touch and no-touch conditions wereat all times pleasant and friendly to customers. Theysmiled during each approach and maintained a constantinteraction distance. All these types of training proce-dures and touching manipulations have been shown inprevious research to be reliable and effective (e.g.,Kleinke 1980; Major, Schmidlin, and Williams 1990;Stier and Hall 1984).

Because of concern for possible distortions in thefield, rigorous measures of control were also adminis-tered during the experiment. Each experimenter wasaccompanied by a member of the research team, whosupervised and recorded the entire procedure from adistance. It should be noted that, in addition to re-porting the experimenter's performance, each observerwas asked to report any unusual subject reactions, suchas interaction break-offs or evidence of restlessness ordiscomfort. Also, between-experimenters tests wereconducted within each experimental condition to revealpossible experimenter effect beyond the designed ma-nipulation. Those experimenters who felt severely un-comfortable with the methodology were allowed to dropout of the experiments.

The methods and results of the three studies will bepresented below, together with a brief discussion ofconclusions reached through the analysis of the pre-dicted effects. A total of 206 subjects participated instudy 1, and 248 and 217 subjects completed studies 2and 3, respectively. According to statistical power tablesprepared by Cohen (1987), these sample sizes shouldbe adequate to test medium effect sizes (ES) at the .05level of significance with a power in the range of .90-.95. These conditions also adhere to the standards de-scribed by Sawyer and Ball (1981) in their discussionof statistical power in marketing and consumer behavior


research. Obviously, the methods and procedures of allthree studies were not identical. The differences arenoted below for each study.

STUDY 1The first study was part of a comprehensive project

investigating patterns of shopping-time behavior. Anattempt was made to use this opportunity to explorewhether touch would significantly affect subjects' shop-ping times and store evaluations. As noted before, theresult of this pilot study prompted us to further inves-tigate possible touch effects on consumer behavior. Ourresults have been described in detail in a companionnote (Hornik 1992). To review briefly, 286 shopperswere approached by three male and three female ex-perimenters as they entered a large bookstore. As a loneshopper entered the store, he/she was greeted by an ex-perimenter who handed him/her a four-page catalogspecifying items, prices, discounts, credit, etc. Duringalternate one-hour periods, the experimenter eithertouched the subject lightly on the upper arm or did not.

At the end of the shopping trip, all shoppers had toproceed to the cash register to stamp their parking cou-pon for exit time. The experimental subjects were ap-proached by interviewers, who recorded their shoppingtime and the amount on the sales slip and asked eachshopper to fill out a brief store-evaluation card, whichentitled him/her to participate in a drawing.

Three measures were used to evaluate the influenceof touch on shoppers' behavior: shopping time, storeevaluation, and amount of purchase. Three ANOVAswere performed for each measure to analyze the resultsof the 2 (touch vs. no touch) X 2 (experimenter's sex)X 2 (subject's sex) factorial design. These analyses re-vealed a positive main effect caused by touch for alldependent measures. The influence on mean shoppingtime (22.11 minutes for the touch group and 13.56minutes for the no-touch group) was significant{F{ 1,245) = 8.27, ;7 < .01, r = . 18).' Touched shoppersresponded more favorably toward the store (X = 3.2 vs.X = 2.1,/-(1,241) = 3.92,/7<.05, r = . 18) and purchasedmore than no-touch subjects ($15.03 vs. $12.23,F{ 1,245) = 4.86, p < .05, r = . 14). Analysis by subjects'gender showed that touch was more influential on fe-male shoppers than on male shoppers but affected onlyone dependent measure—shopping time (f(l,115)= 3.18,p<.08,/•= .16).2

The pilot study results supported the predicted effectof touch not only on shopping time but also on the way

'The following formula was used, throughout the paper, to calculatethe magnitude of all experimental effects (effect size):

r = (eta) =+ (rf/error) '

Rosenthal and Rosnow (1984, p. 356).^Effect size measure led me to not dismiss this interaction effect as

insignificant and unworthy of further attention.

consumers evaluated the store, as well as the effect ontheir purchasing behavior. These results originated thepropositions of studies 2 and 3: touch might influenceconsumers' perceptions of the toucher and facilitatehigher compliance, respectively.

STUDY 2

The second study was designed to investigate the in-fluence of touch on customers' evaluations of thetoucher during a consumption activity. This was op-erationalized using a common expression of apprecia-tion to a server—tipping (Lynn and Grassman 1990).

