- Email: [email protected]
Physiologic correlates of direct and averted gaze
BiologicalPsychology 0 North-Holland
10 (1980) 189-199 Publishing Company
PHYSIOLOGIC CORRELATES
OF DIRECT AND AVERTED GAZE *
Wilberta L. DONOVAN and Lewis A. LEAVITT Waisman Center on Mental Retardation and Human Development, University of Wisconsin, Madison, WI 53706, U.S.A. Accepted
for publication
1 July 1980
The present study was designed to explore the effects of reciprocated and unreciprocated gaze upon skin conductance and heart rate response. A paradigm used previously to investigate maternal response to infant gaze was selected which allowed us to consider possible order effects, sex differences and differential rates of habituation in response to direct and averted gaze. Twenty-four males and 24 females were shown 10 set silent images of an adult face on a videotape monitor of an adult face. In one condition, the en face position offered eye contact; in the other condition, eye contact was not possible. Each subject viewed a sequence of six identical episodes of one condition followed by six episodes of the other condition. Analysis of skin conductance response revealed a significant effect of stimulus sex @ < 0.05) with the male stimulus figures eliciting the greater skin conductance response. HR responses to the two conditions varied not only as a function of stimulus sex but also as a function of the viewer’s sex. The most distinctive data came from a male viewing another male. For this group alone, responding was consistent to unreciprocated gaze viewed first @ < 0.01) and viewed second tp < 0.05); and for this group alone responding to reciprocated gaze viewed first habituated tp < 0.05).
1. Introduction Gaze behavior has long been recognized as an important nonverbal process operative during social interaction. The pace of social interaction is regulated in part both by eye contact and gaze aversion. Gaze behavior is used to gather information about another’s reactions and in establishing and defining the particular nature of dyadic relations (Exline, Gray and Schuette, 1965). Gaze behavior appears to be especially important for male nonhuman primates where mutual gaze between males initiates mutual threat displays which reaffirm or reorder their places in the * This research was supported in part by funds from NICHD grant HD08240 awarded the second author and implementation grant HD03352 awarded the Waisman Center. The authors wish to thank C. Gillman and B. Orchard for computer assistance and B. Anderson and E. Harrington HI for assistance in the construction of stimuli. Preliminary results of this study were reported at the meetings of The Society for Psychophysiological Research, San Diego, 1976. Address requests for reprints to: L. Leavitt or W. Donovan, Waisman Center, University of Wisconsin, Madison, WI 53706, U.S.A. 189
190
W.L. Donovan and L.A. Leavitt / Physiologk correlates of direct and averted gaze
dominance hierarchy. It has been suggested that the averted gaze is a sign of submission and a means of avoiding confrontation (Hall and Devore, 1965). An additional interesting feature of gaze behavior is its relation to arousal. Several studies have noted differential effects of eye contact and unreciprocated gaze upon arousal. Frequency and amplitude of galvanic skin responses have been found to be greater when normal adult subjects’ gazes were reciprocated as compared to unreciprocated gaze (Nichols and Champness, 1971); EEG arousal was found to be higher for direct than for averted gaze (Gale, Spratt, Chapman and Smallbone, 1975; Gale, Kingsley, Brookes and Smith, 1978) and Kleinke and Pohlen, (1971) report that heart rate (HR) averaged over a 15 min period was significantly higher for subjects in the gaze condition when compared to subjects in the no-gaze condition. The present study was designed to explore both skin conductance (SC) and HR correlates of direct and averted gaze. Our study differs from those reported earlier in several important ways. First, we elected to manipulate direct and averted gaze by displaying the stimulus figures via a television monitor as opposed to other investigators’ use of a live confederate instructed to manipulate his gaze behavior. This allowed us to give detailed consideration to experimental procedure, including the regulation of parameters such as onset, intensity, duration of the stimuli, and consistency of facial expression. Comparable precision with a live confederate would be difficult to attain. Furthermore, since spontaneity, expected during social interaction, is lacking with the confederate, that situation may appear somewhat contrived and artificial to the viewer. Second, past investigations of physiologic response to gaze behavior focused primarily on males viewing another male, whereas our study employed stimulus figures and viewers of both sexes. Third, in the single study (Kleinke and Pohlen, 1971) exploring changes in HR as a function of gaze behavior, investigators looked at changes in tonic level over the 15 min experimental period. We were particularly interested in phasic changes in response to gaze behavior and used cardiac deceleration following stimulus onset as an index of the Orienting Response (OR) (Graham and Clifton, 1966) and a physiologic reflection of attentional processes (Lacey, 1967). Direction of HR response to nonsignal stimuli has been emphasised as differentiating the OR and the Defensive Response (DR) (Graham and Clifton, 1966). Since earlier research (Leavitt and Donovan, 1979) indicated that maternal cardiac acceleration to c,ontinued encounters with infant gaze could be predicted by scores on the Locus of Control Inventory (Rotter, 1966), it was of interest to determine if a similar relationship existed for adult pairs. In sum, the present investigation was designed to determine whether: (1) the analysis of phasic cardiac changes differentiated response to direct and averted gaze, (2) there are differences between direct and averted gaze in their ability to command continued attention from the viewer, (3) whether physiologic response to gaze behavior is a function of the sex of stimulus figure and/or sex of viewer, and (4) whether cardiac responses were correlated with perceived locus of control.
