Shyness and Physical Attractiveness in Mixed-Sex Dyads · Shyness Shyness (or social anxiety) is defined as "the discomfort and inhibition experienced in the presence of others that

INTERPERSONAL RELATIONS AND GROUPPROCESSES

Shyness and Physical Attractiveness in Mixed-Sex Dyads

Stella Garcia, Linda Stinson, William Ickes, and Victor BissonnetteUniversity of Texas at Arlington

Stephen R. BriggsUniversity of Tulsa

Male and female strangers' shyness and physical attractiveness strongly affected their own and theirpartners' behavior during initial, unstructured interactions. First, dyad-level measures of behav-ioral involvement and perceived interaction quality were independently predicted by the men'sshyness and the women's physical attractiveness. Second, shy men exerted avoidant control overmutual gazing by denying their female partners (but not themselves) opportunities to initiate andterminate mutual gazes. Third, as physical attractiveness of the men increased, conversationsincluded fewer references to 3rd-party individuals, with the women appearing to be primarilyresponsible for setting the "exclusive" tone of these conversations. Fourth, as physical attractivenessof the women increased, the men reported an increasing percentage of metaperspective thoughtsand feelings that reflected the symbolic adoption of their female partners' perspective.

First impressions matter. In some instances (e.g., during jobinterviews and blind dates), the images created during initialencounters govern one's access to significant social and mate-rial rewards as well as to subsequent interactions. Even whencontinued encounters are ensured (e.g., when first meeting anew neighbor or co-worker), these same images color one's sub-sequent perceptions of the other person's friendliness, desirabil-ity, and competence. The enduring impact of first impressionsseems exaggerated relative to the limited amount of informa-tion available in any single encounter.

Certain personal characteristics influence initial impres-sions more than others. Characteristics that are readily observ-able should have the greatest initial impact. For example, raceand gender are conspicuous features that have been shown toexert considerable influence on interpersonal behavior andperceptions (e.g., Ickes, 1984; Ickes & Barnes, 1977; Ickes,Schermer, & Steeno, 1979). In this study, we examined the im-pact of two other dispositional variables that are salient early inan interaction: physical attractiveness and shyness.

Both of these characteristics are readily discerned. Evidenceof their observability was reported by Funder and Dobroth

Portions of these results were presented at the 1990 InternationalConference on Interaction and Close Relationships at Nags Head,North Carolina, and at the 1990 meeting of the Midwestern Psychologi-cal Association, Chicago.

We thank Dorothy Anderson, Kristine Arnold, Ray Bristow, AndreaDeBolt, Susan Ferry, Jenny Gatlin, Mujahid Khan, Laurie Luce, CarolMarangoni, Marianne Mickan, Keith Milstead, and Jaquetta Talleyfor their assistance in collecting and coding the data.

Correspondence concerning this article should be addressed toStella Garcia, Department of Psychology, University of Texas, Arling-ton, Texas 76019-0528.

(1987). In this study, judges rated 100 Q-sort items on ninedimensions related to the subjective ease with which a trait canbe judged. These nine dimensions were then summed to pro-duce a composite index of easy visibility. Among the most ob-servable items on this index were one that referred directly tophysical attractiveness and several others that were related tothe trait dimension of outgoing versus shy (is., talkativeness,social poise and presence, relaxed, and assertive behavior). In asubsequent study, Funder and Colvin (1988) also found moder-ate correlations among independent raters for several of thesesame items—physical attractiveness, talkativeness, social poise,and so on—when strangers were rated after only a short periodof interaction. Thus, individual raters judged both physical at-tractiveness and shyness to be readily observable, and pairs ofindependent observers concurred when rating specificstrangers on these two dispositional variables.

There have been several studies of the influences of shynessand physical attractiveness on interpersonal behavior and per-ceptions. After briefly reviewing the findings of these studies,we describe the goals of the current investigation, in which weexplored the individual- and dyad-level influences of both ofthese variables.

Physical Attractiveness

There is considerable evidence for the social advantages ofbeing physically attractive (eg., Aron, 1988; Berscheid, 1985;Berscheid, Dion, Walster, & Walster, 1971; Hatfield & Sprecher,1986; Huston, 1974; Walster, Aronson, Abrahams, & Rottman,1966). Complementing an expanding body of evidence for thestereotype that "what is beautiful is good" (e.g., Bassili, 1981;Berscheid & Walster, 1974; Dion, 1981; Dion, Berscheid, &Walster, 1972; Miller, 1970) are studies demonstrating that (a)

Journal of Pereonality and Social Psychology, 1991, Vol. 61, No. 1,33-49Copyright 1991 by the American Psychological Association, Inc. 0O22-3514/91/S3.00

35

36 GARCIA, STINSON, ICKES, BISSONNETTE, BRIGGS

the mere belief that another person is attractive can have apowerful influence on social behavior (Goldman & Lewis,1977; Snyder, Tanke, & Berscheid, 1977), (b) there is a distinctadvantage in being associated with a physically attractive per-son (Sigall & Landy, 1973), and (c) interactions with beautifulindividuals are almost always aesthetically pleasant (Dion et al.,1972). Moreover, the social advantages of physical attractive-ness appear to be widespread, affecting not only dating andromantic relationships (e.g., Walster et al, 1966) but occupa-tional status (e.g, Cann, Siegfried, & Pearce, 1981) and the out-come of courtroom proceedings as well (e.g, Thornton & Ryck-man, 1983).

Within the last decade, Reis and his colleagues have empha-sized the need to study the manner in which physical attractive-ness influences naturally occurring social interactions. Of par-ticular relevance for the present study, Reis, Nezlek, andWheeler (1980) examined the daily social encounters of attrac-tive and unattractive college students. They found that for men(but not for women), physical attractiveness was significantlycorrelated with the number of mutually initiated opposite-sexinteractions that subjects reported. They also found that forboth men and women, physical attractiveness was significantlycorrelated with ratings of interaction quality and satisfaction,particularly in interactions with people of the opposite sex. In asubsequent study, Reis, Wheeler, Spiegel, Kernis, Nezlek, andPerri (1982) found that, relative to their less attractive counter-parts, attractive men rated their interactions as being more inti-mate and disclosing, and attractive women reported beingmore satisfied and pleased with their social encounters.

Perhaps the most important question raised by these studiesconcerns the processes by which physical attractiveness can in-fluence naturally occurring interaction behavior. In the Reis etal. (1980, 1982) studies, attractive men and women reportedtheir interactions with opposite-sex partners to be particularlysatisfying and rewarding. In addition, the strength of this effectwas found to increase over time (i.e, throughout the academicyear) for both freshman and senior subjects. The researchersinterpreted these data as consistent with the notion that attrac-tive individuals, being aware of both the social value placed ontheir looks and the potential use of this asset, have more optionsthan unattractive individuals in their choice of interactionpartners. This reasoning suggests that physically attractive peo-ple might have more satisfying and rewarding opposite-sex in-teractions than unattractive people simply because they areable to attract more desirable interaction partners.

Note that this explanation posits a relatively indirect effect ofattractiveness on the quality of opposite-sex interactions—thatis, one that is mediated by the desirable qualities and behaviorof the interaction partners whom physically attractive individ-uals are able to attract. It therefore diverges from the more tradi-tional explanation that physical attractiveness has a relativelydirect effect on the quality of opposite-sex interactions by evok-ing favorable cognitive, affective, and behavioral reactions fromthe partners of physically attractive individuals that are notevoked from the partners of relatively unattractive ones. Thismore traditional explanation provides the explicit or implicitrationale for studies that experimentally manipulate the physi-cal attractiveness of a target person to study the effects of thisvariable on the subject's perceptions and behavior (e.g, Byrne,

London, & Reeves, 1968; Kleck & Rubenstein, 1975; Snyder etal, 1977).

Because these two explanations are not mutually exclusive, itwould be useful to test for the effects of men's and women'sphysical attractiveness on naturally occurring heterosexual in-teractions in a context in which the assignment of partners isessentially random and not determined in any way by their levelof attractiveness. Clearly, any effects that emerged in this con-text could reflect only the direct impact of the men's andwomen's physical attractiveness on their own and theirpartner's behavior in the immediate interaction. They couldnot reflect the more indirect impact of attractiveness on theselection of interaction partners.

To test this proposed direct effect of physical attractiveness,male and female subjects in this study were randomly paired asinteraction partners without regard to their respective levels ofphysical attractiveness. The random pairing of the dyadmembers should have enabled us to determine which effects ofattractiveness, if any, can be accounted for without recourse to aselection-of-partner interpretation.

Shyness

Shyness (or social anxiety) is defined as "the discomfort andinhibition experienced in the presence of others that derivesdirectly from the interpersonal nature of the situation" (Jones,Briggs, & Smith, 1986). It is typically assessed with variousself-report inventories, all of which are highly correlated (Joneset al, 1986). However, shyness can be assessed from other per-spectives as well: (a) Participants in the interactions can reporttheir feelings and impressions (e.g, tense-relaxed, awkward-atease, and outgoing-inhibited), (b) independent observers canrate the participants on comparable dimensions, and (c) inde-pendent observers can code the videotapes for specific verbaland nonverbal behavioral markers of shyness (Briggs & Smith,1986).

People who score high on shyness and social anxiety scalesreport relatively more negative thoughts about themselves dur-ing dyadic interactions (Ickes, Robertson, Tooke, & Teng,1986), particularly if these interactions involve a member of theopposite sex (Bruch, Gorsky, Collins, & Berger, 1989; Cacioppo,Glass, & Merluzzi, 1979; Hill, 1989; Turner, Beidel, & Larkin,1986). Compared with their nonshy counterparts, shy individ-uals report being more tense and self-preoccupied during dya-dic interactions and focus specifically on the likelihood of be-ing negatively evaluated (Asendorpf, 1987,1989; Melchior &Cheek, 1990). They also describe themselves as inhibited, awk-ward, unfriendly, and incompetent relative to those who scorelow on measures of shyness (Cheek & Buss, 1981; DePaulo,Kenny, Hoover, Webb, & Oliver, 1987). In general, shy individ-uals tend to perceive and interpret their own social abilitiesnegatively. Indeed, one recent reviewer has gone so far as tosuggest "that shy individuals suffer from a pervasive lack ofbeneffectance in their processing of self-relevant information"(Cheek & Melchior, 1990, p. 68).

