1 \ WW VIHHIWW“WWWIHIHWIIWI‘ 11111111111111111unwillumululul “Esra 1293 10612 6877 This is to certify that the thesis entitled PERCEIVED MUTUAL EYE CONTACT AS RELATED TO PERSONAL LIKING presented by JEFFREY MKHUDLWANA BEKA HADEBE has been accepted towards fulfillment of the requirements for M.A. Psychology Jegree in MM Major professor Date 3/1/83 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution v...”- ‘r. \. HERABY “mm g; a :21 5:2th 1 Univfaersty 1 3 _¢_,~,*__J’ t MSU BEIURNING MATERIAg§z Place in book drop to LJBRARJES remove this checkout from .u-cu—In— your record. FINES will be charged if book is returned after the date stamped below. 3:321? 215% r; 1 2: ’1 1 PERCEIVED MUTUAL EYE CONTACT AS RELATED TO PERSONAL LIKING By Jeffrey Mkhudlwana Beka Hadebe A Thesis Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1983 /37- 9’57 ABSTRACT PERCEIVED MUTUAL EYE CONTACT AS RELATED TO PERSONAL LIKING By Jeffrey Mkhudlwana Beka Hadebe Linkages between subjectively-estimated mutual eye contact (SEMEC) and liking were studied within seven small mixed-sex groups concerned with enhancing members' inter- personal communication skills that met about 20 times (mostly 90-minute sessions) during l0 weeks. After brief SEMEC training, each participant's SEMEC with each other group member was collected for every group's 4th, lOth, and 18th sessions. Similar but more inclusive Liked-Disliked ratings were also collected following total group experiences of 22g- and 43g-hours. Employing a 3 (occasions) l 3 (degrees of liking) L 3 (sex: FF, FM, & MM) ANOVA design was seven (one per group) entries per cell, a strong positive bond ob- tained between dyad's total SEMEC and Liking despite these data's divergent collection times base periods, and instabilities. This SEMEC-Liking bond agrees with findings from more technology-bound laboratory studies and its greater ecological representativeness encourages future explorations with appropriate subjective methods. This study is dedicated to my family ACKNOWLEDGEMENTS I would like to extend my sincere gratitude to Professor John R. Hurley, my committee chairman, without whose continued patience, support, and advisement this thesis would have never been written. I also wish to acknowledge the other committee members, Professors Donald Grummon, and Dozier Thornton for their professional insight and assistance. ii TABLE OF CONTENTS LIST OF TABLES INTRODUCTION . REVIEW OF RELEVANT LITERATURE Studies of Eye Contact Studies of Liking/Disliking and Eye Contact. Eye Contact and Individual Differences . Eye Contact and Sex Differences Eye Contact and Personality Differences Hypothesis METHODOLOGY Participants Interpersonal Groups Measures PROCEDURE Training Session . Data Collection General Data Analysis Scheme . Level of Significance and Post-hoc Measures. FINDINGS . The Range of Means Level of Significance and Post-hoc Measures. DISCUSSION . iii Page 10 n: to H‘ < 16 17 18 21 22 22 23 24 24 24 26 26 29 31 4O 4O 42 APPENDIX A . APPENDIX B . APPENDIX C APPENDIX D . REFERENCES TABLE OF CONTENTS(Continued) iv 52 53 54 55 S6 LIST OF TABLES Table 1 Raw MEC Estimates Dyadic Partners (X & Y) Within Each Group and Standard Deviation 2 Correlations of MEC Estimates Within Pairs by Group: Raw and Percentage (Z) Data 3 Pearson Correlations Between Individuals' Estimates Over Time. 4 Summary of Three- -Way ANOVA of Percentage MEC Estimates . . . . . . . . . . . . Page 32 34 38 39 INTRODUCTION The purpose of this study was to explore the relation- ships between amount of mutual eye contact (MEC), familiar- ity, and personal liking. MEC is defined as an event in which two peOple claim to have looked at each other's eyes while communicating, verbally or nonverbally. These events were measured during meetings of small, interpersonally—oriented groups for college undergraduates. Familiarity is defined by the number of hours that these groups had met. An early session, perhaps the third of a series of 20, would represent low familiarity; a middle session, perhaps the tenth, would represent more familiar— ity; and a late or near-final session in this series would represent most familiarity. Personal liking/disliking was measured by self-reports from the group members. It was hypothesized that in the earlier sessions of these groups, the participants would have relatively little MEC, for such sessions often include anxieties about getting acquainted with each other. By the middle sessions, the participants would be much more familiar with each other. It seemed probable, therefore, that their like or dislike of one another would have begun to stabilize. This increased degree of familiarity along with the crystalliza- tion of their likings or dislikings, was assumed to be reflected by changes in the amounts of MEC. Near the groups' end, familiarity would even be greater, personal bonds and/or alienations even more intense, and partners'MEC might be more stable. Because the philOSOphical orientation (Egan, 1976) of these small groups emphasizes empathy, self-disclosure, mutual acceptance, and constructive relationships, it was hypothesized that increased familiarity and personal liking for each other would be associated with increased MEC, while decreased personal liking would be associated with declining MEC between dyadic partners during these group sessions. Virtually all studies reviewed below were con— ducted in laboratories, under experimental conditions. The present study differed from them in that it was done in a more naturalistic environment where participants were more freely interacting with each other. Participants interacted in an almost natural way. Review of Relevant Literature Studies of Eye Contact Extensive research has been done on the role of eye contact (EC), mutual glance, or gaze, but virtually all of it was conducted in laboratories under experimental conditions. These studies appeared mainly between 1965 and 1975. Mehrabian (1967), a major contributor to these works, hypothesized (1) that relatively more immediate head-orientation and EC elicit stronger positive or nega- tive impressions of the communicator from the addressee, and (2) that a relatively more immediate body-orientation provides the addressee with either greater positive or negative impressions about the communicator. Immediate head- and eye-orientation was defined as an event in which the communicator's head and eyes were directed toward (H1) or away from (H2) the addressee during communications. Immediate body-orientation was defined as an act which the communicator's body slightly leaned toward (Bl) or away from (32) the addressee during communications. About 116 female university undergradutes participated in this study. The results generally supported the first hypothesis. That is, immediate head— and EC-orientation by the communicator tended to intensify both negative and positive impressions of addressees. Mehrabian (1967), like Ellsworth, Carlworth, and Henson (1972) who found that EC elicited avoidance behav- ior in humans, did not fully explicate the conditions under which the EC tended to evoke adverse effects. According to Ellsworth and Carlsmith (1968), it is essen- tial to realize that the impression given by EC also depends on the messages provided by other communication cues. In support of this point. Ellsworth and Langer (1976) wrote: The stare does have some intrinsic properties: it elicits attention, arousal, and a sense of interpersonal involvement inferred and the response perceived to be appropriate upon con- textual cues. Consequently, the behavioral response is a function of the stare and its context, or a stare in context (p. 122). This observation also supported Argle and Dean's (1965) affiliative conflict theory which states that Intimacy = EC, physical proximity, intimacy of the tOpic, amount of smiling, etc. According to Kleinke, Staneski, and Berger (1975), EC, along with physical attraction and other variables, has positive effects on interpersonal interaction. Kendon (1967) remarked that within a friendly atmosphere, EC may be perceived as a desire for affiliation and liking, while in a threatening and/or highly competitive situation, EC may be perceived as a challenge and be negatively related to liking. This observation was supported by Langer and Abelson (1972) who reported that if help is to be received, the manner in which a request is presented is more influ- ential than the nature of the favor itself. In his Encoding Experiment IV, Mehrabian (1968a) requested the participants to imagine themselves as vari- ous addressers, encoder-communicators, interacting with different addressees. They were asked to stand in a way they felt and thought they would if they were actually addressing the described addressee. Mehrabian observed that: a. Males used more EC than females in their attempts to positively impress the addressee. b. Males differed from females by showing greater BC with individuals of high status than with those of low status. c. Participants showed more BC with addressees they "liked" than with addressees they "dis— liked." This last conclusion supported findings by Exline (1963) and Kleinke, Deautels, and Knap (1977). Mehrabian (1968b) also investigated the effects of the communicator's posture, orientation, and distance from the addressee on his/her attitude toward the addressee. The findings indicated that the communicators showed more BC with the addressees they "liked" than with those they "disliked" (Exline, Gray, & Schutte, 1965). The plotted curve