W.L. Donovan and LA.
Lea&t
/Physiologic
correlates of direct and averted gaze
191
2. Method
Forty-eight college students, 24 of each sex, participated in the experiment. Subjects were recruited from psychology classes and were given class credits for their participation.
The stimuli were black and white 10 set silent images of an adult face displayed on a 50.8 cm videotape monitor. In one condition, the en face position offered eye contact (Toward), in the other condition eye contact was not possible (Away). In the Toward condition the monitor displayed a full-face view of the head and shoulders of an adult. This stimulus figure which filled the monitor screen was against an undifferentiated light background. The stimulus was similar for the Away condition except the head was turned approximately 45” from midline during filming so that both eyes were visible, but eye contact was not possible. The figures had been filmed with the camera at eye level. In neither condition were the figures smiling. The construction of a stimulus tape involved the editing of a film so that a sequence of six identical episodes of one condition was followed by six episodes of the other condition. A blank screen followed each stimulus presentation. Four tapes were constructed. To avoid confounding any findings with any single adult image, two adults of each sex were filmed for each of the two conditions in each of the two possible orders. The four adults selected for filming were young adults and not known to any of the subjects. Subjects were seated approximateiy 1.22 m from the monitor. 2.3. Apparatus An Offner Type R Dynograph was used to record HR and SC. Cardiac activity was recorded via Beckman biopotential m~iature eIectrodes. Active leads were at the left ankle and right forearm. The ground electrode was at the right ankle. The signal was transduced and amplified by the Dynograph and via a Vetter FM Model-3 recorder adaptor was recorded on magnetic tape on a Teat 3300 S recorder. The R-R intervals of the electrocardiogram were timed to the nearest millisecond off line using a PDP-12 computer and then converted to HR in bpm for each I see interval. Analysis of second by second HR responses were made for the 10 set interval following stimulus onset. Skin conductance was monitored via Beckman biopotential skin electrodes (8 mm diameter) placed according to the unpolar method; the active electrode was placed on the palmar surface of the second phalanx of the middle finger of the right hand and the indifferent electrode over an abraded forearm site near the right elbow. K-Y sterile lubricant jelly was used and the electrodes
192
W.L. Donovan and L.A. Leavitt / Physiologk correlates of direct and averted gaze
were held in place with micropore surgical tape. A Beckman type 9842 GSR coupler with a time constant of 10 set applied a constant voltage of 0.5 V to the electrodes with the anode at the finger. The 1.27 cm memorex videotape containing the sequences stimuli was played on a Panasonic Model 3020 VTR and a Model AN-69V monitor displayed the stimuli. With the equipment in operation, the illuminance at the subject’s face was 117 cd/m2 as measured by a Tektronix J-16 Photometer using a Model 56511 illuminance probe. To ensure a standard background, white noise produced by a GrasonStadler Model 901 B Noise Generator was delivered via a 10.16 cm KLH model 708 speaker within the chamber. The sound level at the site of the subject’s head measured 32 dB (A) on a General Radio 1551C sound level meter. 2.4. Design Five factors were manipulated in a 2 X 2 X 2 X 2 X 6 factorial design. The first three factors, Stimulus Sex, Subject Sex and Order were, respectively, the sex of the person portrayed on the monitor, the sex of the viewer, and the order in which each condition was viewed. All were between-subject factors. The fourth and fifth factors, Condition and Trials were, respectively, the Toward and Away conditions and the six episodes of each. These were within-subject factors. A Seconds (second by second change) was an additional within-subject factor for analysis of HR data. The design was selected to allow consideration of possible order effects, sex differences and differential rates of habituation in response to direct and averted gaze. The subjects were randomly assigned to one of the eight between-subjects conditions involved (N= 6 for each group) with the restriction that the Toward condition was viewed first by half the subjects and that the sex of the stimulus figure was the same as that of the viewer for half the subjects. 