Independent observers also rate people with higher scores onshyness scales as appearing more shy, anxious, awkward, andinhibited during dyadic interactions (Asendorpf, 1987, 1989;Cheek & Buss, 1981; Jones & Briggs, 1984; Pilkonis, 1977).

SHYNESS AND PHYSICAL ATTRACTIVENESS 37

Objective ratings show that these individuals actually do talkless, initiate fewer conversational sequences, avert their gazemore often, smile less, show less facial expressivity, and aremore likely to engage in nervous self-touching (Cheek & Buss,1981; Daly, 1978; Mandel & Shrauger, 1980; Pilkonis, 1977).And while they are perceived by others to be relatively shy aswell, the impression they make on others is never as negative asthat implied by their own self-rated ability (Bruch et al., 1989;DePaulo et al, 1987; Jones & Briggs, 1984).

Although these results converge to form a reassuring pattern,there are some characteristic methodological limitations in thestudies from which they derive. First, most of the researchershave either asked subjects to role play a dyadic encounter (e.g.,Curran, Wallander, & Fischetti, 1980; Turner et al., 1986), haveused a confederate as a dyad member (e.g, Bruch et al., 1989;Pilkonis, 1977), or have overtly audio- or videotaped the inter-action (Asendorpf, 1987; Leary, Knight, & Johnson, 1987;Leary, Kowalski, & Bergen, 1988). Whereas procedural featuresof this sort are often necessary to achieve the specific goals of astudy, they constrain the naturalness of the resulting interac-tions and may themselves exert considerable influence on inter-personal behavior and perceptions.

Second, previous studies have often selected only those sub-jects who score extremely high or low on a measure of shyness(e.g, Asendorpf, 1987; Bruch et al., 1989; Cheek & Buss, 1981).The advantage of selecting participants with extreme scores isthat it increases the likelihood of uncovering a relationshipgiven a limited sample size. However, this increase in power isachieved by artificially inflating the magnitude of the relation-ship. Thus, studies that are based on groups of extreme scorerscorrectly assess whether an effect exists, but they do not provideaccurate estimates of the size of the effect being studied (Alf &Abrahams, 1975; Fedlt, 1961).

Third, previous studies have often relied on summary im-pressions to assess the role of cognitions in the experience andexpression of shyness. The most commonly used measure inthis regard is the Social Interaction Self-Statement Test (Glass,Merluzzi, Biever, & Larsen, 1982), which requires subjects toprovide summary ratings that estimate how frequently they ex-perienced positive (facilitative) or negative (inhibitory) thoughtsduring an interaction (e.g., Bruch et al, 1989; Turner et al,1986). However, assessing the specific thoughts and feelings thatsubjects experienced during an interaction may enable a richer,more detailed understanding of the dynamic relationship be-tween cognitions and behaviors in an unfolding social en-counter. To date, only a few studies have examined individualdifferences in shyness or social anxiety using a video-recon-struction strategy in which participants are prompted by a vid-eotape to recall the specific thoughts and feelings they experi-enced during the interaction (Asendorpf, 1987,1989; Ickes etal, 1986).

Fourth, because previous research has typically studied onegender at a time, additional work is needed on the question ofhow gender affects the cognitive, affective, and behavioral ex-pression of shyness. Some recent evidence suggests that boththe gender of the individual participants and the gender com-position of the dyad can moderate the effects of shyness in ini-tial dyadic encounters. For example, it appears that high scoreson shyness are more strongly related to negative self-statements

for men than for women (Bruch et al, 1989) and that mixed-sexinteractions are more socially threatening than same-sex inter-actions (DePaulo, Dull, Greenberg, & Swaim, in press; Russell,Cutrona, & Jones, 1986; Turner et al, 1986).

Overview and Objectives

The purpose of this study was to examine the effects of shy-ness and physical attractiveness on the thoughts, feelings, andbehavior that occur in initial, unstructured interactions be-tween two strangers of the opposite sex. Because the literaturewe just reviewed does not provide the basis for a theoreticallyunified account of these effects, the goals of our study weremore inductive and exploratory than deductive and confirma-tory.

In general, however, we expected to confirm previous find-ings indicating that (a) shyness has a pervasive inhibitory effecton the behaviors expressed in initial, mixed-sex interactions; (b)this behavioral inhibition is accompanied by certain cognitiveconcomitants of shyness (i.e, negative, self-focused thoughtsand feelings) that could interfere with interaction-relevant cog-nition and behavior; and (c) interaction with physically attrac-tive members of the opposite sex is affectively pleasant andtherefore motivates attempts to establish more intense and com-munal relationships with them. More specific questions to beinvestigated in this study (e.g, determining the direct, as op-posed to the indirect, effects of physical attractiveness) havealready been noted.

The expanded dyadic interaction paradigm (Ickes et al, 1986;Ickes & Tooke, 1988; Ickes, Bissonnette, Garcia, & Stinson,1990) was used in this study because it provides an establishedmethod for determining the manner in which dispositional fac-tors operate in initial interactions and contribute to the forma-tion of first impressions. It also provides a viable technique forassessing the links between naturalistic social behavior and nat-uralistic social cognition in unstructured, dyadic interactions.Finally, it enables the investigator to combine some of the bestfeatures of both naturalistic and laboratory research (cf. Clark&Reis,1988).

Essentially, the paradigm involved two phases of data collec-tion. The first phase involved the unobtrusive recording of spon-taneous, naturalistic social interactions that were relatively freeof the task and situational demands associated with most inter-action paradigms. The resulting videotapes were recorded in acontrolled observational setting and could be reliably coded toyield a range of behavioral measures that could be analyzed atboth the individual and dyadic levels. (For detailed reviews ofthe paradigm and the observational coding categories, seeIckes, 1982,1983; Ickes et al, 1986; Ickes & Tooke, 1988; Ickes,Bissonnette, Garcia, & Stinson, 1990). In the second phase ofthe paradigm, each dyad member was instructed to observe thevideotaped interaction immediately after the actual interac-tion. We asked the participants to list the thoughts and feelingsthat were experienced during the interaction as they viewed thevideotape.

Three advantages of the expanded dyadic interaction para-digm are directly relevant to this research. First, mixed-sexdyads were formed wherein partners were randomly pairedwithout regard to their respective levels of physical attractive-


ness or shyness. Randomly assigning strangers into pairs en-sured that selection factors did not influence subsequent per-ceptions or interpersonal behaviors. It also permitted more ac-curate estimates of effect magnitudes for physical attractivenessand shyness.

Second, by using an unstructured interaction paradigm, wecould examine the effects of the men's and women's attractive-ness and shyness on their naturally occurring cognitions andemotions as well as on their spontaneous interaction behavior.In other words, we could take an unobtrusive and fine-grainedlook at the influence these predictor variables naturally exertedon the cognitive, emotional, and behavioral components of ini-tial, mixed-sex interactions. (A detailed description of the mea-sures corresponding to each of these components is provided inthe Method section.)

Third, unlike most other studies in which only 1 dyadmember's perceptions and behavior are assessed, this study en-abled us to determine how the men's and women's physicalattractiveness and shyness affected both their own and theirpartner's behavior during a period of initial interaction. Dyad-level influences of attractiveness and shyness could also be stud-ied (Le, according to the logic outlined by Kenny [1988,1990]and Kenny & La Voie, 1985).

To the best of our knowledge, this study is the first to exam-ine systematically the effects of physical attractiveness and shy-ness on the thoughts, feelings, and interpersonal behavior ofmen and women interacting for the first time. Understandingthe separate contributions of shyness and attractiveness wouldobviously be complicated to the extent that these predictor vari-ables are themselves correlated. Previous research indicatesthat shyness correlates substantially with self-ratings of attrac-tiveness but has near zero correlations with ratings of physicalattractiveness by independent observers (Bruch, Giordano, &Pearl, 1986; Cheek & Melchior, 1990; Jones et al, 1986). Be-cause our focus was on physical attractiveness as perceived byothers (i.e., independent raters), no relationship with shynesswas expected.

The data reported in this article were collected from 38mixed-sex dyads whose responses were the focus of a previousarticle on empathic accuracy (Ickes, Stinson, Bissonnette, &Garcia, 1990). Because no experimental manipulations wereimposed on these dyads, we were able to analyze the same dataset to determine the influences of the men's and women's shy-ness and physical attractiveness without having to worry thatsuch influences were moderated by a manipulated condition ortreatment variable. As a consequence, the findings of the studydo not overlap, and are not qualified by, those reportedpreviously.

Method

Subjects

The subjects were 38 women and 38 men enrolled in introductorypsychology classes at the University of Texas at Arlington. They wererecruited through a modified version of the usual "sign-up sheet" pro-cedure in which separate sign-up sheets were used to schedule non-pre-selected male and female volunteers into mixed-sex dyads. Each pair-ing for a given session was essentially random (cf. Ickes et al., 1986). Theseparate sign-up sheets bore the names of different experiments and

d 1

i—z—i

L| £ )

6

\• 1

3_

I

8

/

» 1\

t/

a

a

1 13

7\ •

Figure 1. Schematic diagram of the laboratory setting.

directed subjects to report to different waiting areas before their partici-pation in the study.

Setting and Equipment

As indicated in Figure 1, the observation room in which the dyadicinteraction was recorded was furnished with a long couch (1) and anaccompanying coffee table (2) that concealed an FM wireless micro-phone. A video camera focused on the area of the couch and coffeetable was concealed in a darkened storage room (IS) across a hallwayfrom the observation room. This arrangement permitted the experi-menter to record the dyadic interaction through the open doorways ofthe storage and observation rooms with minimal likelihood of thecamera's being detected.