of these findings looked parabolic. They do not agree with Kimble and Olszewiski's (1980) observation that the actor's messages with negative or positive emotions tended to generate more EC than messages with neutral or ambivalent feelings. They also observed that emotionally intense actors' messages (negative or positive) elicited more EC than weaker or lower intensity (negative or positive) messages. The reasons underlying the dis- crepancy between these two studies are not obvious. Mehrabian's work was quite likely to receive more credi- bility, however, due to his extensive prior research in this sector. Mehrabian's (1968a) experiment was much more involved than that of Kimble and Olszewski (1980), but this greater complexity might have caused spurious results. One answer to this disagreement may be that an increase in emotional intensity (negative or positive) may accom- pany increasing EC up to a certain point, beyond which emotional intensity continues to increase while EC starts decreasing. Thus, the correlation between amounts of emotional intensity and EC may be curvilinear. It is essential to note that the above explanation may not always be relevant. For example, just before they start fighting, boxers have been observed to stare unrelentingly at each other with the apparent intent of intimidating their opponent. The importance of EC during a dyadic interaction was expressed by Simmel (1908) who wrote: By the glance which reveals the other, one dis- closes himself. By the same act in which the observer seeks to know the observed, he surrend— ers himself to be understood by the observer. The eye cannot take unless at the same time it gives. The eye of a person discloses his own soul when he seeks to uncover that of another (p. 646). While EC may help interactants to get some insight about each other, the lack of EG may also be used to conceal certain information. In the similar vein, Simmel said: Shame causes a person to look at the ground to avoid the glance of the other. The reason for this is certainly not only because he is thus spared the visible evidence on the way in which the other regards his painful situation, but the deeper reason is that the lowering glance to a certain degree prevents the other from compre- hending the extent of his confusion. The glance in the eye of the other serves not only for me to know the other but enables him to know me. Upon the line which unites the two eyes, it conveys to the other the real personality, the real atti- tude, the real impulse (p. 606). It is essential to note, however, that this observation is not always true because pathological liars, e.g., pyschOpaths, tend to maintain large amounts of EC while telling lie after lie. This observation was supported by Nielsen (1962) who said: One man's reaction to another is reflected in his eyes. The mirror of the mind, as in no other part of the face; however, not every person is expressive with his eyes and some people reveal themselves only in unguarded moments. Similarly, not everybody is able to translate the language of the eyes into ordinary, verbal descriptions (p. 128). According to Schack (1852), as cited by Nielsen (1962), "the fine and subtle expressions and meanings of the eye are difficult to catch in a phrase, they must be grasped with one's feelings, realized by way of imagination, seen and understood in nature" (p. 51). Nielson (1962) con- tinued further: Yet, without necessarily being able to describe the phenomena, people are extremely sensitive to the way in which they use their eyes, and very sensitive to whether another person stares at them. We sense very quickly if a person looks directly at us, and may get a queer and uncom- fortable feeling of being "looked at," "looked over and assessed," or "seen through." We may get this feeling very strongly even if we our- selves do not look in the direction of the other person, but merely sense his presence in the per- ipheral field of vision and indirectly know that he is looking at us (p. 129). The meaning of EC is also culture-bound. Thus, among the Nguni and Sotho natives in South Africa, it is a sign of disrespect, and even defiance, for a child to maintain EC while talking with an adult, or for a wife while talking with her husband. However, when they are together, away from other people, a wife can maintain EC while talking with her husband. In such cases, the lower- ing of EC during dyadic interaction is not indicative of any hidden information as it is generally believed in Western cultures. This is consistent with Argyle and Dean's (1965) observation, that "there are cross-cultural differences, varying from taboos on eye contact, to much greater amounts of intimacy than the common in Western countries" (p. 290). Studies on Liking/Disliking and Eye-Contact This section reviews studies on liking/disliking which are related to eye contact (BC), the main focus of this study. Exline and Winters (1965) explored the rela- tionship between affective relations and EC. They hypothe- sized: (1) that in interpersonal interactions, the par- ticipant's EC sends cues to others as to whether s/he is liked or disliked by the other; (2) that an observer can infer whether or not interactants like each other from their mutual eye contact (MEC): and (3) that persons tend to increase EC while talking to preferred interviewers, but to decrease EC while talking to nonpreferred inter- viewers. Their findings supported all three hypotheses. In another study, Exline and Winters (1965) reported that males decreased EC during critical interviews, but increased EC during praising interviews. Scherwitz and Helmreich (1973) studied interactive effects of EC and verbal content on interpersonal attrac- tion in dyads. They found that under high evaluative conditions, confederates who had low EC with the partici- pants were most liked; those who had an average amount of EC were next liked; and those who had the highest EC were least liked. Under low evaluative conditions, however, participants showed more liking for confederates with the highest amount of BC with them and showed least liking for 10 those with whom they had least EC. The latter condition appears more relevant to the groups of the present study. Kleinke and Pohlen (1971) investigated affective and emotional responses as a function of the other person's gaze and cooperativeness in a two-person game. They observed that their participants "rated the cooperative confederate significantly higher than the competitive confederate on liking preference" (p. 311). This hypo- thesis was supported and the findings also agreed with Exline's (1963) observation that "groups composed of persons more disposed toward relationships of communion than control engage in more mutual visual interaction than groups not so disposed" (p. 4). A study of Ellsworth and Carlsmith (1968) investi- gated effects of EC and verbal content on affective responses to dyadic interaction. It was found that in a dyadic interaction, increased EC during negative verbal content resulted in negative evaluations, while increased EC during positive verbal content led to positive evalua— tions. This supported the earlier findings of Argle and Dean (1965), as well as those of Kleinke, Meeker, and LaFong (1974), which found that "the gaze was an important variable affecting subjects' responses on rating forms. Gazing couples were .nated more favorably on all ten items, compared with non-gazing couples" (p. 370). 11 Argyle and Dean's (1965) affiliative conflict theory discusses the equilibrium development equation for inti- macy. The authors maintain that if one element (e.g., physical proximity manipulations) is missing from the equations in a dyadic interaction, EC may be employed to satisfy affiliation needs. To test this proposition, Pellegrini, Hicks, and Gordon (1970) hypothesized that, with physical proximity held constant, assigned approval- seeking behavior will increase EC while assigned approval- avoidance behavior will have the opposite effect. The hypothesis was supported. Earlier, Rosenfeld (1965) found that participants who played an approval-seeking role sat closer to a con- federate than did those assigned to an approval-avoiding role. This observation was later supported by the find- ings of Storm and Thomas (1977). Rosenfeld's findings for proximity could not be compared with those of Pellegrim et al. (1970), however, since his experiment included no control group. The effect of distance between friends on EC in equilibrium theory was further investigated by Russo (1975) who found that the amount of EC increased with dis- tance. Phrased differently, the amount of EC decreases as interactants move closer and closer toward each other (Argyle & Dean, 1965). Rubin (1970) hypothesized that dating couples, strongly in love, would have more BC with each other than would less 12 attracted couples. The participants were dating University of Michigan undergraduates who were taking an introductory course in psychology that required experimental participa- tion. A love scale was develOped and used to distinguish couples strongly in love from those less so. A third "apart group," composed of participants paired with opposite-sex strangers, was added. Two observers who recorded the participants' visual behavior sat behind a one-way mirror. Each observer pressed a button connected to a cumulative clock each time the individual he was watching looked across the table at the partner's face. Each clock provided measures of individual