2.5. Procedure Subjects were tested individually in a sound-attenuated chamber. They were instructed to watch the monitor displaying the stimuli throughout the session. A camera was placed directly above the stimulus display monitor and by means of another monitor outside the chamber the experimenter was able to monitor the subjects’ gaze behavior. Subjects were aware of the camera but were also assured that a videorecording was not being made. The purpose of the study was described as an investigation of responses to social stimuli, especially responses to other persons. No mention was made, however, of our specific interest in the effects of eye contact upon response. A woman served as experimenter and attached the electrodes. After attaching electrodes a minimum of 5 min elapsed before the session commenced. Subjects were requested to relax and sit quietly during this period and throughout the session. All tapes had been constructed so that a blank screen preceded the first stimulus presentation by 45 sec. The duration of each stimuluswas
W.L. Donovan and L.A. Leavitt /Physiologic
correlates of direct and averted gaze
193
10 sec. The intertrial interval, measured from stimulus offset was randomly varied between 25 and 45 set with a mean of 3.5 sec. Following the session and the removal of the electrodes, subjects left the soundchamber and were seated in the adjoining room. Here they completed the Locus of Control Inventory (Rotter, 1966).
3. Results 3.1. Electrodermal responses The magnitude of the skin conductance response to each stimulus was determined by computing the difference between SC level in pus2 at stimulus onset and the maximum SC reached during the 10 set interval following stimulus onset. An analysis of variance was performed on log transformed scores. Analysis of SC responses revealed a significant effect of stimulus sex with the male (M) stimulus figures eliciting greater SC responses than females (F), [F (1,40) = 4.73, p < 0.051. Means for each group were TM = 0.136, s.d. = 0.16, and XF = 0.077, s.d. = 0.12. The Condition effect was marginally significant, [F (1,40) = 3.36, O.lO>p > 0.051 with the Toward condition eliciting a greater response than the Away condition. The respective means were x = 0.116, s.d. = 0.156 and x = 0.095, s.d. = 0.134. A general decrement in responding was observed as indicated by a significant Trials effect, [F (5,200) = 24.6, p < O.OOl] , accounted for by the approximate 50% response decrement (similar for both conditions) between trials 1 and 2. The significant Condition X Order interaction [F (1,40) = 12.86, p < O.OOl] reflects the increased responsiveness to the Toward condition being viewed first, x= 0.148, s.d. = 0.167, as compared to being viewed second, x = 0.086, s.d. = 0.139. Response difference to the Away condition viewed first and second was negligible,~=0.105,s.d.=0.138andX=0.087,s.d.=0.131. Similarly, response differences across trials to the two conditions were more pronounced upon initial viewing of each, whereas differences between the conditions viewed second were negligible, Condition X Order X Trials interaction [F (5,140) = 2.69, p < 0.051. Mean responses on Trial 1 for the Toward and Away condition were x = 0.3017 and x = 0.2290, respectively, with the percentage drop from Trial 1 to 6 being 66% for the Toward condition, and 75% for the Away condition. Thus two factors contributed to the Condition effect: greater response to the Toward condition viewed first and less response decrement across trials for this condition viewed first. The sex of the stimulus figure influenced response as reflected by the Condition X Order X Stimulus Sex X Trials interaction [F (5,200) = 2.64, p < 0.051. Post hoc analyses were conducted to detect those contrasts that accounted for this interaction. Response to the Away condition viewed second was a function of the sex of
194
W.L. Donovan and L.A. Leavitt 1 P~~_vsiologiccorrelates of direct and averted gaze
the stimulus figure. In the analysis of these A2 responses, the effect of Stimulus Sex was marginally significant, [F (1,22) = 3.1, 0.10 > p > 0.051. Mean responses of subjects viewing the male and female were _? = 0.118, s.d. = 0.145; x = 0.055, s.d. = 0.105, respectively. Since the Condition X Order X Stimulus Sex X Trials interaction suggested that the two sexes elicited differential responding across trials as a function of gaze, further post hoc analyses analyzed responses of subjects viewing the male vs. the female stimulus figures to the Toward condition viewed first (Tl) vs. second (T2), and to the Away condition viewed first (Al) vs. second (A2). Only the male stimulus figure elicited continued responding during T2 and A2. This was reflected in the Tl and T2 comparison with the male stimulus figure, and the Al and A2 comparison with the male stimulus figure: both comparisons revealed significant Order X Trials interactions, [F, (5,100) = 3SO,p< 0.051, and [F (5,100) = 2.76,~ < 0.051, respectively. In contrast, response decrement across trials was similar for all conditions involving a female viewing a female figure. Thus it appears that our initial finding of a main effect for Stimulus Sex is, in part, because the male stimulus figures continue to elicit responding especially across the latter trials (T2 and A2 responses). 