Adjacent to the observation room was a control room used to housethe video and audio equipment operated by the experimenter and aresearch assistant. Along one wall of the control room was the experi-menter's control station. Here the experimenter could sit in front of asingle table (10 in Figure 1) that supported two identical videocassetterecorders that were both connected (by means of a V-adaptor) to thevideo camera in the storage room across the hall. Each videocassetterecorder was also connected to one of two 25-in. color television moni-tors (11 and 12) that were oriented to face through one-way mirrors intoidentical test cubicles. The test cubicles (13 and 14) were each equippedwith a help button that allowed subjects to signal the experimenter (bymeans of a red light) if they needed some assistance or additional infor-mation. Each test cubicle also contained (a) a remote start/pause switchconnected to the videocassette recorder and television monitor systemunique to that cubicle and (b) a supply of thought-feeling coding forms(to be described later).

Procedure

Following the directions given on their respective sign-up sheets, the2 subjects who were scheduled for each session reported to differentwaiting areas in the psychology building. These areas were physicallyisolated from each other but were both in reasonably close proximity tothe suite of research rooms just described. The separate sign-up sheetsand waiting areas were used to help ensure that (a) 2 friends orpreviously acquainted individuals would not sign up for the same ses-sion and (b) each scheduled pair of subjects would not meet and inter-act before their session began.

Collection of the videotape data. After checking the operation of thevideo equipment, the experimenter collected the 2 subjects from theirrespective waiting areas and, while conducting them to the observation


room, observed their reactions for any sign that they were alreadyacquainted with each other.1 Once satisfied that the subjects were in-deed strangers, the experimenter escorted them into the observationroom, asked them to place any books or other belongings on the table(5 in Figure 1), and directed them to be seated on the couch (1). Theexperimenter then reached for copies of the informed-consent formfrom a box on one of the bookcases (6) but pretended to discover thatonly one copy of the form was left. Explaining the need to get addi-tional copies of the form, the experimenter left the observation room,promising to return "in a few minutes." At this point, a research assis-tant in the control room started the videotaping. Exactly 6 min later, atthe end of the observation period, the experimenter returned to theobservation room and the videotaping was terminated.

After probing for any evidence of suspicion, the experimenter con-ducted a partial debriefing.2 The subjects were told that they had beenvideotaped for the purpose of studying their naturally occurring inter-action behavior. The experimenter described the videotaping proce-dure, explained the methodological importance of not telling the sub-jects about the taping in advance, and informed them that their writtenconsent was required for the tape to be used as data. To assure thesubjects that their rights to privacy had been protected, the experi-menter explained that the videotape had not yet been reviewed orstudied in any way. If 1 or both of them exercised the right to have thetape erased immediately, instead of releasing it for use as data, thecontent of the interaction would remain their own private concern.(None of the subjects in this study requested to have the tape erased.)

Collection of the thought-feeling data. Once the subjects' writtenconsent had been obtained, the experimenter described the next partof the study. In this part, the subjects were asked to view the videotapeand write down the thoughts and feelings they experienced during the6-min interaction period. When the subjects had been seated in theirindividual cubicles (13 and 14 in Figure 1), the experimenter used videoinstructional tapes to give them detailed instructions regarding thethought-feeling data collection procedure. The two videocassette re-corders were then activated, and the subjects used their start/pausecontrols independently to start and stop their respective copies of thevideotape.

The subjects were instructed to stop the videotape at each pointduring the interaction that they recalled having had a specific thoughtor feeling. They recorded each thought or feeling on a thought-feelingcoding form by entering (a) the time the thought or feeling occurred(available as a digital readout on the upper left part of the videotape),(b) whether the entry was a thought or a feeling (coded "I was thinking:"or "I was feeling:" on the coding sheet), (c) the specific content of thethought-feeling entry (written in sentence form to complete the initialphrase), and (d) whether the entry was presumed to be positive, nega-tive, or neutral in its overall affective tone (coded +, - , or 0 on thecoding sheet).

The instructions the subjects received explicitly encouraged them toreport all of the thoughts and feelings they remembered having had asaccurately and honestly as possible. The importance of being com-pletely candid was emphasized, and the subjects were assured that thethoughts and feelings they recorded would not be seen by their interac-tion partners. On the other hand, the instructions cautioned them toreport only those thoughts and feelings they distinctly rememberedhaving during the observation period. They were not to report anythoughts and feelings that occurred to them for the first time whileviewing the videotape.3

Collection of the final self-report measures. In the final part of thestudy, the subjects were asked to remain in their respective cubicleslong enough to complete a posttest questionnaire. This questionnaireincluded items that assessed the subjects' perceptions of (a) the qualityof their interaction (i.e., "forced, awkward and strained" vs. "smooth,natural and relaxed") and (b) the degree to which they liked each other.

When each subject had completed the questionnaire and returned it tothe experimenter, the experimenter then debriefed the subjects morefully; asked them not to discuss the study with potential future subjects;and then thanked them, gave them point credit for their participation,and released them.

Behavioral Measures

Three types of behavioral data—static behaviors, dynamic behav-iors, and personal pronoun usage—were coded from the videotapes byindependent judges who were kept unaware, insofar as possible, of anyindependent or predictor variables being studied. The static behaviors,which either occurred only once or did not vary much across time,included measures of the degree of each subject's body orientationtoward his or her partner and the openness of each subject's bodyposture. The interrater reliabilities for these measures were .70 and .89,respectively.

The temporally more variable dynamic behaviors included the totalfrequency and duration of verbalizations (i.e., speaking turns), directedgazes, mutual gazes, expressive gestures, and expressions of positiveaffect (i.e., smiles and laughter). Also included was a measure of thenumber of verbal reinforcers (e.g., yeah, right, and uh-huh) provided byeach dyad member during the course of the interaction. The interraterreliabilities of these measures ranged from .75 to .99, with a mean of.90. (For the operational definitions of these dynamic behavior mea-sures, see Ickes & Turner, 1983, pp. 214-215).

Two independent judges coded personal pronoun usage from thevideotapes. They used tally marks to record the number of first-, sec-ond-, and third-person pronouns (broken down into the subcategoriesof singular and plural) occurring in each dyad member's conversation.These values were then divided by the total number of personal pro-nouns used by the dyad member to create percentage scores thatpreviously have been shown to be free of any confound with individualdifferences in the amount of talking (Ickes, Reidhead, & Patterson,1986). These percentage score measures provide a common metric forassessing the degree to which the content of the dyad members' conver-sations reflect a focus on self (first-person pronouns), one's currentinteraction partner (second-person pronouns), or some other person or

1 If the subjects in a given session displayed any sign that they werealready acquainted with each other, the experimenter cited "equipmentproblems" and gave each subject an experimental point credit withoutrequiring any further participation in the study. If the subjects werestrangers, the experimenter continued the procedure without introduc-ing them to each other or providing any instruction or encouragementfor them to interact or get to know each other.

2 Four dyads were dropped from the study after the probe for suspi-cion revealed that 1 or both of the dyad members had located the videocamera while being taped. This level of suspicion was so unusual in ourresearch that we were motivated to investigate it further. In doing so,we found that a common element in three of the four cases was atten-dance at a lecture given by an introductory psychology instructor whodiscussed the unobtrusive measurement of behavior by means of aconcealed video camera. With the instructor's cooperation, this lec-ture is no longer given during the first several weeks of the semester.

3 Following the collection of their own thoughts and feelings, thesubjects were asked to view the videotape a second time for the pur-pose of listing their inferences about the content and valence of each oftheir partners reported thoughts and feelings. Because the findingspertaining to these measures of empathic accuracy are reported in anearlier article (Ickes, Stinson, Bissonette, & Garcia, 1990) and haveminimal relevance to the findings reported here, the empathic accu-racy assessment is not discussed in the present article.


persons not involved in the current interaction (third-person pro-nouns). The interrater agreement rates for these measures were .97, .94,and .92, respectively.

Self-Report Measures

Three types of self-report data were obtained that are relevant to thisinvestigation: (a) the subjects' responses on a self-report measure ofshyness (Cheek & Buss, 1981), (b) the dyad members' own reportedthoughts and feelings, and (c) their responses to the previously de-scribed items on the posttest questionnaire. To minimize their per-ceived relevance to this study, the 9 items on Cheek and Buss' shynessmeasure were embedded in a research survey that included over 100items tapping a range of different personality traits of widely varyingcontent. This research survey was administered to the subjects at theend of each experimental session under the rationale that the data werebeing collected for another project. For this sample, the internal consis-tency value of the shyness measure was .80.

Thought-Feeling Measures

The subjects' own thought-feeling data were later coded by twojudges who determined whether the thought-feeling entry representeda direct perspective (subject's own perspective) or a metaperspective(subject's representation of his or her partner's perspective; cf. Laing,Phillipson, & Lee, 1966). The same two judges also determined thetarget or object of the perception expressed by the entry (self, partner,other person(s), environmental object, event, or circumstance). Therates of interrater agreement for these categorical classifications were.99 for perspective and .87 for target—values consistent with thosepreviously reported by Ickes et al, 1986.

The coded thought-feeling data for the subjects within each dyadwere subsequently entered into a software program called COLLECTYOUR THOUGHTS (Ickes, Bissonnette, Garcia, & Stinson, 1990).This program produces as output various summary indexes ofthought-feeling content. In this study, these measures included thepercentages of each subject's reported thought-feeling entries that werecoded as thoughts, feelings, positive entries, negative entries, self (rele-vant) entries, partner entries, other(s) entries, environment entries, di-rect perspective entries, and metaperspective entries. A few more fine-grained measures that were of special theoretical interest (e.g., the per-centages of negative self-entries and positive partner entries) were alsocomputed.