Crusco and Wetzel (1984) studied the effect of touch-ing on the tipping behavior of single customers. Thisexperiment is not only a replication but also a significantextension of their study. First, the current experimentinvestigated a possible server-customer gender inter-action effect. Second, as suggested by others (e.g., Ca-ballero and Pride 1984), this study looked into the pos-sible influence of server attractiveness on customers'behavior. Third, this study focused on the reaction totouch by mixed couples. An attempt was made to com-pare the results to those of Crusco and Wetzel's (1984)study by focusing on the effect of touch, by a waiter orwaitress, on the customer in the presence of a compan-ion of the opposite sex.̂ Fourth, this study containsadditional dependent variables to measure the custom-ers' evaluations of the servers.

Experimenters

The experimenters were four part-time waiters andfour part-time waitresses employed by a large restau-rant. They were selected from a pool of 27 servers onthe basis of their physical attractiveness, following apretest on this factor.

Physical attractiveness has been an important topicof research in the social sciences. There is ample evi-dence to suggest that an attractive person may be moreeffective in the interpersonal context of selling than anunattractive one (Caballero and Pride 1984). The resultsof studies showing that a physically attractive personfacilitates behavior and engenders a higher rate of com-pliance have been interpreted using various theoreticalexplanations such as social adaptation theory (Kahleand Homer 1985)."

A two-stage procedure was used here to select theeight experimenters, four of whom (two waiters andtwo waitresses) were scored as highly attractive and fourof whom (two waiters and two waitresses) were scoredlow. First, two weeks prior to the study, a group of cus-

•'The decision to investigate mixed-couple customers was alsoprompted by the fact that most customers dining at the restaurantbeing studied were mixed couples.

•"A comprehensive review on this topic can be found in Beney(1987); Debono and Teleca (1990).


tomers were asked to rate various attributes of the res-taurant, including the physical attractiveness of itsservers. The procedure and measure were adopted fromSimpson, Lerma, and Gangestad (1990). Specifically,the measure was a seven-point scale, with seven rep-resenting high attractiveness. Twelve servers (six maleand six female) were selected for the attractivenesstreatment.^ The second stage was a pretest of the ex-perimental conditions. The servers selected as experi-menters were those who scored high/low on attractive-ness and who also received a high evaluation by thetrainers on the way they conducted the experimentalmanipulations. Experimenters were not informed of theexperimental hypotheses. Each was asked to approachapproximately the same number of randomly assignedsubjects.

Subjects and DesignA total of 248 mixed-couple diners participated in

the second experiment. Subjects were assigned ran-domly to each cell in a 2 (touch vs. no touch) X 2 (serv-er's gender) X 2 (high vs. low on attractiveness server)X 2 (customer's gender) design.

Dependent MeasuresTwo measures were employed in this study to eval-

uate the influence of touch on consumers' evaluationsand reactions toward the servers. First and most im-portant, an objective measure of customer appreciationof server was obtained. Following Lynn and Grassman's(1990) study confirming the expectation that customersgive tips to convey their appreciation and to buy socialapproval and equitable relationships, tip size as a per-centage of the bill was taken as the measure of appre-ciation.*

The second dependent variable was a general eval-uation of the server obtained from subject response toa nine-point bipolar semantic differential scale. A sim-ilar scale was used for the third dependent variable,which aimed to assess overall attitude toward the res-taurant. These last two responses were obtained by in-terviewers stationed near the restaurant exit as part ofa routine survey conducted during the previous weeks.Five pairs, two from the touch and three from the no-touch condition, declined to complete the form. Theywere therefore used only for the first dependent mea-sure.

ProcedureAll experimenters went through thorough training

sessions as described previously. The experiments were

'The experimenters selected received a high interjudge reliabilityscore. The correlation coefficient obtained for ratings by males vs.by females on all 12 was .87.

^Although the person touched was recorded, obviously, no datacould be obtained on who actually made the tip decision.