3.2. Cardiac responses Analysis of second by second HR responses was made for the 10 set interval following stimulus onset. Responses were determined by expressing the mean HR during each of these 10 periods of 1 set each following stimulus onset as a deviation from HR during the prestimulus 1 set period. An analysis of variance performed on these deviations scores revealed a significant main effect for Seconds (second by second change) [F (9,360) = 7.32, p < O.OOl]. * Response was one of deceleration across the 10 set period with a peak deceleration of -2.3 bpm occurring at poststimulus second 6. The significant Stimulus Sex X Trials X Seconds interaction [F (45,180O) = 1.73, p < 0.0251, reflected trial differences in this deceleratory response to the male and female figures. The initial deceleratory response to male figures habituated across trials. In contrast, response to the female figures on Trials 1 and 6 was of greater magnitude than on intermediate trials (e.g., Trial 2). Other significant higher order interactions, Subject Sex X Condition X Trials X Seconds [F (45,180O) = 1.98, p < 0.0251, and the Condition X Order X Stimulus Sex X Subject Sex X Seconds, [F (9,360) = 2.75,~ < O.OOS], suggested the need to compute separate ANOVAs and trend components for each of the 16 groups. For three of the four groups viewing the Toward condition first, cardiac deceleration was consistent and did not habituate; F view F (Seconds = 2.20,~ < 0.05), F view M (linear Seconds X cubic Trials = 9.22,~ < 0.05) M view F (linear Seconds = * Due to the likelihood of heterogeneity of convariance being present in the data when rapid sampling of physiologic responses is involved, the degrees of freedom (df) for each significance test were reduced by one-half (Wilson, 1974).
W.L. Donovan and L.A. Leavitt /Physiologic
correlates of direct and averted gaze
195
13.9 1, p < 0.05), for M view M, the initial deceleratory response habituated by Trial 6 (linear Trials = 7.49, p < 0.05). A comparison of responses of same-sex dyads showed that reciprocated gaze from a female elicited a deceleratory response in female subjects which did not habituate, whereas an initial deceleratory response of comparable magnitude in males to the male figure habituated. This difference in response to Tl was significant as reflected in the Groups (same-sex dyads) X linear Trials interaction, [F (1 JO) = 5.38, p < 0.05. Although M view M was the only group that habituated to Tl, it was also the only group displaying a consistent response to Al (linear Trials = 19.62, p < 0.01). The deceleratory response on Trial 1 of males to another’s averted gaze was of considerable magnitude, with deviations at poststimulus seconds 5,6 and 7 being as large as 9 bpm. In addition, response to T2 and A2 differed as a function of the viewer’s sex and the sex of stimulus figure. A2 response was consistent only for M view M (Secons = 2.66, p < 0.05). In contrast, response to T2 was consistent for all groups, M view M (quad Seconds = 10.07, p < 0.05), M view F (Seconds = 2.42, p < 0.05), F view M (linear Seconds 9.96,~ < 0.05) except F view F (n.s.). In summary, data from both physiologic measures suggest that information concerning sex of stimulus figure and sex-of-viewer/sex-of-figure match increases our ability to predict response to direct versus averted gaze behavior. Analysis of skin conductance response revealed a significant effect of Stimulus Sex with the male stimulus figures eliciting the greater skin conductance response. Male stimulus figures elicited greater SC responses than females primarily because the male figure continued to elicit responding across the latter trials. HR responses to the two conditions varied not only as a function of Stimulus Sex but also as a function of the viewer’s sex. Only for males viewing another male, was responding consistent to unreciprocated gaze viewed first and viewed second; and for this group alone, cardiac deceleration to reciprocated gaze viewed first habituated. 3.3. Questionnaire data Scores on Rotter’s Locus of Control Inventory were used to classify subjects as Externals or Internals. Mean score for females was 11 .l (range 2-20) and for males 9.5 (range 2-21). Overall, scores were not systematically related to HR response.