Physical Attractiveness Measure

The physical attractiveness of each subject was assessed by 11 raters(5 men and 6 women) who viewed a clip of the first 20 s of each dyadicinteraction with the sound turned off. Each dyad member's attractive-ness was rated on a 10-point scale ranging from extremely unattractive(1) to extremelyattractive'(10). The internal consistency of these ratingswas .93, a value that was sufficiently high to justify using the mean ofthe ratings as the measure of physical attractiveness.

Results

Data-Analytic Strategy

Note that we have four predictor variables in our study, eachof which could plausibly predict a given interaction behavior.The men's talking, for example, could be influenced not only by(a) their own attractiveness (i.e., attractive men might be moreconfident and socially assertive) but also by (b) their femalepartner's attractiveness (which might provide a greater incentive

for the men to talk), (c) the men's own shyness (which mightinhibit their talking), and (d) the women's shyness (which mightlead to the failure of the men's talking to be reinforced). Becauseall four predictors could plausibly determine a given behavior,we had to take all four into account simultaneously (i£, in amultiple regression framework) and to statistically control orpartial out the effects of the three remaining predictors whentesting for the unique effect of each.

Before analyzing these data, we had to compute the correla-tion between the male and female partners' scores on a givencriterion variable to determine whether that variable should beanalyzed at the individual or dyadic level of analysis. If the dyadmembers' scores were independent of each other, the variablewas analyzed at the individual level (ie, the four predictorswere related to the man's individual behavior and to thewoman's individual behavior, as indicated in the lower portionof Figure 2). However, if the dyad members' scores were nonin-dependent, their scores on the variable were combined through

DYAD BEHAVIOR

(Group Mean)

DYAD MEMBERS'

RESPONSES ARE

INTERDEPENDENT

DYAD MEMBERS'

RESPONSES ARE

INDEPENDENT

Figure 2. Data-analytic strategy: Partial correlations computed at thedyad level (top) or at the individual level (bottom) of analysis.


averaging to create an aggregated, dyad-level variable for subse-quent analysis. (In this case, the four predictors were related tothe dyad-level variable, as indicated in the upper portion ofFigure 2).

Using guidelines for establishing nonindependence that wereslightly more stringent than those suggested by Kenny and LaVoie (1985), we denned as dyad-level measures all those depen-dent measures for which the correlation between the dyadmembers' scores exceeded .22 (all ps <. 15). The actual values ofthe correlations denning the set of dyad-level measures rangedfrom .24 to .78 (see Table 1, last column).

Once the two sets of criterion variables—dyad-level mea-sures and individual-level measures—had been identified, mul-tiple regression analyses were used to test the effect of eachpredictor variable on each criterion variable, with the effects ofthe three remaining predictor variables statistically controlled.4

The results of these analyses are reported in Tables 1,2, and 3.The partial correlations estimating the effects of each predictorvariable on various dyad-level measures of thoughts, feelings,and behavior are reported in Table 1. The partial correlationsestimating the effects of the men's shyness on individual-levelmeasures of their own and their female partner's thoughts, feel-ings, and behavior are reported in Table 2. Finally, the partialcorrelations estimating the effects of the men's and women'sphysical attractiveness on individual-level measures of themen's and women's thoughts, feelings, and behavior are re-ported in Table 3.

A straightforward interpretation of these data would obvi-ously be compromised if shyness and physical attractivenesswere found to be significantly correlated for either the men orthe women in this study. Consistent with research reported byothers (eg., Bruch et al, 1986; Cheek & Melchior, 1990; Jones etal, 1986), however, the data showed that self-reported shynessand other-rated physical attractiveness were uncorrelated forthe men, r(36) = .01, the women, r(36) = -.02, and for all sub-jects combined, r(74) = .01, ns.

Correlates of the Men's Shyness

The partial correlations estimating the effects of the men'sshyness on various dyad-level behaviors are reported in the firstcolumn of Table 1. These data indicate that the men's shynesswas negatively correlated with such aspects of the dyadmembers' verbal behavior as the frequency and duration oftheir speaking turns and the number of questions they askedeach other. The men's shyness was also negatively correlatedwith aspects of the dyad members' nonverbal behavior that in-cluded the frequency and duration of their positive affect (i£.,smiles and laughter), their directed gazes, and their mutualgazes. Considering the strong and pervasive inhibitory influ-ence of the men's shyness on these primary measures of interac-tional involvement, it is hardly surprising that the men's shynesswas also negatively correlated with the dyad members' ratings ofthe perceived quality of their interaction.

The partial correlations estimating the effects of the men'sshyness on their own and their female partners' individual-levelbehaviors are reported in Table 2. The data in the first columnof Table 2 indicate that the men's shyness was positively corre-lated with the percentage of negative thought-feeling entries

they reported having experienced during the interaction, and,in particular, with the percentage of negative feelings that con-cerned themselves. On the other hand, the men's shyness wasnegatively correlated with the percentage of thoughts and feel-ings they reported having had about their female partners. Fi-nally, the more shy the men were, the less open their body pos-tures were and the fewer expressive gestures they used.

These individual-level correlations converge with the dyad-level correlations in suggesting that the greater the men's levelof shyness, (a) the more uncomfortable and self-conscious theyfelt, (b) the more inhibited they were in their own nonverbalbehavior, and (c) the less interactional involvement (both verbaland nonverbal) they displayed in the dyadic interaction.

The data in the second column of Table 2 reveal that themen's shyness was positively correlated with the percentage ofself-entries, particularly with the percentage of negative self-en-tries that their female partners reported. Moreover, the men'sshyness appears to have inhibited such aspects of the women'snonverbal behavior as their gestures and their degree of bodyorientation toward their male partners. These effects of themen's shyness on the women's behavior tend to parallel similareffects of the men's shyness on their own behavior. Our assump-tion is that the direct effects of the men's shyness on their ownbehavior were responsible for the indirect effects of the men'sshyness on the behavior of their female partners.

Note, however, that the men's shyness had some distinctiveeffects on the women's gazing behavior that were not apparentin the men's. These effects took the form of strong negativecorrelations between the men's shyness and the frequency withwhich the women initiated and terminated mutual gazes withthem. Because the corresponding correlations for the men werenear zero and nonsignificant (-.01 and -.12, respectively; referto the first column in Table 2), it is clear that the men's shynesshad essentially no effect on their own initiations and termina-tions of mutual gazes but dramatically limited the women's abil-ity to initiate mutual gazes (by meeting their male partners'gazes) or to terminate mutual gazes (by breaking off the mutualgazes by looking away before their male partners did). Theserelationships are graphically depicted in Figure 3.

A follow-up analysis conducted at the individual level re-vealed (a) a significant negative partial correlation (r = — .53, p <.002) between the men's shyness and the duration of time theyspent looking at their partners but (b) a nonsignificant negativepartial correlation between the men's shyness and the frequencyof their gazes (r = -.23). Given the entire pattern of data fordirected and mutual gazes, there is only one reasonable infer-ence that can be drawn. This inference is that the more shy themen were, the more they managed to keep their femalepartners from establishing and subsequently breaking eye con-tact with them by looking at the women only in very briefglances and then looking away quickly if they found the womento be already looking at them. Through this process, the shymen could control the onset, offset, and overall level of mutual

4 The results of moderated multiple regression analyses, designed totest selected dependent measures for statistical interactions involvingsubsets of the four predictor variables, are described later in this arti-cle.


Table 1Correlations of the Men's and Women's Shyness and Physical Attractiveness With Dyad-LevelBehaviors, Controlling for the Effects of the Other Three Predictors

Dyad-level measure

Thought-feeling content(% positive entries)

Posttest self-reportPerceived interaction qualitySubject likes partner

Verbal behaviorVerbalizations

FrequencyDuration

No. questions askedNo. conversation sequences initiated% third-person singular pronouns

Nonverbal behaviorsPositive affect

FrequencyDuration

Direct gazesFrequencyDuration

Mutual gazesFrequencyDuration

Shyness

Men's

—.49*****

-.42***-.45***-.38**-.39***

-.34**-.36**

-.41***-.45****

-.44****-.39***

Women's

-.27

-.30-.27-.14-.30

-.26-.22

-.36**-.31

-.38**-.25

Attractiveness

Men's

.46*****

.28

.09

.28

.30

.10

.12-.54*****

.38**

.32

.11

.32

Women's

-.05

.35**

.41**

.53*****

.48***

.48***

.50***-.06

.25

.42**

.36**

.35**

Male-female

correlation

.29*

.30*

.41***

.78*****

.24*

.41***

.28*

.49***

.39***

.42***

.54*****

.52*****

1.00*****1.00*****

Note. For each of the partial correlations, df= 33. Given the interdependence of the male and female dyadmembers' scores for all of the variables in this table, one would not expect that a given predictor's correla-tion with the male's (individual) behavior would differ significantly from its correlation with the female's(individual) behavior. Pearson-Filon tests confirmed this expectation, indicating no significant differen-tial correlation for any of the findings.•p<.07. **p<.05. ***p<.01. ****p<.005. *****p<.0001.

gaze that occurred during the interaction. By restricting theopportunity of their female partners to meet their own briefglances, the shy men could dramatically limit the women's abil-ity to initiate and terminate mutual gazes but not significantlylimit their own ability to do the same.

The clarity and precision with which the data pinpoint thishypothesized process lead us to speculate that it is a fundamen-tal means by which the men's shyness affected the behavior ofboth dyad members in these initial, mixed-sex interactions. Wefurther speculate that this process may have helped to mediatethe women's negative reactions in such encounters. Thus, inaddition to its direct effect of limiting mutual gaze at the dyadlevel, the shy men's unusual gazing behavior may have beenresponsible for such indirect effects as (a) making the womenfeel self-conscious and uncomfortable, (b) inhibiting the verbaland nonverbal behavior of both dyad members, and (c) leadingthe dyad members to rate their interaction more negatively (i.e.,as "awkward, forced, and strained" rather than as "smooth,natural, and relaxed").