TABLE 1

MEANS (SD) OF MALE AND FEMALE DINERS' TIPPING ANDEVALUATION OF SERVERS AND EVALUATION OFRESTAURANT BY TOUCHING CONDITIONS AND

SERVERS' SEX AND ATTRACTIVENESS

Experimenters"

Male (HA);Touch

No touch

Male (LA):Touch

No touch

Female (HA):Touch

No touch

Female (LA):Touch

No touch

Total:Touch

No touch

Tipping

Male

16.35(6.49)14.41(5.13)

16.04(6.20)14.07(4.78)

16.92(6.70)14.67(5.27)

16.63(6.71)14.52(4.87)

16.51(6.49)14.49(5.17)

Female

17.96(8.23)14.51(6.12)

16.21(8.11)14.12(5.84)

19.82(8.61)15.25(6.37)

16.39(8.29)14.27(5.97)

18.84(8.21)14.51(6.18)

Evaluation ofserver*

Male

3.27(1.21)2.42(.86)

2.88(1.03)1.83(.77)

3.31(1.09)2.54(.89)

2.79(.90)1.82(.76)

3.07(1.08)2.20(.91)

Female

3.49(1.27)2.48(.85)

2.94(1.14)1.93(.80)

3.53(1.38)2.57(.97)

2.69(1.02)1.95(.81)

3.25(1.13)2.27(.82)

Evaluation ofrestaurant"

Male

2.78(1.07)2.13(.79)

2.55(.91)2.07(.77)

2.81(1.06)2.46(.95)

2.60(.92)2.11(.77)

2.69(1.03)2.14(.84)

Female

2.98(1.12)2.17(.81)

2.73(1.03)2.11(.79)

3.04(1.09)2.31(.88)

2.72(.96)

2.14(78)

2.84(1.10)2.21(.82)

'HA = high attractive, LA = low attractive."Extremely good (+4), neutral (0), extremely poor (-4)."Superior restaurant (+4). neutral (0), inferior restaurant (-4).

conducted during evening hours over five weekdays. Inhalf the cases the server touched either the male or fe-male subject on the arm for one second toward the endof the meal while asking if everything was all right.

Results of Study 2

Table 1 contains the means for all experimental cells,while Table 2 reports all significant results obtained bythe ANOVA for each of the three dependent measures.

Rate of Tipping. The highest average tip (19.82percent) was obtained from the female diners in thetouched condition served by an attractive waitress. Thelowest average tip (14.07 percent) was from male dinersin the no-touch condition served by a waiter rated lowon attractiveness. This figure was close to the averagenormal tip of 13.8 percent for mixed pairs in this res-taurant. The mean scores also show that touched malediners tipped on the average 2 percent more than no-touch males, whereas the mean differences betweentouched and no-touch females is much higher—over 4percent.


NOTE.—SS = sum of squares.*p<.1.*p < .05."p < .01.

TABLE 2

SUMMARY OF ANOVAS FOR RESPONSE MEASURES: STUDY 2

Source

Touch (T)Server gender (S)Customer gender (C)Server attractiveness (A)T X CT X AT X C X A

df

1111112

Percent

SS

8.72.13.22.96.33.45.8

tipping

F

6.4"1.44.0*3.7*4.1*2.13.4+

SS

9.82.82.62.47.45.04.2

Evaluation ofserver

F

7.6**1.92.11.94.4*4.1*2.4

SS

6.11.2

.9

.34.72.82.2

Evaluation ofrestaurant

F

4.3*.6.4.5

3.3+1.92.3

The ANOVAs reveal a significant main effect fortouching on tipping (/='( 1,214) = 6.4, ;7 < .01, r = . 17).In addition, customer's gender and server's physical at-tractiveness are also statistically significant (f(l,214)= 4.0,p < .05, r = .14; f(l,214) = 3.7,/? < .05, r = .13,respectively). The only main effect not found to signif-icantly influence customer tipping was the server's gen-der. Table 2 also shows that, among the interactions,the following were found significant on tipping: touchby customer gender (7 (̂1,98) = 4.1, /? < .05, r = .20)and the three-way interaction of touch by customer'sgender and server's attractiveness (F(2,98) = 3.4, p< .01, r = .25). No other significant interaction effectswere revealed in the data for the tipping measure.

Independent F-tests show that servers who touchedthe female diner of the pair received more appreciationthan those touching the male diner. This result is es-pecially apparent when the touching server is an at-tractive waitress (F(l,58) = 3.91, p < .05, r = .25).

Evaluation of the Server. In accordance with thestudy predictions, the mean scores shown in Table 1clearly reveal that touch produced higher evaluationscores of servers by both male and female diners. Thehighest average score was obtained from female cus-tomers touched by an attractive waitress (3.53). Thelowest average score (1.82) was obtained from malesubjects served, but not touched, by unattractive femaleservers.