4. Discussion
Data from the present study extend earlier findings by suggesting that the variables of sex of stimulus figure and sex-of-viewer/sex-of-stimulus-figure match must be considered when predicting physiologic response to gaze behavior. Although sex differences were not significant in the Nichols and Champness study (1971), reliable sex differences were obtained in our study. Consistently, the male
196
W.L. Donovan and L.A. Leavitt /Physiologic
correlates of direct and averted gaze
stimulus figures elicited a greater response. The male figures differed from female figures especially in their ability to elicit continued responding across the latter trials. Unfortunately, Nichols and Champness do not report the number of persons serving as confederates (i.e., stimulus figures). If a single confederate was used, it is possible that the finding of no sex differences was a function of the single confederate’s particular characteristics. To guard against this possibility, two adults of each sex served as the stimulus figures in the present study. Also, in contrast somewhat with the Nichols and Champness (1971) report of increased SC during eye contact compared to averted gaze, our data indicate that increases in SC during eye contact were only marginally significant. This finding is perhaps due to the use of filmed subjects in our study. Kleinke and Pohlen (1971) measured males’ cardiac change in response to eye contact with another male. Looking at tonic changes in HR level as a function of gaze behavior, they reported that males in their direct gaze condition had significantly higher mean HR (78.5 bpm) than males in the averted condition. Unfortunately, a comparison between resting heart rate in males for the two gaze groups prior to testing was not reported. * For the comparison, we looked for possible changes in tonic level in our males engaged in eye contact with another male but did not find similar increases in tonic level following six trials of eye contact. Our primary focus, however, was on possible phasic changes in response to variations in gaze behavior. Phasic changes following stimulus onset are frequently used as a physiologic index of stimulus receptivity. Direction of HR change has been used to distinguish between processes which facilitate stimulus intake as opposed to those processes which reduce the effects of stimulation. Current theory suggests that a protective-defensive system is associated with cardiac acceleration (the cardiac component of the DR, Graham and Clifton, 1966) presumed to limit the effects of stimulation (Lacey, l967), whereas cardiac deceleration, a component of the OR, appears to be associated with an orienting-attentional system presumably operating to enhance stimulus intake. Although the analysis of cardiac response did not reveal a significant main effect for Conditions, the Condition X Trials X Subject Sex X Seconds interaction suggests that direct and averted gaze may differentially affect HR response. Results of the separate trend analyses revealed that seven of the possible eight Toward groups yielded significant effects, whereas only two Away groups exhibited response consistency. Perhaps the most interesting data come from those males who viewed another male. This group alone responded consistently to Al and A2. Apparently, a male’s averted gaze is potent in attracting attention, at least from another male. The sig-
* In the present study, we found a significant difference in mean resting HR prior to any testing (x = 87.67) and (x = 70.78) for the Toward and Away groups respectively. Thus following the viewing of six trials of Toward the mean HR was higher, x = 85.11, than for those viewing six trials of Away, x= 66.35. However, the difference between resting HR after eye contact, x= 85.11, was not significantly different from that group’s prestimulus resting HR, x= 87.67.