gazing. In addition, a third clock was also activated each time observers pressed their buttons simultaneously. This clock provided a measure of mutual gaze. The ANOVA results supported the main hypothesis. These findings were consistent with the idea that EC may be a projection of the exclusive and absorptive component of romantic love (Rubin, 1970), and with Freud (1955) who said "the more (we) are in love, the more completely (we) suffice for each other" (p. 140). In a more poetic spirit, Slater (1963) referred to "the oblivious lovers, who are 'all wrapped up in each other,‘ and somewhat careless of their social obligations" (p. 349). Rubin (1970) further commented: 13 One way in which this oblivious absorption may be manifested is through eye contact. As the popular song has it, "Millions of people go by, but they all disappear from view--"cause I only have eyes for you" (p. 272). According to Argyle and Cook (1976) and Exline and Fehr (1978) gaze and mutual gaze are one form of nonverbal behavior which has a great meaning in human communication. This observation was supported by Libby and Yaklevich (1973) who remarked that nurturant individuals tend to express their interest and love of others by maintaining BC with them irrespective of whether or not the same amount of love and interest is being shown in return. Ellsworth and Ludwig (1972) concluded that EC is a dyadic inter- action both indicates and calls for a certain amount of involvement. Fugita (1974) investigated the effects of anxiety and approval on visual interaction. He hypothesized: (a) that when a low-status person interacts with two higher-status individuals, one approving and the other nonapproving, s/he will have more BC with the approving than with the nonapproving person; (b) that when a high-status individual interacts with two low-status persons, one approving and the other nonapproving, s/he will have more EC with the nonapproving person; and (c) that, as a dyadic interaction progresses forward, and as the topic under discussion gradually gets clearer to the interactants, mutual EC will 14 increase or decrease depending on anxiety and approval factors. The results supported only the first hypothesis. This observation also supported findings by Efran and Broughton (1966) and Efran (1968) that people make more BC with those from whom they expect approval, and this tendency depends on the status of the person from whom approval is eXpected. Otteson and Otteson (1979) hypothesized that children more frequently looked at by their story-telling teacher remember more from the story than those who receive less gaze. Two types of stories were told: complex and simple. The hypothesis was supported. However, there were sex differences. With boys, the hypothesis was supported irreSpective of whether the story was complex or not. Of the complex and simple stories paired with high teacher's gaze frequency, girls tended to recall more from the latter than former. Teacher's gaze might have enhanced childrens' story recall because it tends to be associated with the students' self-esteem (Jones & Cooper, 1971), to call for active participation in class (Caproni, Levine, O'Neal, McDonald, & Garwood, 1977), and to be perceived as part of an evaluation (Minor, 1970). Other studies have found EC to have more or less similar effects on interpersonal inter- action. EC may convey needs for intimacy or affiliation (Argyle & Dean, 1965) or immediacy (Mehrabian, 1967). It 15 may also be perceived as a form of approval (Fugita, 1974) and approbation (Efran, 1968). Jones and COOper (1971) hypothesized that individuals who received more EC from an experimenter would feel happy about themselves, while those who received less EC would feel the opposite. The participants were 80 high school males arranged in 40 pairs. In an experimental room one member of the pair was told to read instructions adOpted from Rosenthal (1966) to the other, while the experimenter observed and rated the participants' EC from behind a one- way mirror. Privately, the experimenter told 20 readers to keep a high frequency of EC with their partners while reading instructions. The remaining 20 were told to main- tain less EC with their listener-partner. At the end of the experiment, the listening partners were requested to rate their own feelings on a scale with polar-adjectives designed to assess their feelings. The hypothesis was supported. The participants were then given neutral photographs to rate as successful or unsuccessful. Those participants who had received more EC from their partners gave significantly more positive ratings than did those who had received less EC. Citing Goffman and Simmel about the role of EC in social interaction, Friedman (1967) pointed out that "a prosaic deduction from these paeans to the mutual glance would be that of exchanging glances 16 with the experimenter, the subject is given a 'good' experience; he feels accepted; he feels that their inter- action has been 'warm'" (p. 55). Eye Contact and Individual Differences Several studies have reported large individual dif- ferences in the amount of EC during dyadic interactions. Kendon (1967) found that there are subjects who spend as much as 65% of their speaking time in looking at Q (the other participant in a dyadic interaction), or as little as 20% of it. Or there are subjects who look at Q while they are listening for only 30% of their listening time, or they look for over 80% of it (p. 26). Weisbrod's (1965) participants were found to spend more than 70% of their time speaking time looking at others, but only 47% of their listing time looking at a speaker. Kendon and Cook (1969) studied subjects in daydic conver- sations. They found that EC patterns were fairly consist- ent aspects of the interactants' behavior, however, this varied as a function of the identity and also possibly with the participants' liking for the other partner. Libby (1970) studied EC and direction of looking at stable individual differences. Participants were indi— vidually asked 54 questions and were requested to main- tain BC with the interviewer until she finished asking each question. Four ocular responses by an interviewer to specific questions were identified: 17 1. Maintained eye contact--scored 1 if a partici- pant maintained EC with the interviewer through- out the duration of his response; scored 0 if he did not. 2. Broke EC-—scored 1 if a participant broke EC with the interviewer by looking away before the interviewer finished speaking; scored 0 if he did not. 3. Directional responses in vertical plane--scored 1 if a participant looked up after the inter- viewer finished speaking; scored -1 if he looked down; and scored 0 if neither occurred. 4. Directional response in horizontal p1ane--scored 1 if a participant looked at the right after the interviewer finished speaking; scored -1 if he looked at the left; and scored 0 if neither occurred (pp. 305-306). All four looking variables demonStrated.inter- observer reliability, intra-individual consistency, and temporal stability over the course of the interview. Eye Contact and Sex Differences Exline and Thibaut (1961) investigated visual inter- action in relation to Machiavellianism and an unethical act. They found that in verbal or nonverbal dyadic inter— actions, women engaged in more mutual EC than men. This 18 observation agreed with Exline, Gray, and Schutte's (1965) findings. Like Exline and Winters (1965), Kendon and Cook (1969) found that women look more and were looked at more, but only under certain conditions, e.g., women look more while listening if they like the person, whereas men look more while speak- ing if they like the person. There is evidently no universal tendency for women to look more or be looked at more (p. 492). However, the tendency of men to look more at females was frequently observed. Argyle, Lalljie, and Cook (1968) reported that women were less comfortable talking to a participant whom they could not see nor maintain mutual BC with. Exline (1963) reported that (1) visual information plays a greater role in the social field of women than of men, and (2) womens' visual activity is more sensitive to social field condi- tions than is men's. Ellsworth and Ludwig (1972) concluded that in general, these studies indicate that in a neutral or positive interaction at a given level of intimacy, females engage in more eye contact and possibly depend more on visual feedback than do males (p. 380). Eye Contact and Personality Differences Exline (1963) found that women with a high need for affiliation maintained more mutual EC when talking than did women with a low need for affiliation. This finding was 19 supported by Exline, Gray, and Schutte (1965) who found that participants high on measures from Schutz's (1958) Fundamental Interpersonal Relations Orientations (FIRO) test had more BC with interviewers than did participants low on both dimensions. However, Kendon and Cook (1969) found that "of 90 correlations involving FIRO only five were significant, almost exactly what would be expected by change. There was no tendency for affiliation, wanted or given, to be related to amount of looking" (p. 493). Gray (1971) also used FIRO to investigate the amount of EC as a function of need for affiliation, need for dominance, and sex of the participant. He found that dominant female participants maintained more EC with an interviewer than did dominant male participants. However, dependent participants of each sex maintained equal amounts of EC