The ANOVAs shown in Table 2 yielded a main effectfor touch (F(l,211) = 7.6, ;?<.01, r = . 19), with serversin the high attractiveness conditions obtaining thehighest evaluations. As with the tipping measure, theservers' gender did not produce significant differencesin the way they were evaluated. Two significant inter-action results were obtained from the data: first, touchby customer gender and second, touch by server's at-tractiveness (F(l,92) = 4.4, p < .05, r = .21; F(l,92)= 4.1, ;7 < .05, r = .21, respectively). A separate F-test

showed that servers who touched the female customerreceived significantly higher evaluation scores thanthose touching the male diner (F(l,113) = 5.15, p< .01, r= .21).

Evaluation of Restaurant. Mean restaurant evalu-ation scores and ANOVA results are also shown in Ta-bles 1 and 2. The mean scores show an obvious im-provement in restaurant evaluation by both male andfemale customers who were touched by servers. As inthe previous two cases the highest mean evaluation wasobtained from female diners touched and served by anattractive waitress (3.04). The lowest score was obtainedfor male customers served by a nontouching and lessattractive waiter (2.07).

The ANOVA results reveal a significant main effectfor touching on restaurant evaluations (/"(1,211) = 4.3,p < .05, r = .14). Notable among the interactions is thesignificant {F{ 1,92) = 3.3, p < . 1, r = . 19) influence oftouch by customer gender. However, neither the genderof the touched diner nor the gender or attractiveness ofthe touching server seem to affect the way customersevaluated the restaurant. A comparison of the evalua-tions between the no-touch experimental subjects andprestudy diners revealed that averages of restaurant rat-ings were very close.

Overall, the results of study 2 reveal that a casualtouch by a waiter or waitress might be an importantpositive factor in the way mixed-couple diners will tipand evaluate the server and might also enhance the waythe couples evaluate the restaurant.

It should also be noted that in order to check forpossible unusual behavior of experimenters and sub-jects, two estimates were obtained. First, the writtenand oral notes of the observers on experimenters' andsubjects' conduct were analyzed. Second, postexperi-mental sessions with experimenters on their own andothers' behavior during the experiments were also re-corded. No evidence was found that the effects could


be attributed to rival explanations that were not in-cluded in the experimental manipulations.

STUDY 3The third experiment intended to evaluate whether

touch facilitates higher customer compliance to a mar-keting request. Specifically, the study was designed totest shoppers' reactions to a product demonstrator whotouched them while asking them to sample and buy thenew product. As in the previous two studies, the thirdexperiment was conducted in a field setting, which pro-vided more genuine consumer responses as dependentmeasures.

Stimulus and Procedure

Lone adult shoppers in a supermarket were ap-proached by four trained female demonstrators whowere unaware of the study's objectives.^ The demon-strators (experimenters) were selected from a pool ofpart-time employees who performed best in the trainingsessions. They were dressed in company uniforms bear-ing the company's name and logo, which were also dis-played on the exhibition counter.

The demonstrators, positioned in a location withinthe supermarket frequently used for new product dem-onstrations and sampling, intercepted shoppers as theypassed the exhibition counter and asked them to tastea new snack. During the alternate one-hour periods,the demonstrators either touched the subjects lightlyon the upper arm while making the request or did nottouch them. They were asked to approach an equalnumber of male and female shoppers in each experi-mental condition. Whether the shopper agreed or de-clined to taste the product, s/he was given a discountcoupon to be redeemed at the cash counter and wasshown the location of the product. No further inter-action between the subject and the demonstrator tookplace.

An observer recorded the subjects' gender, the touchcondition, approximate ages, and whether they tastedthe product or rejected the offer. As in the two previousstudies, observers who were fully aware of the studyobjectives were asked to report the demonstrators' con-duct as well as any unexpected behavior on the part ofthe shoppers.

Results of Study 3Observers' reports and study results were first ana-

lyzed to check f̂ or possible rival explanations and be-tween-demonstrators effect on the dependent variables.Specifically, the professional observers' written and oral

'No male demonstrators were available. We considered using malestudents instead but decided to use only the experienced female dem-onstrators, in order to maintain a genuine shopping situation.