W.L.D5~5~~nand LA
Leavitt
~~~ysiolo~.c
correlates
of direct and averted gaze
197
nificant linear Trials reflects habituation to Al, whereas the signifcant seconds effect to A2 reflects consistent responding and suggests that engaging in mutual gaze with another male (as during Tl) enhances the potency of that stimulus figure in its ability to attract attention (reflected in A2 responses) after mutual gaze has been broken. Possibly, information acquired during mutual gaze triggers the male observer to continue processing information about the other male. Similarly, response during T2 was both large and persistent for males viewing another male. In contrast to its T2 response, this group showed habituation during Tl . Perhaps fam~iarization with the male during episodes not involving eye contact (i.e., Al) allows for initial processing of information which in turn enables the male viewer to prolong responding to (tolerate) eye contact during T2. Males viewing another male also differed from other subjects in their response to Tl . Only for these subjects did response habituated by Trial 6. Based on our earlier data (Leavitt and Donovan, 1979) we speculate that, unlike typical habituation effects with repeated stimulation, response to continued encounters with eye contact may shift from an initial OR to a DR under certain conditions. Our earlier research (Leavitt and Donovan, 1979) has shown that maternal personality variables (Locus of Control, Rotter, 1966) predicted direction of HR response to continued encounters with infant eye contact. For those mothers who perceived little or no contingency between outcomes and their behavior (labeled Externals), and initial deceleratory response during El was replaced by an acceleratory response on Trial 6 and was interpreted by us as reflecting aversion to continued mutual gaze (Argyle and Cook, 1976). In the present study, a similar acceleratory response to continued encounters with eye contact was looked for only for males viewing another male since responses of all other groups continued to be deceleratory. Interestingly enough, the single male labeled as an External (scoring above the median score for males) showed a similar acceleratory response on Trial 6. Eye contact with another male may become aversive for those males who do not feel in control of events. Of course, in a more natural setting, a male may avert his gaze as a means of avoiding eye contact and possible confrontation. This bit of data is intriguing in light of data from the animal literature suggesting that mutual gaze between male primates initiates mutual threat displays which reaffirm or reorder one’s place in the dominance hierarchy, and that averted gaze is a sign of submission and a means of avoiding confrontation (Hall and Devore, 1965). The difference in cardiac response of males viewing another male from the response of the other same-sex or mixed-sex dyads may be due in part to social pressures toward adopting sex appropriate behaviors involving eye contact. Since one of the functions of eye contact is the gathering of information; the observed sex difference may reflect differential emphasis upon dependency and autonomy behaviors for the two sexes in our culture. For example, the consistent finding that women engage in more mutual glances regardless of the partner’s sex’(Exline et al., 1965) may reflect the greater acceptance of this behavior in females, and the extent to which one engages in mutual and even one-way glances may be determined in
198
W.L. Donovan and L.A. Leavitt /Physiologic correlates of direct and averted gaze
part by the adherence to norms concerning the appropriateness of such behaviors. Also, Ellsworth and Ross (1975) report the perception of intimacy in response to direct and averted gaze to be a function of viewer’s sex. Specifically, direct gaze, whether constant or contingent upon partner’s response, appears to promote reports of intimacy between females and reticence between males, while gaze avoidance had the opposite effect. A final comment with regard to the data from earlier studies reporting a differential effect of direct and averted gaze upon arousal. Increases in SC (Nichols and Champness, 1971) and EEG-decreased alpha activity (Gale et al., 1975) were greatest under conditions of direct gaze. It may be noted that in these studies and in the present one, the stimulus figures were silent and trial length varied from 10 set (Nichols and Champness and the present study) to 18 set (Gale et al. study). Although our data do not lend strong support for differences in SC under the two conditions of gaze behavior, the cardiac data do suggest differential responding under direct and unreciprocated gaze. A duration of 18 set (and possibly even of 10 set) is uncommonly long for eye contact to be maintained in silence (see Exline, 1971). This may be a contributing factor accounting for increases in physiologic response to direct gaze. In summary, our data suggest that physiologic response to gaze behavior cannot best be predicted on the basis of a direct-averted gaze dichotomy. Rather the data indicate that variables such as sex of stimulus figure and sex-of-viewer/sex-of-figure match must be considered when assessing physiologic responses of subjects to gaze behavior. Especially interesting were the distinctive cardiac responses to those males viewing another male.
References Argyle, M. and Cook, M. (1976). Gaze and Mutual Gaze. Cambridge University Press: Cambridge. Ellsworth, P. and Ross, L. (1975). Intimacy in response to direct gaze. Journal of Experimental Social Psychology, 11, 592-613. Exline, R.V. (1971). Visual interaction: The glances of power and preference. In: Cole, J.K. (Ed.). Nebraska Symposium on Motivation, 163-206. Exline, R., Gray, D. and Schuette, D. (1965). Visual behavior in a dyad as affected by interview content and sex of respondent. Journal of Personality and Social Psychology, 1,201-209. Gale, A., Kingsley, E., Brookes, S. and Smith, D. (1978). Cortical arousal and social intimacy in the human female under different conditions of eye contact. Behavioral Processes, 3, 271 275. Gale, A., Spratt, G., Chapman, A.J. and Smallbone, A. (1975). EEG correlates of eye contact and interpersonal distance. Biological Psychology, 3, 237-245. Graham, I:. and Clifton, R. (1966). Heart-rate change as a component of the orienting response. Psychological Bulletin, 65, 305320. Hall, K.R.L. and Devore, I. (1965). Baboon social behavior. In: Devore, I. (I’d.). Primate Behavior. Holt Rinehart and Winston: New York.