Correlates of the Women's Shyness

The women's shyness was significantly correlated with onlytwo dyad-level measures, and it was not significantly correlatedwith any of the individual-level measures of their own or theirmale partners' behavior, thoughts, and feelings. As indicated in

the second column of Table 1, the women's shyness was nega-tively correlated with both the number of directed gazes and thenumber of mutual gazes that the dyad members displayed. Incontrast to the men's shyness, the women's shyness did not ap-pear to induce feelings of discomfort and self-consciousness inthe dyad members, inhibit their verbal interaction, or lead themto rate the quality of their interaction negatively. But, like themen's shyness, the women's shyness did have the effect of limit-ing the frequency of gazing behavior at the dyad level.

Correlates of the Men's Physical Attractiveness

As the third column of Table 1 reveals, the men's physicalattractiveness was positively correlated with three dyad-levelmeasures. It was correlated, first, with the percentage of posi-tive thoughts and feelings the dyad members reported and, sec-ond, with the total amount of time they smiled at each other.These findings indicate that interactions involving attractivemen were affectively positive.

The third and most intriguing effect of the men's attractive-ness was its strong negative correlation (-.54) with the dyadmembers' use of third-person singular pronouns in their conver-sation. A plausible interpretation of this last finding is that itrepresents a phenomenon that might be termed the exclusivityeffect. Presumably, by making relatively few references in theirconversation to specific third-party individuals, the members


Table 2Correlations of the Metis Shyness With the Men's and WomensBehavior, Controlling for the Effects of the Other Three Predictors

Measure

Thought-feeling content% feelings% self-entries% negative entries% negative self-entries% negative self-feelings% partner entries

Nonverbal behaviorsMutual gazes initiated

(frequency)Mutual gazes terminated

(frequency)Expressive gestures

(frequency)Expressive gestures

(duration)Degree of body orientationOpenness of body posture

Correlation of men's shyness with

Men's behavior

.47****

.2049****.50****.51****

-.37*

-.01

-.12

-.40**

-.06-.04-.36*

Women's behavior

.23

.37*

.26

.34*

.29-.33*

-.63****1

-.54****"

-.49****

-.38**-.35*

.04

Note. For each of the partial correlations, df= 33.*tdiff(35) = 3.50,p<.01 (Pearson-Filontest). btdiff(35)=2.14,p<.05(Pearson-Filon test). The differences for all other pairs were not signifi-cant.*p<.05. **p<.025. ***p<.01. ****p<.005.

of dyads containing attractive men could create a conversa-tional context that implied that their relationship was an exclu-sive one, and therefore, a more intimate and communal one.This interpretation raises the interesting question, however, ofwhether it was the women or their attractive male partners whowere primarily responsible for setting the exclusive tone ofthese conversations.

The data do not permit a definitive answer to this question. Itmay be worth noting, however, that the men's attractiveness wascorrelated —.51 with their female partners' use of third-personsingular pronouns but was correlated only —.39 with their ownuse of third-person singular pronouns. Because the correlationaccounted for substantially more variance in the women's be-havior (r2 = .260) than in the men's behavior (r2 = .152), themost plausible interpretation may be that the women were set-ting the exclusive tone of these conversations and that the at-tractive men were following their lead.

This interpretation is made even more plausible by the posi-tive correlation (33) between the men's attractiveness and thepercentage of positive partner thoughts and feelings that theirfemale partners reported (Table 3, second column). This corre-lation suggests that the women's feelings of attraction to physi-cally attractive male partners may have motivated them to setan exclusive tone for their conversations to facilitate greaterintimacy and mutual involvement. That the women were press-ing for more intimacy and involvement than their attractivemale partners is further suggested by the finding (reported inthe first column of Table 3) that the more attractive the menwere, the more likely they were to terminate mutual gazes bylooking away while their female partners were still looking atthem.

Correlates of the Women's Physical Attractiveness

As the data in the fourth column of Table 1 show, the women'sphysical attractiveness was positively correlated with severaldyad-level measures of verbal and nonverbal behavior. Withregard to verbal behaviors, the women's attractiveness wasstrongly correlated with the frequency and duration of talkingwithin the dyad, with the number of questions asked, and withthe number of conversation sequences initiated. With regard tononverbal behaviors, it was also positively correlated with thedyad-level measures of the frequency and duration of directed

Table 3Correlations of the Men's and Womens Physical Attractiveness With the Mens and WomensBehavior, Controlling for the Effects of the Other Three Predictors

Measure

Thought-feeling content% metaperspective entries% positive partner entries

Posttest self-reportSubject understands partner

Verbal behaviorsNo. verbal reinforcers

Nonverbal behaviorsOpenness of body postureNo. mutual gazes terminated

Men's attractivenesscorrelated with

Men's Women'sbehavior behavior

.28 .33*

.38*** -.01

Women's attractivenesscorrelated with

Men'sbehavior

.41***

.34**

.01

.39***

.15

Women'sbehavior

-.11*

.16

.34**

.04

.36**

Note. For each of the partial correlations, df- 33.* ?diff(35) =2.31, p<. 05. The difference between the pair of correlations is significant by the Pearson-Filontest. The differences for all other pairs were not significant.*p<.055. **p<.05. ***/>< .025.


20MEAN 16

INITI

AT

ONS

10 -

•

1IJ• •women SSSMan

JLow Moderate

MEN'S SHYNESS LEVEL

High

M 2 0

EAN 16TERM 10INAT 5I

ONS n IIJ

••Woman SSjMen

Low Moderate

MEN'S SHYNESS LEVEL

High

Figure 3. Average number of mutual gazes initiated (top) and termi-nated (bottom) as a function of the men's shyness level. (The meansdepicted were adjusted to control for the effects of the women's shynessand the men's and women's attractiveness.)

gazes and the frequency of mutual gazes. Finally, it was posi-tively correlated with the dyad-level measures of how much thedyad members reported liking each other and how positivelythey rated the quality of their interaction.

Perhaps the most striking effect of the women's attractivenesswas its individual-level correlation with the percentage of meta-perspective thought-feeling entries reported by the men in thisstudy (Table 3, third column). The more attractive the womenwere, the more likely the men were to symbolically adopt theirpartners' cognitive perspective and to "see the interactionthrough her eyes." Consistent with this finding, the women'sattractiveness was also positively correlated with the men's per-ception that they understood their female partners. Finally, thewomen's attractiveness was positively correlated with the open-ness of their male partners' body posture.

The greater number of conversation sequences initiated indyads with physically attractive women hints at an interactionpattern in which the conversation was sporadic and discon-tinuous. It suggests, in other words, that the conversation se-quences were short-lived and died out fairly rapidly and thatnew ones were subsequently initiated that suffered the same

fate. This interaction pattern makes sense if one assumes thatphysically attractive women did not encourage extended conver-sation with their male partners, whom they may have perceivedas "coming on" to them. Such an assumption would also ac-count for our discovery, in subsequent individual-level analyses,that the women's attractiveness was significantly correlatedwith an increased duration of their male partners' talking (par-tial r= .48, p < .005) but not of their own talking (partial r= .25,ns). In addition, it would account for the women's attractivenessbeing positively correlated with their own, but not with theirmale partners' use of verbal reinforcers and termination of mu-tual gazes (Table 3, column 4).

In summary, the most parsimonious explanation of the datain the last two columns of Table 3 may be that the attractivewomen perceived their male partners as "coming on" to them,and to avoid encouraging these overtures, tended to break offeye contact and allow the conversation sequences to die by let-ting the men "talk themselves out" without responding at anylength themselves. This same explanation also accounts fairlywell for the men's reactions to their female partners' attractive-ness. Specifically, the male partners of attractive women triedrepeatedly, but without much success, to engage them in ex-tended conversation. Given these efforts, it is not surprisingthat the men were highly motivated to adopt their femalepartners' cognitive perspective (i.e., through metaperspectivetaking) to figure out how to make these interactions work. Themen's subsequent perceptions that they understood their attrac-tive partners may have been partly veridical, because partnerattractiveness was found to be a modest but significant predic-tor of actual empathic accuracy (Ickes, Stinson, Bissonnette, &Garcia, 1990).5 The same perceptions may have been partlyillusory, however, if the men failed to appreciate that the attrac-tive women did not want to encourage their overtures.

Preliminary Tests of Possible Interaction Effects

Because the four predictor variables in this study were un-correlated, they should operate as nonredundant predictors ofthe behavior and perceptions of the men and women in theseinitial interactions. What we have not yet done, however, is testfor possible interaction effects involving different subsets ofthese four variables (men's shyness, women's shyness, men's at-tractiveness, and women's attractiveness).

Multiplicative product terms could be computed to test forstatistical interactions involving two, three, or all four of thepredictors. However, the sheer number of possible interactions(11 for a four-variable model), when multiplied by the largenumber of dependent measures (at least 60) and further compli-cated by the need to analyze some measures at the dyad leveland others at the individual level, meant that any attempt toreport, describe, and interpret all of the significant interactionsin the context of an already lengthy report was clearly beyondthe scope of this article.

Lacking sufficient space to report and discuss the interaction

5 All of the analyses we have conducted indicate that this is the onlyfinding from the empathic accuracy data reported by Ickes, Stinson,Bissonnette, and Garcia (1990) that has any bearing on the presentresults.


analyses separately for each measure, we elected to report only acouple of preliminary tests of possible interaction effects. Tosummarize the set of relevant variables as efficiently as possible,we took the following steps. First, on the basis of the results of avarimax factor analysis of the primary measures of interac-tional involvement, we summed the standardized scores onfour dynamic behaviors—frequency of verbalizations, directedgazes, positive affect, and expressive gestures—to form a com-posite, dyad-level index of interactional involvement. The load-ings of these four behaviors on the only factor to emerge fromthis analysis were .93, .93, .78, and .63, respectively. Second, thescores reflecting the dyad members' posttest ratings of theirinteraction as "smooth, natural, and relaxed" (positively keyed)and as "forced, awkward, and strained" (negatively keyed) weresummed to form a composite, dyad-level index of perceivedinteractional quality.