with the interviewer. No correlation was found between the affiliation need and the amount of partici- pants' EC with interviewers. Efran and Broughton (1966) studied the effects of expectancies for social approval on visual behavior. Need for approval was measured by the Crown-Marlowe (1964) scale. They found positive correlations between need for approval and amount of EC. However, Efran (1968) could not repli- cate this result. Thus Ellsworth.amiLudwig (1972) con- cluded that "the relationship between affiliation/inclusion/ 20 approval need and eye contact continues to be dubious, and probably depends largely on some third variable, such as the subject's expectations of approval" (p. 381). Hobbs (1969) investigated EC and introversion- extraversion. He found that extraverts maintained more EC than introverts, whether talking or listening. Argyle (1965) observed that dominant and/or socially poised indi- viduals looked more at others than did submissive/or socially anxious persons. Exline and Messick (1967) investigated relationships between dominance-dependence orientation and social rein- forcement on one hand, and the amount of EC on the other. They used FIRO scales to differentiate between dominant and dependent participants. Results showed some inter— action. Dependent participants showed more BC with inter- viewers when given low (as compared to high) amounts of verbal social reinforcement. Their EC was also higher than that of the dominant participants given the same amount of such reinforcement. A number of studies have commented on the artificial- ity of the continuously-staring confederatecn:interviewer. Kendon and Cook (1969) remarked that "most studies have employed a continuously gazing confederate, so that when- ever the subject looks at the confederate, eye contact occurs; this is a rather artificial situation" (p. 481). 21 Libby (1970) also agreed with this observation. He further commented that this is "an unnatural situation which may elicit unnatural responses” (p. 304). Similarly, Vine (1971) argued that in natural interactions we either give EGs (Eye Gazes) or we look at other target in the head region rather rarely. If this is the case, then observers need only be able to discriminate FGs (Facial Gazes) from non-FGs (non-Facial Gazes), which is relatively an easier task (p. 329). Nearly all the studies reviewed here were conducted in laboratories, under experimental conditions. The present work differed in that it was done in a naturalistic environ- ment where individuals interacted with each other much more freely. No variables were deliberately manipulated, so the degree of familiarity was not manipulated. Participants interacted in an almost natural way. There was no continuously-staring confederate or interviewer, as in the foregoing studies. In a way, Vine's (1971) suggestion that "observers need only be able to discriminate FGs from non- FGs" was accommodated in this study. Another point of difference is that the amount of BC was recorded by an observer in these previous studies, whereas in the present work it was directly estimated and recorded by the participants. Hypothesis: There will be positive correlation between amount of mutual eye contact (MEC) and degree of personal liking among dyads during meetings of small interpersonally— oriented groups. METHODOLOGY Participants The persons studied were male and female participants in undergraduate groups oriented toward increasing inter- personal communication skills during Spring term of 1982. Students participated in these growth groups as partial fulfillment of the requirements of a 400 level psychology course at Michigan State University. Enrollment was open to juniors and seniors from any major. This research was conducted concurrently in seven groups. Their composition was as follows: Male to Female Group Male N Female N Total Ratio A 3 4 = 7 0.75 B 3 2 = 5 1.50 C 4 2 = 6 2.00 D 4 4 = 8 1.00 E 4 3 = 7 1.33 F 3 4 = 7 0.75 G _3_ ___ = _§ 0.60 Total 24 24 48 22 23 The Interpersonal Groups These groups were primarily educational, rather than psychotherapeutic, in nature. Their main purpose was to provide a favorable climate in which participants could examine and possibly enhance their interpersonal skills. This included learning about and expressing empathic under- standing, caring, warmth, acceptance, confrontive skills, etc. Twice during the term each participant was also required to describe his/her perception of self and each other group member. These self-descriptions were then compared with the descriptions of how each other member of that group perceived him/her so as to highlight congru- ities and discrepancies. Each group had two 90-minute sessions weekly plus two twelve-hour-long marathon group sessions near the term's third and seventh weekends, totaling about 50 hours of experiential meetings over nine weeks. Each group was led by an undergraduate facilitator or pair of facilitators who had at least one prior term of training in group leadership in addition to having earlier been a member of these groups. Apart from their group sessions, group leaders had an additional two-hour weekly staff meeting with the course instructor. 24 Measures This study used two different tools: the Estimate of of Mutual Eye Contact in Seconds (EMECS) and the Liked- Disliked scale. Data generated by the latter were used to select dyads whose mutual eye contact (MEC) estimates, in seconds, would be analyzed. The EMECS (see Appendix A) was jointly develOped by the researcher and his faculty advisor, John R. Hurley. Its purpose was to yield perceived estimates of the amount of MEC that participants had maintained with one another during predetermined group sessions. The Liked-Disliked measure (see Appendix B) is a scale that precedes Hurley's (1978) eight other scales designed to assess Self-Acceptance vs. Rejection and Acceptance vs. Rejection of Others. These latter measures were not part of the present study. Anchored at its posi- tive pole by Likgd and at its negative pole by Disliked, this instrument was presented in semantic differential form with 10 equally-spaced intervening checkpoints and scored from 0 to 9. Each participant rated how much s/he liked self and every other group member (see Appendix B). Procedure Training Session The purpose of the training session was to enhance the likelihood that group members would make meaningful 25 MEC estimates. All seven interpersonal groups were com- bined in one classroom where they were asked to sit in seven rows. Seven randomly selected participants were each asked to form a new group by choosing one unfamiliar person from each row. This led to the formation of seven training groups. The members of each group were then asked to sit together in some area of the classroom or adjacent locations. The researcher then gave each member an index card on which to list all training group partici- pants' names. He then asked them to talk to one another about what they eXpected from their regular 90-minute groups. This discussion lasted about 15 minutes. Then they were asked to return to their original seats. The researcher then requested them to estimate the seconds of MEC that they had with each other group member during the immediately preceding 15-minute meeting. Estimations of MEC from all members of each group were combined in matrix format on a single index card. The researcher then encircled the dyad or dyads having the greatest MEC estimate discrepancies and also the dyad or dyads showing the least discrepant estimates. At the sub- sequent general class meeting,one week later, each partici- pant received a feedback form that displayed all MEC estimates attributed ("given") to each partner as well as those assigned by (received from) partners in addition to the total estimates "given" by each individual. This 26 informed each participant of all discrepancies between his/her own MEC estimates and those of all training group partners. A general discussion of these data and a review of factors that might have contributed to these discrep- ancies followed. The participants were then advised that similar MEC estimates would be collected toward the end of some of the regular 90-minute sessions of their small groups. They were not told, however, which sessions would be selected for this purpose. Data Collections MEC estimates were later collected from all seven interpersonal groups, each of which met for about 20 sessions. At the end of an early session, the fourth or when each group had met for about six hours, the researcher administered the EMECS and collected these forms. The same process was repeated at the end of a middle session, perhaps the tenth, or when each group had met for about 28 hours. This process was again repeated at the end of a late or near-final session, perhaps the 18th, or when each group had met for about 49 hours. So the EMECS was admin- istered to each group on three occasions. General Data Analysis Scheme Ratings of how much the group members liked each other were collected on two other occasions: (1) following each group's first postmarathon I session (after about 27 22%-hours of group participation), and (2) again after each group's first postmarathon II session (after about 43%-hours of group participation). The latter set of liking ratings was based on fuller acquaintance and knowledge of each other, since the prior ratings had been discussed within each group. Thus, only the second set of liking ratings was used to identify pairs of members within each group who displayed high or moderate mutual ratings of liking or disliking. The instructions used with each administration of the Liked-Disliked scale had specified that these ratings were to be based on the ratee's "actual behavior within the group sessions up to now,’ implying a 22%-hour base for Time I ratings and a partly overlapping 43%-hour data base for Time II ratings. In contrast, MEC estimates were exclusively based on behavior during those 90-minute group sessions that immediately preceded their collections at 6-, 28-, and 49-hours. Dyads representing different degrees of liking were identified as follows: 1. The researcher examined the second set of Liked-Disliked ratings which each group member completed for every other member of that group. As an illustration, Appendix B shows the ratings of one participant (Mark) in Group D. 28 2. These data were next assembled in a matrix that displayed how each partner of all possible dyads rated the other member (see Appendix C). The dyadic part- ners were arbitrarily designated X and Y. Thus, each pair had two ratings, namely, X's rating of Y, and Y's rating of X. For example, Appendix C shows that Mark's (X's) rating of John (Y) = 7 (the second value in the first row), while John's (Y's) rating of Mark (X) also = 7 (the first value in the second row). 