TABLE 3

SHOPPERS' RATE OF COMPLIANCE TO A TASTING ANDPURCHASING REOUEST, BY TOUCH CONDITIONS

AND SUBJECTS' GENDER

Subjects

Tasted:Male(n = 106)Female (n = 111)

TotalPurchased:

MaleFemale

Total

Touch

45/58 = 77.654/59 = 91.5

99/117 = 84.6

33/58 = 56.942/59 = 71.2

75/117 = 64.1

No touch

29/48 = 60.436/52 = 69.2

65/100 = 65.0

19/48 = 39.624/52 = 46.2

43/100 = 43.0

•All significance tests based on normal approximationdistributions:

P,

z-Value"

1.923.033.37

1.812.763.17

to z and

P

.03

.001

.0001

.03

.003

.001

one-tailed

VP(1 -

reports were thoroughly discussed to reveal any possiblerival explanation. This analysis showed that the touchmanipulation seemed to be the only independent factorin this study. Also, experimental results by demonstra-tors did not show significant differences between dem-onstrators and group results. Thus, homogeneity of ex-perimenters and procedure was assumed.

To evaluate compliance, two dependent variableswere used.* The first was the subjects' willingness totaste the product, as obtained from observers' records.The second was the subjects' actual purchasing behav-ior, using redeemed coupons as the purchasing measure.

The touch manipulations combined with subject'ssex resulted in a 2 (̂ touch vs. no touch) X 2 (subject'sgender) design. Table 3 contains the number of maleand female shoppers who complied and who failed tocomply with the tasting and purchasing requests in re-lation to the touch condition.

A standard test for significance of differences in pro-portions was employed, using the normal approxima-tion to the appropriate z transforms (Guilford andFruchter 1973). Statistical significance was calculatedon a one-tail basis, since the theoretical base for thestudy specified the direction for anticipated differences.A review of Table 3 leads to the following conclusions.

Taste Request. The overall percentage of touchedsubjects who complied with the taste request was 84.6percent, whereas only 65 percent of the no-touch (con-trol) shoppers agreed to the taste appeal by demonstra-tors. The difference between the two is significant (z= 3.31, p < .0001). The rate of product sampling showsconsiderable variation between treatments, ranging

*A third dependent variable had been initially designed for thisstudy—store evaluation. It had to be dropped, however, at the lastminute, due to a misunderstanding between the store manager andchain headquarters.


from a high of 91.5 percent for touched female shoppersto a low of 60.4 percent for no-touch male shoppers.

Table 3 reveals significant differences between thetouch conditions for both male and female shoppers {p= .02 and/7 = .001, respectively). The rates of compli-ance between male and female shoppers were also com-pared. While no significatit differetices were obtaitiedfor gender effects in the no-touch condition on the tastemeasure (z = .92, p > A), significant differences wererevealed between touched male and female shoppers (z= 2.13,p< .01).

Purchase Request. As predicted, the number ofoverall coupons redeemed by the treatment (touched)group was significantly greater than for the control (no-touch) group {p < .001). Significant differences werealso found between the touch and no-touch conditionsfor both male and female shoppers. The highest pur-chasing group was touched female (71.2 percent), whilethe lowest was no-touch male (39.6 percent). Finally,touch was more influential on female shoppers thanmale shoppers (z = 1.63, p = .05), while no significantdifferences were revealed between the two sexes in thecontrol group (p> .1).

A chi-square analysis was also performed on the data.The results confirmed those in Table 3. Specifically, thetaste and purchase rates in the different experimentalconditions were compared with the compliance ratesin the control (no-touch) conditions. As anticipated,the taste and purchase rates for the treatment groupswere significantly greater than for the control groupsix\2) = 25.38, p < .01; and xH2) = 21.78, p < .01,respectively).

The data of the third experiment were in accord withthe study's predictions: touched customers compliedmore with the experimenters' requests than did no-touch customers. These results were especially apparentfor touched female shoppers.

LIMITATIONSAs with any controlled studies, the current experi-

ments necessarily imply some compromise of externalvalidity for the sake of increased measurement control.Therefore, the mode of stimuli, the type of respondents,experimenters, and instruments might have affected themagnitude or even the directions of the reported results.Hence, the possibility always exists that different con-sumer settings, different experimenters, or differentmeasurement techniques could have resulted in some-what different findings. Replications to assess these andother questions concerning validity must await furtherresearch. For example, study 2 showed the role of serverattractiveness on recipients' responses. Thus, one can-not completely rule out the influence of this factor onthe findings in studies 1 and 3. Future researchers, uti-lizing touch as an independent variable, would be ad-vised to control for the attractiveness of the touchingsource during experimental manipulations.

Although the practice sessions, as well as the observ-ers' reports, indicate that the treatments were admin-istered in an unbiased fashion, future research mightattempt to impose a stricter test for possible bias. Forexample, researchers may incorporate expectancy con-trol groups in which experimenters are given differentexplanations of the experiments (Rosenthal and Ros-now 1984).