Multiple regression analyses revealed that regressing the com-posite measure of interactional involvement onto the four pre-dictor variables yielded a multiple R of .68. Similarly, regressingthe composite measure of perceived interactional quality ontothe four predictor variables yielded a multiple R of .61. More-over, these findings indicated that a substantial portion of vari-ance in the dyad-level measures of interactional involvementand perceived interaction quality was accounted for by only twoattributes of the male and female dyad members: the men'sshyness and the women's physical attractiveness. For the com-posite measure of interactional involvement, the men's shynessand the women's attractiveness were the only significant individ-ual predictors, with partial correlations of .55 and .50, respec-tively. They were also the only significant individual predictorsof the composite measure of perceived interactional quality,with partial correlations of .49 and .35, respectively.

In general, the various interaction terms failed to contributesignificantly to the prediction equation for either interactionalinvolvement or perceived interactional quality. Out of 22 inter-actions tested, only one was significant (i.e., the interaction ofthe women's shyness and the men's physical attractiveness forthe interactional involvement measure). We do not discuss thissingle interaction because it was neither strong nor readily in-terpretable and because one significant interaction out of 22tested could be expected by chance alone.

Given the strong and pervasive main effects reported in Ta-ble 1 and the general lack of qualifying interaction effects re-ported here, we suggest that an additive model based primarilyon the men's shyness and the women's physical attractivenessmight account quite parsimoniously for their behavioral involve-ment and perceived interaction quality at the dyadic level ofanalysis. On the other hand, a more complex model might stillbe required to account for the men's and women's behaviors andperceptions at the individual level of analysis.

Discussion

The results of this study not only replicate and confirm thoseof previous investigations but add several novel insights as well.Accordingly, our discussion of the results for the shyness andphysical attractiveness variables (a) reviews the ways in whichthey replicate and reinforce the results of earlier studies andthen (b) focuses on the theoretical and practical implications of

the findings in this study that have not been previously re-ported.

Summary and Implications of the Attractiveness Data

Our findings contribute in several ways to understandinghow physical attractiveness influences the unstructured inter-actions of mixed-sex dyads. First, our findings complementthose of earlier studies regarding the social value that peopleplace on physical beauty (Berscheid et al, 1971; Dion et al,1972; Walster et al., 1966; see Patzer, 1985). In general, thesefindings indicate that individuals not only value and desire butalso seek to establish and intensify interactions with those whoare physically attractive. Similarly, the apparent exclusivity ef-fect we found suggests that when women are in the company ofattractive men, they attempt to establish a more intimate andcommunal relationship with their partners—even during theearliest minutes of their first interaction. Moreover, when meninteract for the first time with attractive women, they not onlytry repeatedly to engage their partners in conversation but alsoattempt to adopt their partner's cognitive perspective, presum-ably with the goal of maximizing the success of these overtures.These findings are clearly consistent with the assumption thatinteracting with physically attractive members of the oppositesex is a social investment worth pursuing (cf. Crouse & Mehra-bian, 1977; Sigall & Landy, 1973).

Second, our results also reinforce the general assertion thatinteractions with physically attractive people are affectivelypleasant and qualitatively positive. As the dyad-level correla-tions in Table 1 show, the women's attractiveness was positivelycorrelated with the degree to which the dyad members likedeach other and perceived their interaction as pleasant. Simi-larly, the men's attractiveness was positively correlated with thepercentage of positive thoughts and feelings the dyad membersreported during their interaction as well as with the amount oftime they smiled at each other.

Third, our findings permit greater specificity in the theoreti-cal interpretation of these attractiveness effects. By randomlypairing the male and female dyad members, we eliminated theattractive persons' usual opportunities to select their interac-tion partners. Consequently, our results can be interpreted asreflecting the relatively direct impact of physical attractivenesson the initial interactions of randomly paired, opposite-sexstrangers. This interpretation suggests that it is possible to ac-count for the Reis et al. (1980, 1982) finding that physicallyattractive people rate their opposite-sex interactions as particu-larly satisfying without recourse to a selection-of-partner inter-pretation. Although we cannot rule out the possibility that therelatively direct effect of attractiveness we obtained is furtheraugmented by a more indirect selection-of-partner effect, itdoes seem clear that partner selection is not a necessary causeof the association between physical attractiveness and self-re-ported satisfaction in the interaction of heterosexual dyads.

Fourth, perhaps the most intriguing of the present findings iswhat we have tentatively labeled the exclusivity effect. It has al-ready been demonstrated that individuals seek to establish com-munal relationships with physically attractive people (e.g., Clark& Mills, 1979). What the present findings appear to documentis a previously unrecognized process by which an attempt is


made to create a communal atmosphere in initial, heterosexualinteractions through the restriction of conversational refer-ences to third-party individuals. The present data suggest thatthe women in these interactions may have been primarily re-sponsible for setting a more exclusive conversational tone as afunction of the physical attractiveness of their male partners.Further studies will be needed, however, to determine both theunderlying dynamics and the generality of this effect.

Fifth, the dynamic influence of the female partners' attrac-tiveness on both dyad members' behavior is also deserving offuture research. In the current study, the men who were pairedwith attractive women adopted their partner's cognitive per-spective while repeatedly attempting to engage them in conver-sation. These behaviors appear to have been influenced bythose of the attractive women, who seemed to exercise somecaution in these initial interactions. Women are taught to bewary of strange men, especially those who seem interested onlyin their looks. By breaking off eye contact and not encouraginglengthy conversation, the attractive women may not only haveevidenced the greater distrust of men that Reis et al. (1982) havedocumented but also may have required the men to exertgreater cognitive and conversational effort to make the interac-tions work. Apparently, the interactions involving attractivewomen succeeded in spite of these obstacles, as the dyad-levelcorrelations revealed that posttest measures of liking forpartner and perceived interaction quality were both positivelycorrelated with the women's attractiveness. On the other hand,in less enforced interaction contexts, the same obstacles mightserve as a kind of filtering mechanism by which attractivewomen selectively interact with the most persistent (though notnecessarily the most sincere) men.

Sixth, consistent with popular stereotype, the men in thisstudy appeared to be more greatly affected by the women's at-tractiveness than the women were by the men's. This asymmet-rical influence is most clearly suggested by the data in Table 1,where it can be seen that the significant correlates of thewomen's attractiveness outnumber the significant correlates ofthe men's attractiveness by a ratio of about 3 to 1. Although themen's attractiveness was certainly not irrelevant to the dy-namics of these interactions, as the significant correlations incolumn 3 of Table 1 and columns 1 and 2 of Table 3 indicate, itseffects did not appear to be as pervasive as the effects of thewomen's attractiveness in the present study.

Finally, although researchers have noted the importance ofstudying the long-term effects of physical attractiveness (Hat-field & Sprecher, 1986; cited in Berscheid & Walster, 1974; Reiset al, 1980,1982), relatively little is known about the specificcognitive, emotional, and behavioral correlates of physical at-tractiveness in initial, mixed-sex interactions. The findings ofthe present study provide a fine-grained, detailed look at theserelations, many of which suggest promising directions for fu-ture research.

Summary and Implications of the Shyness Data

The results from the shyness data are broadly consistent withprevious research on the role of shyness in dyadic interactions.Self-reported shyness was related in the direction predicted byprevious research to behavioral measures of interactional in-

volvement, to measures of positive and negative thoughts andfeelings, and to posttest perceptions of the interaction. Theresults also enabled several novel insights, however, that raisesome interesting issues worth pursuing in future research.

Generally, the pattern of findings differed markedly for menand women. The first and most striking gender difference wasthat the women's shyness scores were related to only two of thebehavioral measures (i.e., directed and mutual gazes), whereasthe men's shyness scores were broadly related to measures ofverbal and nonverbal behavior as well as to various types ofthought and feeling entries. Hints of a similar gender differencehave emerged in previous studies involving mixed-sex dyads(Beidel, Turner, & Dancu, 1985; Pilkonis, 1977).

Viewed collectively, these studies lend some support to thespeculation that, because social stereotypes require men to takethe initiative in initial mixed-sex encounters more often thanwomen, "the burden of shyness, particularly when manifestedin one's overt behaviors, may be much greater for males than forfemales" (Bruch et al, 1986, p. 184). Although women's shynessscores have been found to be related significantly to similarbehavioral and cognitive measures in two other studies (Cheek& Buss, 1981; Melchior & Cheek, 1990), it is important to notethat the dyadic partner in both of these studies was also awoman. This pattern of results raises the interesting possibilitythat shyness may be more of an issue for women in same-sexinteractions and more of an issue for men in mixed-sex interac-tions.

A second intriguing gender difference emerged from the anal-yses of the variables pertaining to gazing behavior. The entirepattern of data for these variables suggests that the more shy themen were, the more they managed to keep their partner fromestablishing and maintaining eye contact with them. Appar-ently, shy men can restrict opportunities for their femalepartners to initiate and terminate mutual gazes without signifi-cantly limiting their own ability to engage in the same behavior.This pattern of results leads us to speculate that eye contact is animportant marker variable that can help us understand howshyness affects the behavior of both dyad members in theseinitial mixed-sex interactions. Perhaps shy men use this strat-egy as a way of psychologically avoiding an overwhelming socialsituation when it cannot be physically avoided or escaped. If so,this avoidant gaze control may be adaptive in the short run, byallowing shy men to deal with their anxiety at their own paceand discretion.

On the other hand, the shy men's gaze control may be mal-adaptive in the long run, by creating a pattern of disturbedvisual interaction that leads both the men and the women toexperience negative, self-referenced feelings and to reduce theirlevel of interactional involvement. This interpretation is sup-ported by the findings documenting the influence of the men'sshyness on their female partner's behavior. The men's shynessscores were related to indices of the women's nonverbal behav-iors and thought-feeling content even though the women's ownshyness scores were not related to these indices. Furthermore,the men's shyness scores influenced their partner's interactionalinvolvement as much as it influenced their own.