3. Within this matrix, all male-male, male-female, and female-female dyads were identified and rearranged into three columns, Partners' liking ratings were summed to yield a single aggregate value for each pair. These sums were then used to rank-order the pairs within each column (see Appendix D). 4. Within each column, the researcher selected dyads with the highest, middle, and lowest liking ratings. The "middle" rating was defined by the median. If two or more pairs fell at the median point, the pair with the least interpartner discrepancy was selected. For example, column three of Appendix D shows a median with two differ- ent pairs: Suzi and Marie (5 + 8 = 13), and Leslie and Lisa (6 + 6 = 12). The latter was chosen since the part- ners' ratings had zero discrepancy. If two or more dyads again formed a tie on this respect, one pair was selected 29 by a random method. For example, Appendix D column one shows three different dyads which were at median point: 14); Mark and Paul (8 + 6 = 14); Mark and John (7 + 7 John and Paul (7 + 7 = 14). The discrepancy criterion eliminated the middle dyad. One of the remaining two was selected randomly. 5. This process was repeated in the remaining six groups, resulting in a total of 21 dyads. MEC estimates of the selected 21 dyads were analyzed, using 3 x 3 5 3 ANOVA technique, over three time periods (early, middle, and late), to determine whether time period, degree of personal liking, and sex had any statistically significant effect on perceived MEC esti— mates. The SSPS computer program was employed to calcu- late all pertinent Pearson pmoduct-moment correlation coefficients. Level of Significance and Post-Hoc Measures Since this was an exploratory study, an alpha level of .05 was used to test the significance of an F ratio. To check which means were significantly different from one another, the Least Significant Difference (LSD) tech-. nique was employed using the following formula: LSD where Sd lfi ll MSE = This method 1. 30 = t a/2 Sd standard error of the mean difference mass/J);5 t-distribution table a chosen two-tailed alpha level denominator used to calculate F number of observations in each mean of multiple comparison were chosen because: In it, comparisons need not be orthogonal (independent) It works well with no more than four com- parisons LSD sets the Type I error rate around each contrast (Winer, 1971). FINDINGS Table l gives the mean and the standard deviation of partner's raw MEC estimates by each group for the three data collections. Thus, the 6-hour MEC estimates of Group D's partners X and Y averaged 64 and 46 seconds, respectively, with dramatically different standard deviations of lll and 50. This group showed the largest discrepancy between partners' 6-hour MEC estimates. Group G's 6-hour estimates had the smallest mean discrepancy. The same—time differences of 2's and 7's among these seven groups showed much intergroup MEC variability. Their relatively large standard deviations revealed much variation of partner's MEC estimates within each group, especially at 6-hours. Table l also shows that 6-hour raw MEC estimates con- sistently exceeded those made subsequently. Thus, Group A's 6-hour estimates were quite high (l7 and 5l) as compared to those at 28—hours (means = 6 and 4). In other words, the correlation between the total mean raw MEC estimates and the total number of hours of group experience that preceeded the data collections was somewhat curvilinear. In all remaining groups, the relationship between the sum 31 Table l Raw MEC Estimates of All Dyadic Partners (X & Y) Within Each Group on Each Occasion Groqp Mean Mean Discrepancy (50y) Means Time N of Pairs Grou (erx I 1 00000 [\d'm AAA .—o sea owe. om. mH. uwm. No.1 Nu. mm o «m. mH. moN. mm. Hm. mm. Hm a mm. mN. am. no.1 mm. ma. NN m fiwmml_ «N. oh. mm. no.1 mm. mm o .Nm. . NH.1 NH. SN.1 mm. No.1 ma 0 ma. oo. ma. NH. as. mm. 0H m was. No.1 cos. No. com. mo. HN < N mmm N mmm. N mmm. mufimm no 2 mmmmw musomlmq musomlwm musomlo mama mo ANV omeCoohom cam 3mm "mmouv an mwflmm CH£DH3 mommaflumm om: wo macaumamuwou N memH 35 was statistically significant, although the total set of 144 dyadic partners correlated significantly on each occasion. Overall, there was little stability in partners' raw MEC estimates within or across these time periods. Table 2 also shows correlation coefficients derived from the conversion of raw estimates into percentages. Each percentage represented the portion of his/her total MEC estimates that a person reported giving to each partner in that group. Most of these correlations (16 of 21) exceeded those derived from the raw MEC estimates. For example, using raw data, the correlation coefficient of group G at 49-hours was .29, whereas using percentage estimates, the same dyads' estimates correlated .48 (p_< .05). The percentage data also showed more stability among mutual (pairs') estimates with increasing contact over time, especially at 49-hours. The raw estimates did not show this tendency. However, even these percentage data showed quite limited stability for 6- and 28-hour estimates, but modest stability at 49-hours. Despite these limitations the percentage data had several advantages over raw MEC estimates. The raw estimates had many large interpartner discrepancies that tended to markedly lower their Table 2 correlations. These extreme differences were reduced by converting the raw estimates into percentages. As a result, the 36 correlations between partners increased. In three cases, correlations which had not been statistically significant increased sufficiently to achieve this significance. For example, using raw data, Group A at 49-hours had a -.04 correlation, but this correlation for percentage data increased to .66 (p < .001). Thus, the percentage data yielded several significant correlations between part- ners when raw estimates had not. These increased correla- tions demonstrated that this conversion of individual's raw estimates into percentages yielded greater interpartner consistency. Percentage data also showed increasing stability of mutual estimates at 49-hours. In addition, to get some sense of what was going on in individual groups, the median correlations were also considered. Within each time period, the percentage data yielded a higher median group correlation than did raw MEC estimates. How correlations between individuals' estimates shifted over time is shown in Table 3. For raw MEC, the low stability of individuals' estimates was shown by the following mean correlations: 6- vs. 28-hours = .19 [(.28 + .10) + 2], 6- vs. 49-hours = .01; and 28- vs. 49-hours = .32. The comparable percentage correlations were: 6- vs. 28-hours = .28; 6- vs. 49—hours = .39; and 28- vs. 4'9-hours = .30. These latter data showed more stability among individuals' MEC estimates over time. 37 Table 3 shows that raw data yielded relatively lower correlations than percentages, reflecting lesser stability among individual's estimates over time. On the other hand, percentage data yielded relatively higher correlations. These several advantages of percentage data over raw estimates led the researcher to adopt the former as more promising for the purposes of this study. The correlations between the summed pairs' Likgg- Disliked ratings of partners (3 = 63) from Times I to II was .63. This accounted for about 40% of the total vari- ance of liking ratings. Table 4 shows the summary of the ANOVA findings. The most powerful main effect was for liking and a lesser, but statistically significant, main effect concerned sex. There were not significant effects for time, which is not surprising because the use of percentage MEC estimates eliminated this finding so evident in the raw MEC scores. None of the interactions were significant. The Range of Means Among the nine selected pairs per group, the mean percentage MEC estimates over time were: 6-hours = 31.78; 28 hours = 28.57; and 49 hours = 31.57. So the overall correlation between time (i.e., hours) and MEC estimates 38 umwu vawmuuo3u wcflms OH.V mo ummu vaNmulo3u wcwm: mo.v.mU mss mom osN mss nNN sNN mom QNN nNN osN mNN QNN mom uoN osN NH nNN mom ssH «No mom uoN oo- No oHN mNo Ho- mom oo No- UHN ooH