CONCLUSIONSSubject to these limitations, the results of the three

studies are encouraging. They constitute a contributionto the growing empirical evidence in the social sciencesof the positive influence of interpersonal casual touchon recipient response. Touch connotes many meaningssuch as closeness, warmth, being cared for, contact, andaffection—for the most part comforting feelings. Thisconnection with closeness—the establishment of socialattachment—has been popularized by one major cor-poration with the slogan "Reach out—reach out andtouch someone."

The results of the studies presented here offer em-pirical support for the relationship of touch and threegeneral dependent variables that are of considerable in-terest to consumer researchers. They demonstrate thattouch has a positive influence on consumers' evalua-tions of the external stimuli surrounding the touchingsituation, that it contributes to the customers' positiveregard for the server, and that it facilitates buyers' com-pliance with a request. As such, the meaning of the re-sults is clear: a consumer's response can be influencedby touch.

The data revealed some interesting interactions, es-pecially between touch and customer's gender (studies2 and 3). These interactions indicate that differences incompliance rates are linked to the sex of the customer.Although significant server-gender effects (in studies 1and 2) are not large, systematic differences as a functionof server attractiveness were revealed in study 2, par-ticularly for female customers. The issue of differencesin reaction to touch by the recipient's gender is notconclusive. As in other studies (e.g., Stier and Hall1984), part of the current research (studies 2 and 3)provides evidence that touch has more of an effect onfemale recipients. However, the first study failed toprovide conclusive results on this point, raising thequestion of whether and under what conditions thegender of the recipient might produce a different re-sponse to touch.

Comparing the findings among the three current ex-periments with other studies (e.g., Johnson and Edwards1991) suggests the tentative explanation that, when theresponse immediately follows the touch, female recip-ients tend to respond more than males. The more theexpected response is delayed, the less apparent is theeffect on females and the greater the similarity of theirbehavior to that of male recipients. It is proposed that


these gender differences in reactions to touch possiblyreflect a tendency among women to seek more affection,warmth, and interpersonal intimacy in their social re-lationships than do men (Hall and Veccia 1990). Also,women tend to be more submissive than men (Rose1990). It is possible, therefore, that immediate reactionto touch is one way for women to express their appre-ciation for any signs of affection and interpersonalintimacy and is an expression of their behavioralsubmissiveness. These suggestions, however, deservefurther scrutiny in research on the effects of touch.

There is more to touch than "meets the hands." Be-cause of the ambiguity of touch and the variety of cog-nitive responses to touch discussed in the literature, itcannot be claimed that the three studies reported hereconfirm all the underlying mechanisms offered in thetheoretical sections. Further investigations should in-clude more direct assessments and measurements ofvariables suggested to mediate consumers' reactions totouch. For example, how does touch, in a shopping set-ting, facilitates positive moods, increase attentionalarousal, and enhance the feelings that an interpersonalbond was offered or established? Attempting to link allthese mediators to different consumers' responses arequestions for future research.

Because persuasion can be enhanced by deliberatetouch, its use is of ethical as well as practical concern.Although it is a common practice in marketing to ma-nipulate consumers to secure compliance, the ethics ofthe method is of paramount importance—particularlyin terms of unethical applications of certain modes oftouch. The issue is one of degree. Therefore, only rea-sonable tactile stimulation of consumers should be usedand/or one should rely on other less sensitive persuasiontechniques.

Future ResearchBesides addressing the methodological issues dis-

cussed in the limitations section, future research shouldexplore several areas of interest to the field of consumerresearch. Nonverbal communication is ubiquitous. Itis used simultaneously with much of the verbal com-munication through body language, gestures, facialexpressions, and physical touch. Like other nonverbalbehavior, touching may support or contradict com-munication by others. A serviceperson may explain tocustomers that they do not need to worry about theirdryer, and his or her touch may add confirmation, butit may contradict the verbalization if the servicepersonappears restless, still, or abrupt. Thus, the study of non-verbal components of behavior appears to offer a richfield for researchers interested in identifying the waysin which nonverbal communications, like touch, com-plement and structure the verbal interaction process indifferent consumption situations. The results shouldprovide new ways to increase the understanding andeffectiveness of various interpersonal communications.

Tactile stimulation in interpersonal transactions andservice delivery has been virtually ignored by consumerbehavior researchers. Future research might take a steptoward integrating research on touch with the long tra-dition of research on consumer satisfaction (e.g., Oliverand Swan 1989). Similarly, greater attention should bedirected to investigating possible cognitive and emo-tional reactions of consumers to touch and to thetoucher.