We suggest that the shy men's gazing strategy may havehelped to mediate the negative reactions and impaired perfor-mance of the women in these encounters. Thus, in addition to


its direct effect of limiting mutual gaze at the dyad level, the shymen's unusual gazing behavior may have been responsible forsuch indirect effects as making the women feel self-consciousand uncomfortable, inhibiting the verbal and nonverbal behav-ior of both dyad members, and causing both dyad members torate the interaction more negatively. If the unusual gazing behav-ior of the shy men is indeed a major cause of the problems thatoccur in their initial interactions with women, then an inter-vention designed to normalize the gazing behavior of shy menmight also help to normalize and enhance the quality of otheraspects of their interactions with members of the opposite sex.Conducting empirical tests of the effectiveness of such an inter-vention should be an important goal of future research.

A third and final gender difference worth mentioning in-volves the relationship between shyness and focus of attention.The men's, but not the women's, shyness scores were related toboth partners' independently reporting a larger percentage ofnegative self-entries and a correspondingly smaller percentageof partner-directed entries. Thus, the men's shyness led bothdyad members to become more preoccupied with their ownanxiety and distress and to devote fewer thoughts and feelingsto their partner. These findings complement and extend thoseof a recent study by Bruch et al. (1989) in which shy men re-ported more negative self-thoughts on the Social InteractionSelf-Statement Test (Glass et al., 1982) after a contrived, mixed-sex dyadic interaction than did all other groups (nonshy men,shy women, and nonshy women). On the other hand, becauseshy women have reported being more self-preoccupied thannonshy women during a same-sex interaction (Melchior &Cheek, 1990), these findings again suggest that the nature ofthe interaction (same sex vs. mixed sex) may moderate the waythat men and women experience and express their shyness ininitial encounters.

Conclusions

There are at least four broad conclusions to be drawn fromthe results of this study. First, the results leave no doubt that twoindividual-difference variables, shyness and physical attractive-ness, have a dramatic impact on initial mixed-sex interactions.The strength and pervasiveness of this impact in the presentstudy was evident for behavioral measures of interactional in-volvement as well as for self-report measures of perceived inter-actional quality, liking for partner, and the kinds of thoughtsand feelings experienced during the interaction by the dyadmembers.

Second, the present findings reveal that the impact of shynessand physical attractiveness on an initial, mixed-sex interactionis different between men and women. Therefore, we suggestthat caution be used in any attempt to generalize across genderor from studies of same-sex dyads to studies of mixed-sexdyads. Future research should investigate more systematicallyhow the influences of shyness and attractiveness on initial en-counters are moderated by the gender of each dyad memberand by the gender composition of the dyads.

Third, the results of this study highlight the interdependencyin dyad members' scores on certain behavioral and self-reportmeasures. Our data revealed influences of the men's andwomen's shyness and physical attractiveness at both the dyadic

and individual levels of analysis. Findings such as these clearlyillustrate the limitations of experimental studies that examineonly the influence of a target's attribute (e.g., attractiveness) on aperceiver's perceptions and behavior. Although such studiesmay succeed in determining the causal ordering of the attri-bute-behavior relationship, they often fail to determine the vari-able's dynamic effects on the individual- and dyad-level behav-ior of all those involved in the social encounter. Clearly, futurestudies should recognize, rather than ignore, the problem ofbehavioral interdependence, and researchers should analyzeand report findings at both the dyadic and individual levels ofanalysis.

Fourth, the results offer additional evidence that the ex-panded interaction paradigm is a powerful and exceptionallyinformative method for studying initial interactions as they nat-urally occur. In this study, several data patterns were strikinglydiagnostic of underlying processes. It should be obvious thatthis outcome is due in part to the "triangulated" insights pro-vided by a wide range of behavioral measures and by a similarlywide range of thought-feeling content measures. Less obvious,perhaps, is that this outcome is also due in part to the deliberatelack of situational constraints on these unstructured interac-tions. By forcing the dyad members to rely on individual-differ-ence variables to provide the cues to guide their behavior, thisfeature of the dyadic interaction paradigm allows personality-based processes to stand out as figural against the ground of apsychologically weak situation (Ickes, 1982; Snyder & Ickes,1985).

As a final point, at least two limitations of the present studyshould be noted. First, our assessments of the predictor vari-ables—shyness and attractiveness—were less than optimal andcould be improved in future research. In particular, the shynessmeasure should be administered to subjects well in advance oftheir laboratory interaction, something that was precluded inthis study because of the unavailability of space on the depart-mental pretest. In addition, the objective ratings of physicalattractiveness might be made from more static videotape ex-cerpts in which no interaction was occurring and the target'sdyad partner was masked off. The practical advantages of theselatter refinements could prove to be minimal, however, giventhe nearly perfect (.93) reliability of the attractiveness ratingsand the fact that they were obtained with the sound turned offat a point in the interaction when conversation in most of thedyads had not yet begun.

Second, in any exploratory study in which a large number ofdependent variables are used, the question of Type I error inevi-tably arises. Attempting to answer this question in terms of anexperimentwise error rate is often unsatisfactory, for reasonscited by Lykken (1968). More to the point are questions aboutthe magnitude or effect size of the findings reported in thestudy (Cohen, 1990) and the extent to which these findings areinternally corroborative (i.e., form patterns of theoreticallymeaningful, mutually supportive results) and externally corro-borative (i.e., are consistent with the findings of previous inde-pendent research). Of the significant correlations reported inTables 1, 2, and 3, the average magnitude is greater than .40, avalue that Funder (1987) and Funder and Ozer (1983) have ar-gued is typical of the strongest effects in personality and socialpsychology. In addition, there is considerable evidence (re-


viewed earlier) for both the internal and external corroborative-ness of these results.

Still, the ultimate test of any effect is its replicability. For thisreason, the present findings should not be regarded as defini-tive. Instead, they should be viewed as offering some interestinghypotheses about, and some consistent evidence for, the dy-namic interactional processes that are precipitated in initialmixed-sex interactions by individual differences in shyness andphysical attractiveness.

References

Alf, E. E, Jr., & Abrahams, N. M. (1975). The use of extreme groups inassessing relationships. Psychometrika, 40, 563-572.

Aron, A. (1988). The matching hypothesis reconsidered again: Com-ment on Kalick and Hamilton. Journal of Personality and SocialPsychology, 54, 441-446.

Asendorpf, J. (1987). Videotape reconstruction of emotions and cogni-tions related to shyness. Journalof PersonalityandSocialPsychology,53, 542-549.

Asendorpf, J. (1989). Shyness as a final common pathway for two dif-ferent kinds of inhibition. Journal of Personality and Social Psychol-ogy, 57, 481-492.

Bassili, J. (1981). The attractiveness stereotype: Goodness or glamour?Basic and Applied Social Psychology, 2, 235-252.

Beidel, D. C, Turner, S. M., & Dancu, C. Y (1985). Physiological, cog-nitive, and behavioral aspects of social anxiety. Behavior ResearchandTherapy, 25,109-117.

Berscheid, E. (1985). Interpersonal attraction. In G. Lindzey & E. Ar-onson (Eds.), Handbook of social psychology (Vol. 2, pp. 413-484).New York: Random House.

Berscheid, E., Dion, K., Walster, E., & Walster, G. W (1971). Physicalattractiveness and dating choice: A test of the matching hypothesis.Journal of Experimental Social Psychology, 7,173-189.

Berscheid, E., & Walster, E. (1974). Physical attractiveness. In L. Berk-owitz (Ed.), Advances in experimental social psychology, (Vol. 7, pp.157-215). San Diego, CA: Academic Press.

Briggs, S. R., & Smith, T. G. (1986). The measurement of shyness. InW H. Jones, J. M. Cheek, & S. R. Briggs (Eds.), Shyness: Perspectiveson research and treatment (pp. 47-60). New York: Plenum Press.

Bruch, M. A., Gionlano, S., & Pearl, L. (1986). Differences betweenfearful and self-conscious shy subtypes in background and adjust-ment. Journal of Research in Personality, 20,172-186.

Bruch, M. A., Gorsky, J. M., Collins, T. M., & Berger, P. A. (1989).Shyness and sociability reexamined: A multicomponent analysis.Journal of Personality and Social Psychology, 57, 904-915.

Byrne, D, London, Q, & Reeves, K. (1968). The effects of physicalattractiveness, sex, and attitude similarity on interpersonal attrac-tion. Journal of Personality, 36, 259-21 \.

Cacioppo, J., Glass, C, & Merluzzi, T. (1979). Self-statements and self-evaluations: A cognitive-response analysis of heterosexual anxiety.Cognitive Therapy and Research, 3, 249-262.

Cann, A., Siegfried, W, & Pearce, L. (1981). Forced attention to spe-cific applicant qualifications: Impact on physical attractiveness andsex of applicant biases. Personnel Psychology, 34, 65-73.

Cheek, J. M., & Buss, A. H. (1981). Shyness and sociability. Journal ofPersonality and Social Psychology, 41, 330-339.

Cheek, J. M., & Melchior L. A. (1990). Shyness, self-esteem, and self-consciousness. In H. Leitenberg(Ed-), Handbook of social and evalu-ation anxiety (pp. 47-82). New York: Plenum Press.

Clark, M. S, & Mills, J. (1979). Interpersonal attraction in exchange

and communal relationships. Journal of Personality and Social Psy-chology. 37,12-24.

Clark, M. S., & Reis, H. T. (1988). Interpersonal processes in closerelationships. Annual Review of Psychology, 39, 609-672.

Cohen, J. (1990). Things I have learned so far. American Psychologist,45,1304-1312.