W.L. Donovan and L.A. Leavitt /Physiologic
correlates of direct and averted gaze
199
Kleineke, C.L. and Pohlen, P.D. (1971). Affective and emotional responses as a function of other person’s gaze and cooperativeness in a two-person game. Journal of Personality and Social Psychology, 17, 3088313. Lacey, J. (1967). Somatic response patterning and stress: Some revisions of activation theory. In: Appley, M. and Trumbell, R. (Eds). Psychological Stress: Issues in Research. AppletonCrofts: New York, 14442. Leavitt, L.A. and Donovan, W.L. (1979). Perceived infant temperament, locus of control and maternal physiological response to infant gaze. Journal of Research In Personality, 13, 2677 278. Nichols, K.A. and Champness, B.G. (1971). Eye gaze and the GSR. Journal of Experimental Social Psychology, 7,623-626. Rotter, J.B. (1966). Generalized expectancies of internal versus external control of reinforcement. Psychological Monographs, 80, whole No. 609. Wilson, R.S. (1974). CARDIVAR: The statistical analysis of heart rate data. Psychophysiology, 11,76-85.
10 (1980) 189-199 Publishing Company
PHYSIOLOGIC CORRELATES
OF DIRECT AND AVERTED GAZE *
Wilberta L. DONOVAN and Lewis A. LEAVITT Waisman Center on Mental Retardation and Human Development, University of Wisconsin, Madison, WI 53706, U.S.A. Accepted
for publication
1 July 1980
The present study was designed to explore the effects of reciprocated and unreciprocated gaze upon skin conductance and heart rate response. A paradigm used previously to investigate maternal response to infant gaze was selected which allowed us to consider possible order effects, sex differences and differential rates of habituation in response to direct and averted gaze. Twenty-four males and 24 females were shown 10 set silent images of an adult face on a videotape monitor of an adult face. In one condition, the en face position offered eye contact; in the other condition, eye contact was not possible. Each subject viewed a sequence of six identical episodes of one condition followed by six episodes of the other condition. Analysis of skin conductance response revealed a significant effect of stimulus sex @ < 0.05) with the male stimulus figures eliciting the greater skin conductance response. HR responses to the two conditions varied not only as a function of stimulus sex but also as a function of the viewer’s sex. The most distinctive data came from a male viewing another male. For this group alone, responding was consistent to unreciprocated gaze viewed first @ < 0.01) and viewed second tp < 0.05); and for this group alone responding to reciprocated gaze viewed first habituated tp < 0.05).
1. Introduction Gaze behavior has long been recognized as an important nonverbal process operative during social interaction. The pace of social interaction is regulated in part both by eye contact and gaze aversion. Gaze behavior is used to gather information about another’s reactions and in establishing and defining the particular nature of dyadic relations (Exline, Gray and Schuette, 1965). Gaze behavior appears to be especially important for male nonhuman primates where mutual gaze between males initiates mutual threat displays which reaffirm or reorder their places in the * This research was supported in part by funds from NICHD grant HD08240 awarded the second author and implementation grant HD03352 awarded the Waisman Center. The authors wish to thank C. Gillman and B. Orchard for computer assistance and B. Anderson and E. Harrington HI for assistance in the construction of stimuli. Preliminary results of this study were reported at the meetings of The Society for Psychophysiological Research, San Diego, 1976. Address requests for reprints to: L. Leavitt or W. Donovan, Waisman Center, University of Wisconsin, Madison, WI 53706, U.S.A. 189
190
W.L. Donovan and L.A. Leavitt / Physiologk correlates of direct and averted gaze
dominance hierarchy. It has been suggested that the averted gaze is a sign of submission and a means of avoiding confrontation (Hall and Devore, 1965). An additional interesting feature of gaze behavior is its relation to arousal. Several studies have noted differential effects of eye contact and unreciprocated gaze upon arousal. Frequency and amplitude of galvanic skin responses have been found to be greater when normal adult subjects’ gazes were reciprocated as compared to unreciprocated gaze (Nichols and Champness, 1971); EEG arousal was found to be higher for direct than for averted gaze (Gale, Spratt, Chapman and Smallbone, 1975; Gale, Kingsley, Brookes and Smith, 1978) and Kleinke and Pohlen, (1971) report that heart rate (HR) averaged over a 15 min period was significantly higher for subjects in the gaze condition when compared to subjects in the no-gaze condition. The present study was designed to explore both skin conductance (SC) and HR correlates of direct and averted gaze. Our study differs from those reported earlier in several important ways. First, we elected to manipulate direct and averted gaze by displaying the stimulus figures via a television monitor