More research is needed into contextual factors per-taining to touch in consumer behavior. Given the in-herent ambiguity of touch, the context in which it occursmay be an important determinant of consumers' re-actions. Future research should reflect situations thatare encountered frequently in consumer behavior, likephysical surroundings (at home, in a bank, in a bar),social surroundings (presence of other people), and an-tecedent states (mood).

The effect of server attractiveness was studied herein the one case where data could be obtained. It is clearthat research on touch and consumer behavior requiresgreater familiarity with the attractiveness factor as wellas other moderating variables that might exert influenceon recipients' responses to touch. These might includesuch variables as race, age, and the social status of theinteractants.

Some research on touch demonstrates that peoplewho touch are seen as warm and caring by observers.For example, Kleinke, Meeker, and La Fong (1974) ex-posed college students to videotapes of couples who ei-ther touched or did not touch during the taped inter-action. Touching couples were seen as significantlycloser, more attentive to each other, and more relaxedthan nontouching couples. It might be interesting toinvestigate the meaning consumers extract from ob-serving other people touching, for example, in varioustelevision commercials. Also, it might be worthwhileto expand study 2 to include measures that will inves-tigate the influence of touch on the individual's reac-tions while observing his/her spouse being touched.

Finally, it is suggested that the research paradigmsemployed in the present studies be extended beyondthe subject of tactile communication to apply to thewide variety of nonverbal communication modes.

In conclusion, we believe that the role of touch inconsumer behavior is more important than the presentstate of the art in consumer research would seem toindicate. There is even some evidence to suggest that itmay be more important than verbal communication(Stewart, Hecker, and Graham 1987). It is certainly notless important. Therefore, further testing should be un-dertaken in order to provide a more comprehensivetheoretical foundation and practical guidance in diverseconsumer behavior situations.

[Received June 1991. Revised April 1992.]


REFERENCESArgyle, Michael (1988), Bodily Communication, Madison,

WI: International Universities Press.Assael, Henry (1987), Consumer Behavior and Marketing Ac-

tions. Boston: Kent.Beney, Joseph W. (1987), "Tfie Credibility of Physically At-

tractive Communication: A Review," Journal of Adver-tising, 11 (3), 15-24.

Caballero, Marjorie J. and William M. Pride (1984), "SelectedEffects of Salesperson Sex and Attractiveness in DirectMail Advertising," Journal of Marketing, 48 (January),94-100.

Cohen, Jacob (1987), Statistical Power Analysis for the Be-havioral Sciences, Hillsdale, NJ: Erlbaum.

Crusco, April H. and Christopher G. Wetzel (1984), "TheMidas Touch: The Effects of Interpersonal Touch onRestaurant Tipping," Personality and Social PsychologyBulletin, 10 (December), 512-517.

Debono, Kenneth G. and Christine Teleca (1990), "The In-fluence of Source Physical Attractiveness on AdvertisingEffectiveness—a Functional Perspective," Journal ofApplied Social Psychology, 20 (17), 1383-1395.

Fisher, Jeffrey D., Marvin Rytting, and Richard Heslin (1976),"Hands Touching Hands: Affective and Evaluative Ef-fects of an Interpersonal Touch," Sociometry, 39 (3),416-421.

Guilford, John P. and Benjamin Fruchter (1973), Funda-mental Statistics in Psychology, New York: McGraw-Hill.

Hall, Judith D. and Ellen M. Veccia (1990), "More TouchingObservations—New Insights on Men, Women, and In-terpersonal Touch," Journal of Personality and SocialPsychology, 59 (6), 1155-1162.

Heslin, Richard and Aari Alper (1983), "Touch—a BondingGesture," in Nonverbal Interaction, ed. John M. Wei-mann and Randall P. Harrison, Beverly Hills, CA: Sage,47-75.

Hornik, Jacob (1992), "Effects of Physical Contact on Cus-tomers' Shopping Time and Behavior," Marketing Let-ters, 3 (January), 49-56.

and Shmuel Ellis (1988), "Strategies to Secure Com-pliance for a Mall Intercept Interview," Public OpinionQuarterly, 52(4), 539-551.

Johnson, Kevin L. and Renee Edwards (1991), "The Effectsof Gender and Type of Romantic Touch on Perceptionsof Relational Commitment," Journal of Nonverbal Be-havior, 15(0,43-55.