Crouse, B. B., & Mehrabian, A. (1977). Affiliation of opposite-sexedstrangers. Journal of Research in Personality, 11, 38-47.

Curran, J. P, Wallander, J. L., & Fischetti, M. (1980). The importanceof behavioral and cognitive factors in heterosexual-social anxiety.Journal of Personality, 48, 285-292.

Daly, J. A. (1978). Behavioral correlates of social anxiety. British Jour-nal of Social and Clinical Psychology, 7 7,117-120.

DePaulo, B. M., Dull, W. R., Greenberg, J. M, & Swaim, G. W (inpress). Are shy people reluctant to ask for help? Journal of Personalityand Social Psychology.

DePaulo, B. M, Kenny, D. A., Hoover, G, Webb, W, & Oliver, P. V(1987). Accuracy of person perception: Do people know what kindsof impressions they convey? Journal of Personality and Social Psy-chology, 52, 303-315.

Dion, K. (1981). Physical attractiveness, sex roles, and heterosexualattraction. In M. Cook (Ed.), The bases of human sexual attraction(pp. 3-22). San Diego, CA: Academic Press.

Dion, K., Berscheid, E., & Walster, E. (1972). What is beautiful is good.Journal of Personality and Social Psychology, 24, 285-290.

Fedlt, L. S. (1961). The use of extreme groups to test for the presence ofa relationship. Psychometrika, 26, 307-316.

Funder, D. C. (1987). Errors and mistakes: Evaluating the accuracy ofsocial judgment. Psychological Bulletin, 101, 75-90.

Funder, D. C, & Colvin, C. R. (1988). Friends and strangers: Acquain-tanceship, agreement, and the accuracy of personality judgment.Journal of Personality and Social Psychology, 55,149-158.

Funder, D. C, & Dobroth, K. M. (1987). Differences between traits:Properties associated with interjudge agreement. Journal of Personal-ity and Social Psychology, 52, 409-418.

Funder, D. C, & Ozer, D. J. (1983). Behavior as a function of the situa-tion. Journal of Personality and Social Psychology, 44,107-112.

Glass, C. R., Merluzzi, T. V, Biever, J. L., & Larsen, K. H. (1982).Cognitive assessment of social anxiety: Development and validationof a self-statement questionnaire. CognitiveTherapy and Research, 6,37-55.

Goldman, W, & Lewis, P. (1977). Beautiful is good: Evidence that thephysically attractive are more socially skillful. Journal of Experimen-tal Social Psychology, 13,125-130.

Hatfield, E., & Sprecher, S. (1986). Mirror, mirror: The importance oflooks in everyday life. New York: State University of New \fork Press.

Hill, G. J. (1989). An unwillingness to act: Behavioral appropriateness,situational constraint, and self-efficacy in shyness. Journal of Person-ality, 57, 872-890.

Huston, T. (1974). Foundations of interpersonal attraction. San Diego,CA: Academic Press.

Ickes, W (1982). A basic paradigm for the study of personality, roles,and social behavior. In W Ickes & E. S. Knowles (Eds.), Personality,roles, and social behavior (pp. 305-341). New \brk: Springer-Verlag.

Ickes, W (1983). A basic paradigm for the study of unstructured dyadicinteraction. In H. Reis (Ed.), New directions for methodology of socialand behavioral science (pp. 5-21). San Francisco: Jossey-Bass.

Ickes, W (1984). Compositions in Black and White: Determinants ofinteraction in interracial dyads. Journal of Personality and SocialPsychology, 47, 330-341.

Ickes, W, & Barnes, R. D. (1977). The role of sex and self-monitoring in


unstructured dyadic interactions. Journal of Personality and SocialPsychology, 35, 315-330.

Ickes, W, Bissonnette, V, Garcia, S., & Stinson, L. (1990). Using andimplementing the dyadic interaction paradigm. In C. Hendrick & M.Clark (Eds.), Review of personality and social psychology: Researchmethods in personality and social psychology (Vol. 11, pp. 16-44).Newbury Park, CA: Sage.

Ickes, W, Reidhead, S., & Patterson, M. L. (1986). Machiavellianismand self-monitoring: As different as "me" and "you." Social Cogni-tion, 4, 58-74.

Ikes, W, Robertson, E., Tooke, W, & Teng, G. (1986). Naturalistic so-cial cognition: Methodology, assessment, and validation. Journal ofPersonality and Social Psychology, 51, 66-82.

Ickes, W, Schermer, B., & Steeno, J. (1979). Sex and sex-role influencesin same-sex dyads. Social Psychology Quarterly, 42, 373-385.

Ickes, W, Stinson, L., Bissonnette, V, & Garcia, S. (1990). Naturalisticsocial cognition: Empathic accuracy in mixed-sex dyads. Journal ofPersonality and Social Psychology, 59, 730-742.

Ickes, W, & Tooke, W (1988). The observational method: Studying theinteraction of minds and bodies. In S. Duck (Ed.), The handbook ofpersonal relationships: Theory, research, and interventions (pp. 79-97). New York: Wiley.

Ickes, W, & Turner, M. (1983). On the social advantages of having anolder, opposite-sex sibling: Birth-order influences in mixed-sexdyads. Journal of Personality and Social Psychology, 45, 210-222.

Jones, W. H., & Briggs, S. R. (1984). The self-other discrepancy insocial shyness. In R. Schwarzer (Ed.), The self in anxiety, stress, anddepression (pp. 93-107). Amsterdam: North-Holland.

Jones, W. H., Briggs, S. R., & Smith, T. G. (1986). Shyness: Conceptual-ization and measurement. Journal of Personality and Social Psychol-ogy, 51, 629-639.

Kenny, D. A. (1988). The analysis of data from two-person relation-ships. In S. Duck (Ed.), The handbook of personal relationships:Theory, research, and interventions (pp. 57-77). New York: Wiley.

Kenny, D. A. (1990). Design issues in dyadic research. In C. Hendrick &M. Clark (Eds.), Review of personality and social psychology: Re-search methods in personality and social psychology (Vol. 11, pp.164-184). Newbury Park, CA: Sage.

Kenny, D. A., & La Voie, L. (1985). Separating individual and groupeffects. Journal of Personality and Social Psychology, 48, 339-348.

Kleck, R. E., & Rubenstein, C. (1975). Physical attractiveness, per-ceived attitude similarity, and interpersonal attraction in an oppo-site-sex encounter. Journal of Personality and Social Psychology, 31,107-114.

Laing, R. D., Phillipson, H., & Lee, A. R. (1966). Interpersonal percep-tion: A theory and method of research. New York: Harper & Row.

Leary, M. R., Knight, P. D, & Johnson, K. A. (1987). Social anxiety anddyadic conversation: A verbal response analysis. Journal of Socialand Clinical Psychology, 5, 34-50.

Leary, M. R, Kowalski, R. M., & Bergen, D. J. (1988). Interpersonal

information acquisition and confidence in first encounters. Personal-ity and Social Psychology Bulletin, 14, 68-77.

Lykken, D. T. (1968). Statistical significance in psychological research.Psychological Bulletin, 70,151-159.

Mandel, N. M., & Shrauger, J. S. (1980). The effects of self-evaluationstatements on heterosocial approach in shy and nonshy males. Cog-nitive Therapy and Research, 4, 369-381.

Melchior, L. A., & Cheek, J. M. (1990). Shyness and anxious self-preoc-cupation during a social interaction. Journal of Social Behavior andPersonality, 5,117-130.

Miller, A. G. (1970). Role of physical attractiveness in impression for-mation. Psychonomic Science, 19, 241-243.

Patzer, G. L. (1985). The physical attractiveness phenomenon. New"ibrk: Plenum Press.

Pilkonis, P. A. (1977). The behavioral consequences of shyness. Journalof Personality, 45, 596-611.

Reis, H. T, Nezlek, J, & Wheeler, L. (1980). Physical attractiveness insocial interaction. Journal of Personality and Social Psychology, 38,604-617.

Reis, H. T, Wheeler, L., Spiegel, N, Kernis, M., Nezlek, J., & Perri, M.(1982). Physical attractiveness in social interaction: II. Why doesappearance affect social experience? Journal of Personality and So-cial Psychology, 43, 979-996.

Russell, D, Cutrona, C, & Jones, W H. (1986). A trait-situational analy-sis of shyness. In W H. Jones, J. M. Cheek, & S. R. Briggs (Eds.),Shyness: Perspectives on research and treatment (pp. 239-249). New\brk: Plenum Press.

Sigall, H, & Landy, D. (1973). Radiating beauty: The effects of having aphysically attractive partner on person perception. Journal of Person-ality and Social Psychology, 28, 218-224.

Snyder, M., & Ickes, W (1985). Personality and social behavior. In G.Lindzey & E. Aronson (Eds.), The handbook of social psychology.(3rd ed., Vol. 2, pp. 883-947). New York: Random House.

Snyder, M., Tanke, E., & Berscheid, E. (1977). Social perception andinterpersonal behavior: On the self-fulfilling nature of social stereo-types. Journal of Personality and Social Psychology, 35, 656-667.

Thornton, B., & Ryckman, R. M. (1983). The influence of a rape vic-tim's physical attractiveness on observers' attributions of responsibil-ity. Human Relations, 36, 549-562.

Turner, R. G, Beidel, D. C, & Larkin, K. T. (1986). Situational determi-nants of social anxiety in clinic and non-clinic samples: Physiologi-cal and cognitive correlates. Behavior Research and Therapy, 24,56-64.

Walster, E., Aronson, V, Abrahams, D, & Rottman, L. (1966). Impor-tance of physical attractiveness in dating behavior. Journal of Person-ality and Social Psychology, 4, 508-516.

Received July 27,1990Revision received February 18,1991

Accepted February 22,1991

Documents

Shyness and Physical Attractiveness in Mixed-Sex Dyads · Shyness Shyness (or social anxiety) is defined as "the discomfort and inhibition experienced in the presence of others that