oH oo osH mHo oo oo so osN omw oNN mH mo N m N m musoslms mHSOSIwN mmumaaumm mwmucmopmm mom QNN uoN mom nsN osH NH mNm pom Dom omH oo so- now No oo- NH omN mom HH uoN «mm N m mpsonuo mom oHN uoN «No oo- Ho- No moN omH so- oo pom moN mom 2mm Ho NH mom UNH- mo N. a. N N muaonlmq ummu vonmulo3u wcfim: Ho.v.mo ummu vaHmu103u wcfima Hoo.v.mm oo oH No- oo oHN osH ooH mHo No NH oo- pom pom NH Ho mom mNm «No oH osH uoH mom N N muaonlwm mmumefiumm 3mm .omuuNEO mHmeHomo HHo mmumaflumm .meSvN>NUCH cmmzumm maowumamhhoo Cowpmwm m manme Naoe-os m evesoeuom m M wNDOLIo “wusosuom musonlo Table 4 . Summary of 39 Three-Way ANOVA of Percentage MEC Estimates Source df MS F Probability Main Effects Time 2 202.89 1.29 .279 Sex 2 721.59 4.58 .012 Liking 2 1402.05 8.90 .001 Two-Way Interactions Time x Sex 4 159.69 1.01 .402 Time x Liking 4 22.10 .14 .967 Sex x Liking 4 177.25 1.13 .346 Three-Wgy Interactions Time x Sex x Liking 8 151.54 .96 .468 Total 188 174.56 40 was slightly curvilinear for these selected dyads. The mean percentage of male-male MEC estimates among the 63 selected pairs was highest (34.51); that of female-female estimates was somewhat lower (29.19); and that of male- female estimates was the lowest (28.22). There was a strong positive relationship between percentage MEC estimates and the increasing degrees of liking. At low, medium, and high degrees of liking, the mean percentage estimates were 25.83, 30.84, and 35.25, respectively. Level of Significance and Post-Hoc Measures In each significant main effect, the Least Signifi- cant Difference (LSD) technique was employed to identify means which were significantly different from one another. The critical value was found to be 4.61 for all comparisons. Regarding liking, only two comparisons were signifi- cantly different. These were: medium liking (30.84) versus low liking (25.84) = 5.0 > 4.61 and high liking (35.84) versus low liking (25.84) = 9.41 > 4.61. The third comparison, high liking (35.25) versus medium liking (30.84) = 4.41 < 4.61, closely approached significance. Regarding sex-pairing, only two comparisons differed significantly. These were: male-male (34.61) versus 41 female-male (28.22) = 6.29 > 4.61, and male-male (34.61) versus female-female (29.19) = 5.32 > 4.61. Thus, male- male mean MEC estimates differed significantly from the other two means (female-male and female-female). DISCUSSION Nearly all prior studies of eye contact or gaze have been conducted in laboratories under experimental condi- tions. The present work departed sharply from this tra- dition in being conducted within a more naturalistic setting while individuals interacted rather freely. In this study no variable was deliberately manipulated, including the degree of liking. There was no continuously- staring confederate or interviewer, as in the traditional type of eye contact study. In a way, Vine's (1971) sug- gestion that "observers need only be able to distinguish FGs (Facial Gazes)" was accommodated in this study. Another notable difference was that the amount of EC has traditionally been recorded by an "objective" observer, whereas in the present work participants' own MEC esti— mates were used. Thus, the present study dealt with esti- mates, rather than with objectively rooted measures of MEC. Subjective MEC estimates may well be more meaning- ful than objective estimates, especially in naturalistic settings, because it is the interactants themselves who can best say whether or not they had maintained MEC. This seems true when one considers the particularly elusive 42 43 character of MEC. For example, one individual may main- tain EC with her partner in the hope that the other person is doing the same, only to later discover that the other was just trying to give the impression of fixing his eyes on the former's, with his true attention being elsewhere. Another point of departure was that the times of the collection of liking and MEC estimates data did not only differ but also overlapped. These discrepancies and over- laps were an added special feature of the present study. They extend the existing body of knowledge because such data have traditionally been related to behavioral data collected in close temporal proximity. For example, Kleinke and Pohlen (1971) collected MEC data and then imme- diately administered their liking preference tool to par- ticipants. Similarly, Jones and Cooper (1971) administered a self—liking questionnaire and photographic ratings imme- diately after gathering MEC data from the participants. "Objective" MEC reports may often be even less relia- ble because of the inherent difficulty of judging whether partners, at any moment, have MEC or are just looking at each other's eyebrows, midnose, etc. Subjective MEC esti- mates have some similar limitations because an estimator may not always be aware that he/she has maintained MEC with a partner. Even when aware, she may sometimes under— or over-estimate her reports and/or forget moments of 44 genuine MEC. So the present work departs in two major ways from the more traditional and conventional approaches to the study of MEC. The only evidence of the validity of the subjective MEC estimates used here resides in two factors: (a) agreement between partners about the extent of their mutual eye contacts; (b) evidence that this sub- jective MEC measure related in reasonable ways to independent measure of personality or behavior, as represented by the Liked-Disliked ratings. Partners' raw MEC estimates often differed sharply, indicating that this measure yielded little dyadic agreement and high variability. These raw data also showed relatively large differences among the seven groups within each period and also among grand means at different times. Thus, these grand means at 6-, 28-, and 49-hours were, in seconds, 31.78, 11.67, and 9.46, respectively. Raw estimates also evidenced little stability over time, as the correlations of individuals' MEC estimates from one observation to the next for a given partner were generally quite low (mean 3 = .17). The mean raw MEC estimates of groups were also quite variable, as group C's estimates were consistently high, while Group F's three means were low. These differences not only suggested the influence of important effects of both groups and/or experience, but also highlighted the instability of raw estimates over time. At 6-hours, 45 these group means ranged from 13.5 to 55 seconds. Thus, the raw MEC estimates proved problematic. To achieve a more stable measure, these raw estimates were transformed into percentages which yielded notably less variability among individuals and dyads as well as being intrinsically more stable across groups and times. These percentages preserved differentiations among the estimates that individuals assigned to each of their part- ners. The correlations of individuals' percentage MEC estimates over time averaged modestly higher (mean 3': .32) than the corresponding raw estimate value of .17. These percentages were also analyzed at the level of the mean percentage estimate received by each individual from all others. These 6-hour means correlated .60 (p_< .001) with mean 28-hour MEC, and .57 (p_< .001) with mean 49-hour MEC. The latter two MEC estimates correlated .44 (p < .01). These percentage MEC estimates were more stable over time than their raw estimate counterparts: 6-hour means corre— lated .50 (p < .001) with 28-hour MEC, and .14 (p < .05) with 49-hour MEC. The two latter MEC estimates correlated .51 (p < .001). Thus, the average of these correlations was .54 for percentages versus .38 for the raw estimates. Before discussing the input of Liked-Disliked ratings at Times I and II, it seems important to again note the discrepancies between the time of their administration and 46 that of the MEC data collections. The liking data were collected after about 22%-hours (Time I) of group partici— pation and again (Time II) after about 43k-hours (Time II) of group interaction. MEC estimates, however, were col- lected toward the end of each of the three selected 90- minute group sessions: 6—, 28-, and 49-hours. In addition to these temporal differences at their points of adminis- teration, they also had different base periods. Thus, each MEC collection was based only on behaviors during the 90 immediately preceding minutes while Liking I data were based on a much longer period (22k-hours) of group partici- pation and Liking II data had the even more extended 43%- hour base period. The instructions had specified that Liking ratings were to be based on the ratee's "actual behavior with the group sessions up to now," implying a 22g-hour base for the Time I ratings and an overlapping 43%-hour data base for Time II ratings. Linkages observed between such divergently grounded sets of data would appear especially meaningful. The means of dyads‘Liking ratings at Times I and II were 13.50 and 14.11, respectively. The total Liking rat- ings of partners correlated .63 (p < .001) from Time I to Time II. Both findings indicate that pairs' liking held reasonably stable while modestly rising over the interven- ing four-week period that included about 21 hours of active 47 group participation. Despite the restricted temporal sta- bility of both the MEC and Liking measures, a strong posi- tive linkage held between partners' percentage MEC esti- mates and partners' liking. This agreed with Exline and Winter's (1965) report that persons tend to increase eye contact while talking to preferred interviewers. It also seemed congruent with Rubin's (1970) observation that dating couples strongly in love manifested more MEC than did less attracted couples. A lesser, but statistically significant, main effect concerned sex. The mean percentage of male-male MEC esti- mates among the 63 selected pairs exceeded that of both other pairs (male-female and female-female). This showed that males maintained more MEC with other males than with females or than did females to other females. This find- ing might be a chance deviation since it barely exceeded .05 significant level. It also seems incongruent with the outcomes of related studies. Kendon and Cook (1969), like Argyle and Winters (1965), "found that women look more and were looked at more" (p. 492) by other participants in their study. Exline and Thibaut (1961) found that in verbal or nonverbal dyadic interactions, women engaged in more MEC than men. This observation agreed with Exline, Gray, and Schutte's (1965) findings. In daily life, then, females seem to engage in more EC and possibly depend more on visual feedback than males do. 48 Although time did not emerge as a significant effect from the variance analysis, this is obviously attributable to the use of percentage MEC estimates that totally sup- pressed this factor. The raw estimates showed that time was highly pertinent. Thus, the raw grand means at 6-, 28-, and 49-hours were 31.78, 11.67, and 9.46, respectively. The ANOVA findings permitted no valid statement about the effect of time for it was based on the percentage MEC measure that precluded the possiblity of any time effect. This happened because the transformation of the raw data into percentages created the same common base (100%) for each respondent on each occasion. It is also true that the effect of time on the raw MEC data was confounded with the participants' experience in the usage of Estimation of Mutual Eye Contact in Seconds (EMECS) instrument. The groups' shifts of mean raw MEC estimates (see Table 1) suggest that the general pretraining MEC estima- tion session that all class members received early in this term was quite limited in its effectiveness. It is also known that, at the beginning of their respective data collections, these groups received nonuniform instructions for MEC estimation. Thus, just before making each of their MEC estimates, Group B received the described 5- second MEC estimation exercise. This likely played a role in the relatively low MEC estimates provided by this group 49 (means in seconds, of 26.5 at 6-hours, 8.7 at 28—hours and 6.1 at 49-hours). Compared to the means of other groups, Group B's were generally low and showed more sta- bility over time. Although it is not known precisely what additional training was provided to each of the other groups, this training doubtlessly varied from group to group. Such differences may have played some role in the emergence of one group (B) as consistently low in MEC estimates while another group (C) was consistently high on all occasions. The present study's character was relatively uncon- trolled in the sense that the MEC measures were relatively crude and administered under somewhat irregular conditions that varied in terms of preparation and instruction from group to group; and also the MEC estimates were sub- jective in nature. Despite all this, the MEC estimates nevertheless related to liking at Time II in the expected way. These findings augment the tentative evidence of validity of these subjective estimates that rested on the agreement between partners' estimates. Con— sequently, these findings appear to be something of a break- through with respect to demonstrating the efficacy of subjective MEC estimates in relationship to naturalistic behaviors. They suggest that with further training and preparation, similar measures might be used to study a 50 variety of other social behaviors, instead of largely confining MEC studies to the relatively sterile and eco- logically unrepresentative context of the laboratory. In addition, the present evidence suggests that feedback about initial incongruities in partners' MEC estimates assisted their stabilization. One specific next step in this research would be to try to distinguish between the effects of training and feedback. The present feedback procedures appeared to have some appreciable impact in the sense that MEC estimates regularly declined and it is not clear that this would have happened without feedback. But in the present research design, this was confounded with both the amount of MEC estimation experience and also with increasing familiarity with partners and measures. It is unclear whether the MEC estimates would continue to remain about the same, orrise or fall, if these latter variables were controlled. Another area deserving research exploration concerns the broader behavioral correlates of MEC, as there is much evidence suggesting that MEC may relate to hostile versus affirmative behaviors. Another study might test the effect of training on subjective MEC estimates, where some groups may receive thorough practice in estimating MEC while control groups receive none. Training might be used as an independent variable with three levels: standardized, nonuniform, and none. Additional studies might separate 51 familiarity from training, and others might include dif- ferent programs for different groups, e.g., groups equal in familiarity but different in training. This study demonstrated that these apparently crude estimates yield meaningful information. Now the study can be extended to see how they relate to more general measures of interpersonal behavior, such as a more compre- hensive set of behavior ratings, e.g., Hurley's measures of (1978) Acceptance vs. Rejection of Others and Self- Acceptance vs. Rejection, or other related measures like Benjamin's (1979) Structural Analysis of Social Behaviors, etc. Yet another approach might consider other nonverbal behaviors, such as facial expressions, gestures, etc. Another interesting avenue of research might be an attempt to answer the question: "What are the general character- istics of individuals who markedly appear to over-estimate their MEC versus those who markedly under-estimate it?" 52 HH coumx camao H cmumx Hfifim ooc< uu< o>mo gog Huuoh +og No on om s os on N am gN MN oN H H sH NH oH o o N o m s m N H o muonamz msouu A.HHoo oUMHHQOHaam osu xomno paw we Hoodoo maummc .Nonuo cm oxmfi on among: 90% MHV .wEmc Sumo £uw3 wcwpcoamouuoo Sou onu CH oommm oumfiuaoumam onu xooso Am .mpcooom SN on Ema: cowumfiwumm mNLH .cowmmom ouncfiauom wasp wcwusp comuoo some suNB pocflmucwma so» xcflnu 50% uomuaoo who Hmdudfi mo unboem gnu mumEHumm AN .oEmc ozo Hook mHoHNoom AH .mHoQEoE osonw N50» mo moam: msu m3o£m unmannumsfi mwsu mo nEDHoo pcmgnuwma one "mGOHuoDuumoH Amomzmv mpcoomm CH oomucoo oxm Hmsusz mo moumENuwm < xwpcmoo< 53 .xpsum mHfiu 5H pom: mmB Ao>onm owmv mHmom pmxHHmH91poxHH one cho swoonuHm .anopw some on Cw>Hw .mumeooanHE mamQ:OH mLu mo omen umuHm mzu wouSuHumcoo mcowuosnumcw mwo£H« :3 o H H H H H H H H H H o c: gmomgggggggggmv 3 3.3: N a a o N. a w o w Q m m Co 83 H g o g g a g Q g g g on Hsmm A m m o o o m o o @ m M HE 385 m o o o o o o o o © o Auv 6335 o o o o 0 © 0 o o u H H Ago egos o g g g g g g g @ g g g Amy \. xumz m m w m. m m m. @ m m. .mMmcommou Hmcomuom Hack wuHoNHOm onom wmxHHmHQ womHmN mmxmd ofiu .N0H>mmwm wmwhmmm sofiflz monom nosuo oxHHoD .ume can on pmonm wcficpau muomon mama‘somo so mwcfluwn Ham ouoamaoo .uoumH mHoQEoE QSOHw Ham zuHS coupsm SHHDM on HHHS mwcHumn ommfiH .AmvuopmmH pom MHom wGHpDHocN .mumnsme anouw Ham oumm .mHmom Sumo How mmcHumH onmeom Mo mwamu Hflmm osu mm: 90% MHwammmD umoE on HHH3 mwcwumu ommfiH .30: cu m: mCOHmmom moouw ofiu cHnuH3 N0H>m£mn HmUuUm .mnmnama Loom mo CowmmmumEH HmCOmpwm.NnoN.mucommudmu umon umsu mamom some mo wofimnuxm onu cmosumn Houuoa mnu oHoNHoco .owma mcHBOHHom map auHS wcwpumum .oEm: :30 Much mHoHHocm .poumHH coon m>m£ wwfimc .mumnfiofi anonw HHm umSu muoc mama ummH m.umeoonNaHE_mH£u co "mGOHuosuumoH «monouo cw H0H>m£mm mo uwcflumm m xfipcoaa< 54 Appendix C Ratings of all dyadic partners (X & Y) within each group on the Liked (9) versus Disliked (0) Scale Ratees or Person X o -H m w x c: r—1 «4 H m °H H .1: U) N :3 m H c: m o a) :3 m :H m m 2: 'p NJ to a. .4 22 c: PERSON Y R Mark 7 7 5 8 8 6 8 9 A John 7 7 6 '7 7 5 '7 8 T Leslie 6 6 7 9 9 6 9 9 E Suzie 3 2 8 8 9 4 5 9 R Paul 6 7 8 9 9 7 8 9 S Lisa 6 6 6 7 8 6 6 3 Marie 7 7 6 8 8 5 7 9 Dan 6 4 7 £3 9 5 6 $9 Descending Rank-Order of Dyadic Partners' 55 Appendix D (X & Y) Mutual Liking Ratings Within Group D Column 1 Column 2 Column 3 E X X Z 25 X Paul & Dan Suzi & Paul Leslie & Suzi 9 + 9 = 18 9 + 9 = 18 9 + 8 = 17 Mark & Dan Leslie & Paul Leslie & Marie 9 + 6 = 15 9 + 8 = 17 9 + 6 = 15 John 8 Paul Suzi & Dan Suzi & Marie 1 7 + 7 = 14 9 + 8 = 17 5 + 8 = 13 Mark & John Paul & Marie Leslie & Lisa 7 + 7 = 14 8 + 8 = 16 6 + 6 = 12 Mark & Paul Leslie & Dan Lisa & Marie ' 8 + 6 = 14 9 + 6 = 16 6 + 5 = 11 John & Dan Mark & Marie Suzi & Lisa 8 + 4 = 12 8 + 7 = 15 4 + 7 = 11 Marie & Dan 9 + 6 = 15 Paul & Lisa 7 + 8 = 15 Lisa & Dan 8 + 5 = 13 Mark & Lisa 6 + 6 = 12 John & Leslie 6 + 6 = 12 Mark & Suzi 8 + 3 = 11 John & Lisa 5 + 6 = 11 John & Marie 7 + 4 = 11 Mark & Leslie 5 + 5 = 10 John & Suzi 7 + 2 = 9 56 REFERENCES Aiello, J. R., and Jones, 5. E. Field study of proxemic behavior of young school children in three sub-cultural groups. Journal of Personality and Social Psychology, I97I, 12, 351-356. Argyle, M. Social Interaction. London: Metheun 8 Company, I969. Argyle, M., and Cook, L. Gaze