Jones, Stanley E. and Elaine Yarbrough (1985), "A Natural-istic Study of the Meanings of Touch," CommunicationMonographs, 25 (March), 19-56.

Kahle, Lynn R. and Pamela M. Homer (1985), "Physical At-tractiveness of the Celebrity Endorser: A Social Adap-tation Perspective," Journal of Consumer Research, 11(March), 954-961.

Kleinke, Chris L. (1977), "Compliance to Requests Made byTouching Experimenters in Field Settings," Journal ofExperimental Social Psychology, 13 (2), 218-223.

(1980), "Interaction between Gaze and Legitimacyof Request on Compliance in a Field Setting," Journalof Nonverbal Behavior, 5 (1), 3-12.

, Frank G. Meeker, and C. La Fong (1974), "Effect ofGaze, Touch, and Use of Name on Evaluation of EngagedCouples," Journal of Research in Personality, 1 (3), 368-373.

Lynn, Michael and Andrea Grassman (1990), "RestaurantTipping: An Examination of Three Rational Explana-tions," Jour«a/o/£co«om/c ftyc/io/o^y, 11 (1), 169-181.

Major, Brenda (1981), "Gender Patterns in Touching Behav-ior," in Gender and Nonverbal Behavior, ed. Clara Mayoand Nancy M. Henley, New York: Springer, 15-37.

, Ann M. Schmidlin, and Lynne Williams (1990),"Gender Patterns in Social Touch: The Impact of Settingand Age," Journal of Personality and Social Psychology,58 (May), 634-643.

Mehrabian, Albert (1981), Silent Messages, Belmont, CA:Wadsworth.

Montagu, Ashley (1971), Touching, New York: ColumbiaUniversity Press.

Oliver, Richard L. and John E. Swan (1989), "Consumer Per-ceptions of Interpersonal Equity and Satisfaction inTransactions," Journal of Marketing, 53 (April), 21-35.

Parks, Malcolm (1977), "Relational Communication: Theoryand Research," Human Communication Research, 3 (3),372-381.

Patterson, Miles L., Jack L. Powell, and Mary G. Lenihan(1986), "Touch Compliance and Interpersonal Affects,"Journal of Nonverbal Behavior, 10(1), 41-50.

Pattison, John (1973), "Effects of Touch on Self-Explorationand the Therapeutic Relationship," Journal of Consultingand Clinical Psychology, 40(2), 170-175.

Reite, Martin (1990), "Touch, Attachment, and Health: IsThere a Relationship?" in Touch: The Foundation of Ex-perience, ed. Kathryn E. Bernard and T. Berry Brazelton,Madison, WI: International Universities Press, 195-228.

Rose, Susan A. (1990), "The Sense of Touch," in Touch: TheFoundation of Experience, ed. Kathryn Bernard and T.Barry Brazelton, Madison, WI: International UniversitiesPress, 299-324.

Rosenthal, Robert and Ralph L. Rosnow (1984), Essentialsof Behavioral Research: Methods and Data Analysis, NewYork: McGraw-Hill.

Sawyer, Alan G. and A. Dwayne Ball (1981), "StatisticalPower and Effect Size in Marketing Research," Journalof Marketing Research, 28 (August), 275-290.

Scherer, Klaus R. and Paul Ekman (1982), Handbook ofMethods in Nonverbal Behavior Research, Cambridge:Cambridge University Press.

Simpson, Jeffrey A., Margaret Lerma, and Steven W. Gan-gestad (1990), "Perception of Physical Attractiveness—Mechanisms Involved in the Maintenance of RomanticRelationship," Journal of Personality and Social Psy-chology, 59(6), 1192-1201.

Stewart, David W., Sid Hecker, and John L. Graham (1987),"It's More than You Say: Assessing the Influence ofNonverbal Communication in Marketing," Psychologyand Marketing 4 (4), 303-323.

Stier, Deborah S. and Judith A. Hall (1984), "Gender Dif-ferences in Touch: An Empirical and Theoretical Re-view," Journal of Personality and Social Psychology, 47(2), 440-459.

Stone, Gregory P. (1954), "City Shoppers and Urban Iden-tification: Observations on the Social Psychology of CityLife," American Journal of Sociology, 60 (1), 36-45.

Willis, Frank N. and Helen K. Hamm (1980), "The Value ofInterpersonal Touch in Securing Compliance," Journalof Nonverbal Behavior, 5 (1), 49-55.

tactile stimulation and consumer response · tactile stimulation and consumer response jacob...

Documents