and Mutual Gaze. Cambridge, Cambridge University Press, I976. Argyle, M., and Dean, J. Eye contact, distant, and affiliation. Sociometry, I965, 283 289-304. Argyle, M., Lalljee, and Cook, M. The effects of visibility on inter- action in a dyad. Human Relations, I968, 21) 3-I7. Asher, V., and Snortum, J. Eye contact in children as a function of age, sex, social and intellectual variables. Developmental Psychology, I97], 4) 479. Benjamin, L. S. Structural analysis of differentiation failure. Psychiatry, 1979,4_2, 2—22. Caproni, V., Levine, 0., O'Neal, E., McDonaki P., and Garwood, G. Seating position, instructor's eye contact availability, and student participation in a small seminar. Journal of Social Psychology, I977, 1922 3l5-3I6. Chaikin, A., Sigler, E., and Derlaga, V. Nonverbal mediators of teacher expectancy effects. Journal of Personality_and Social Psythology, I974, 29) l44-I49. Crowme , J. P., and Marlowe, J. The Approval Motive: Studies in Evaluative Dependence, N.Y.: Wiley,l964. Efran, J. S. Looking for approval: Effects on visual behavior of approbation from persons differing in importance. Journal of Personality and Social Psychology, I968, 19) (I) 21-25. Efran, J. S., and Broughton, A. Effect of expectancies for social approval on visual behavior. Journal of Personality and Social Psychology, I966, 43 (I): l03-I07. Egan, G. Interpersonal Living: A skills/contract approach to human relations training in groups. Brooks/Cole Publishing Co., California, I976. 57 Ellsworth, P. C., and Carlsmith, M. J. Effects of eye contact and verbal content on effective responses to a dyadic interaction. Journal of Personality and Social Psychology, I968, I0, (I) 15-20. ‘— EIIsworth, P. C., Carlsmith, J. M., and Henson, A. A stare as a stimulus to flight in human subjects: A series of field ex- periments. Journal of Personality and Social Psycholggy Ellsworth, P. C., and Langer, E. J. Staring and approach: an in- terpretation of the stare as a nonspecific activator. Journal of Personality and Social Psychology, I976, 21, (I) II7-I22. Ellsworth, P. C., and Ludwig, L. M. Visual behavior in social inter- action. The Journal of Communication. I972, 223 375-403. Exline, R. V. Explorations in process of person perception: Visual interaction in relationship to competition, sex, and need for affliliation. Journal of Personality I963, 2L, l-20. Exline, R., and Feh, B. Application of semiosis to the study of visual interaction. In a Siegman and S. Feldstein (Eds.), Nonverbal Behavior and Communication. New York: Wiley, I978, pp. II7-I57. Exline, R. V., and Messick, D. The effects of dependency and social reinforcement upon visual behavior during an interview. British Journal of Social and Clinical Psychology, I967, 63 256-266. Exline, R.V.Gray, D., and Schuette, 0. Visual behavior in a dyad as affected by interview content and sex respondents. Journal of Personality and Social Psychology, I965, 12 (3), 20l-209. Exline, R. V., Thibaut, J., Brannon, C" and Gumpert, P. Visual inter- action in relation to Machiallianism and an unethical act. American Psychologist, I96I, I6, 396. (Abstract) Exline, R. V., and Winters, L. C. Affective relations and mutual glances in dyads. In S. S. Tomkins and C. E. Izard (Eds.), Affect, Cognition, and Personality, N. Y.: Spring, I965. Freud, S. Group Psychology and the analysis of the ego. In The Stan- dard Edition of the Compjete Psychol_gical Works of Sigmud Freud, Vol. l8; London: Hogarth, I955. Friedman, N. The social nature ofypsychological Research: The Psycho- logical Experiment as a Social Interaction. New York, Basic Book, I967. 58 Fugita, S. S. Effects of anxiety and approval on visual interaction. Journal of Personality and Social Psychology, l974, 22) (4), 536-592. Goldings, H. J. On the avowal and projection of happiness. Journal of Personality, I954, 223 30-47. Gray, S. L. Eye contact as a function of sex, race, and interpersonal needs. Dissertation Abstracts. Sept. Vol. 32 (3-8), I842,I97I. Hall, E. The Hidden Dimension. N. Y.: Double Day, I966. Holmes, 0. 5. Dimensions of projections. Psychological Bulletin, 92, 248-268. Hurley, J. R. Toward seeing ourselves as others see us. In L. R. Wolberg, M. L. Aruson, and A. R. Wolberg (Eds.), Group Therapy I978, N.Y.: Stratton Medical Books, I978. Jones, R. and Cooper, J. Mediation of experimenter effects. Journal of Personality_and Social Psychology, I97I, 29, 70-74. Kendon, A. Some functions of gaze direction in social interaction. Acta Psychologica, I967, 26) 22-63. Kendon, A., and Cook, M. The consistency of gaze patterns in social interaction. British Journal of Psychology. I969, 693 48I-494. Kimble, C. E., and Olszewiski, D. A. Gaze and emotional expression: The effects of message positivity-negativity and emotional intensity. Journal of Research in Personality, I980 (March) 14) (I), 60-69. Kleinke, C., Desautels, M., and Knap, 8. Adult gaze and visual responses of pre-school children. Journal of Genetic Psychology, I977, I3I 32I-322. Kleinke, C. L., Meeker, F. B., and LaFong, C. Effect of gaze, touch and use of name on evaluation of ”engaged” couples. Journal of Research in Personality, I974, 2, 368-373. Kleinke, C. L., and Pohlen, P. D. Affective and emotional responses as a function of other person's gaze and co-operativeness in a two-person game. Journal of Personalityyand Social Psychology, 1971. 11. (3). 308-313. Kleinke, C. L., Staneski, R. A., and Berger, 0. Evaluation as a function of interviewer gaze, reinforcement of subject gaze and interviewer attractiveness. Journal of Personality and Social Psychology, l975,_3l, IIS-I22. 59 Langer, E. J., and Abelson, R. P. The semantic of asking a favor: How to succeed in getting help without really dying. Journal of Personality and Social Psychology, I972, 243 26-32. Libby, W. L. Eye contact and direction of looking at stable individual differences. Journal of Experimental Research in Personality) I970, 43 303-3l2. Libby, W. L., and Yaklevich, 0. Personality determinants of eye contact and direction of gaze aversion. Journal of Personality and Social Psychology, I973, 21, I97-206. Mehrabian, A. Orientation behavior and nonverbal attitude communication. Journal of Communication, I967, 17, 324-332. Mehrabian, A. A relation of attitude to seated posture, orientation, and distance. Journal of Personality and Social Psychology, 1968.19. (I). 26—30. (E Mehrabian, A. Inference of attitude from the posture, orientation and distance of a communicator. Journal of Consulting and Clinical Psychology, I968, 32, (3), 296:3081 (5) Mehrabian, A. Relation of attitude to seated posture, orientation and distance. Journal of Personality and Social Psychology, I968, 19, 26-30. m Minor, M. Experimenter-expectancy effect as a function of evaluation apprehension. Journal of Personality and Sosical Psychology, I970. I_5. 326-332. Mobbs, N.A. Eye contact and introversion-extraversion. Reading,B.A. Thesis, Cited by A. Kendon and M. Cook, The consistency of gaze pattern in social interaction. British Journal of ngchology, 969, 69) 48I-494. Nielsen, G. Studies in self-confrontation. Munkssaard, Denmark, I962. Otteson, J. P., and Otteson, C. R. Effect of teacher's gaze on children's story recall. Perceptual and Motor Skills, I979, 59) 35-52. Pellegrini, R. J., Hicks, R. A., and Gordon, L. The effect of an approval seeking induction on eye contact in dyad. British Journal of Social and Clinical Psychology, I970, 9, 373-374. Rosenfeld, H. M. Effect of an approval-seeking induction on interpersonal proximity. Psychological Reports, I965, 113 I20-l22. Rosenthal, R., and Jacobson, L. Pygmalion in the classroom. N.Y.: Holt, I968. 60 Rosenthal, R. Experimenter Effects in Behavioral Research. New York: Appleton-Century, I966. Rubin, Z. Measurement of romantic love. Journal of Personality and Social Psychology, I970, 16, (2), 265-273. Russo, H. Eye contact, interpersonal distance, and the equilibrium theory, Journal of Personality and Social Psychology, I975, 313 497-502. Scherwitz, L., and Helmreich, R. Interactive effects of eye contact and verbal content on interpersonal attraction in dyads. Journal of Personality and Social Psychology, I973, 22, 6-I4. Schutz, W. C. Fundamental interpersonal Relations Orientations (FIRO) measure. Holt, Rienhart and Winston, I958: Simmel, G. Sociology of the senses: Visual interaction In R. E. Park and E. W. Burger, (Ed.), Introduction to the Science of Sociology. Chicago: University of Chicago Press, I92l. Simmel, G. The sociology of the senses. In R. E. Parks and E. W. Burgess (Eds.), Introduction to the Science of Sociology. 6 impr. Chicago: I929. Slater, P. E. On social regression. American Sociological Review, I963, 2g. 339-364. Storm, M. D., and Thomas, G. E. Reactions to physical closeness, Journal of Personality and Social Psychology I977, 35) 4I2-4I8. Vine, I. Judgement of Direction of Gaze: An interpretation of discrepant results. British Journal of Social and Clinical Psychology, I97I, l_0_9 320-33] . Wardwell, E. Children's reactions to being watched during success and failure, (Doctoral dissertation, Cornell University, I960). Dissertation Abstracts International, I960, 29, 474I. Weisbrod, R. R. Looking behavior in a discussion group. Term paper sub- mitted to psychology 546 under the directRXIof Professor Hongabaught, Cornell University, Ithaca, N.Y. I965. Winer, B. J. Statistical principjes in experimental design. McGraw Hill Book Company, New York, 2nd Ed., I97l. 11c“; an TE IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIILIIIIIII“ 3 29 0612