as opposed to other investigators’ use of a live confederate instructed to manipulate his gaze behavior. This allowed us to give detailed consideration to experimental procedure, including the regulation of parameters such as onset, intensity, duration of the stimuli, and consistency of facial expression. Comparable precision with a live confederate would be difficult to attain. Furthermore, since spontaneity, expected during social interaction, is lacking with the confederate, that situation may appear somewhat contrived and artificial to the viewer. Second, past investigations of physiologic response to gaze behavior focused primarily on males viewing another male, whereas our study employed stimulus figures and viewers of both sexes. Third, in the single study (Kleinke and Pohlen, 1971) exploring changes in HR as a function of gaze behavior, investigators looked at changes in tonic level over the 15 min experimental period. We were particularly interested in phasic changes in response to gaze behavior and used cardiac deceleration following stimulus onset as an index of the Orienting Response (OR) (Graham and Clifton, 1966) and a physiologic reflection of attentional processes (Lacey, 1967). Direction of HR response to nonsignal stimuli has been emphasised as differentiating the OR and the Defensive Response (DR) (Graham and Clifton, 1966). Since earlier research (Leavitt and Donovan, 1979) indicated that maternal cardiac acceleration to c,ontinued encounters with infant gaze could be predicted by scores on the Locus of Control Inventory (Rotter, 1966), it was of interest to determine if a similar relationship existed for adult pairs. In sum, the present investigation was designed to determine whether: (1) the analysis of phasic cardiac changes differentiated response to direct and averted gaze, (2) there are differences between direct and averted gaze in their ability to command continued attention from the viewer, (3) whether physiologic response to gaze behavior is a function of the sex of stimulus figure and/or sex of viewer, and (4) whether cardiac responses were correlated with perceived locus of control.
W.L. Donovan and LA.
Lea&t
/Physiologic
correlates of direct and averted gaze
191
2. Method
Forty-eight college students, 24 of each sex, participated in the experiment. Subjects were recruited from psychology classes and were given class credits for their participation.
The stimuli were black and white 10 set silent images of an adult face displayed on a 50.8 cm videotape monitor. In one condition, the en face position offered eye contact (Toward), in the other condition eye contact was not possible (Away). In the Toward condition the monitor displayed a full-face view of the head and shoulders of an adult. This stimulus figure which filled the monitor screen was against an undifferentiated light background. The stimulus was similar for the Away condition except the head was turned approximately 45” from midline during filming so that both eyes were visible, but eye contact was not possible. The figures had been filmed with the camera at eye level. In neither condition were the figures smiling. The construction of a stimulus tape involved the editing of a film so that a sequence of six identical episodes of one condition was followed by six episodes of the other condition. A blank screen followed each stimulus presentation. Four tapes were constructed. To avoid confounding any findings with any single adult image, two adults of each sex were filmed for each of the two conditions in each of the two possible orders. The four adults selected for filming were young adults and not known to any of the subjects. Subjects were seated approximateiy 1.22 m from the monitor. 2.3. Apparatus An Offner Type R Dynograph was used to record HR and SC. Cardiac activity was recorded via Beckman biopotential m~iature eIectrodes. Active leads were at the left ankle and right forearm. The ground electrode was at the right ankle. The signal was transduced and amplified by the Dynograph and via a Vetter FM Model-3 recorder adaptor was recorded on magnetic tape on a Teat 3300 S recorder. The R-R intervals of the electrocardiogram were timed to the nearest millisecond off line using a PDP-12 computer and then converted to HR in bpm for each I see interval. Analysis of second by second HR responses were made for the 10 set interval following stimulus onset. Skin conductance was monitored via Beckman biopotential skin electrodes (8 mm diameter) placed according to the unpolar method; the active electrode was placed on the palmar surface of the second phalanx of the middle finger of the right hand and the indifferent electrode over an abraded forearm site near the right elbow. K-Y sterile lubricant jelly was used and the electrodes
192
W.L. Donovan and L.A. Leavitt / Physiologk correlates of direct and averted gaze
were held in place with micropore surgical tape. A Beckman type 9842 GSR coupler with a time constant of 10 set applied a constant voltage of 0.5 V to the electrodes with the anode at the finger. The 1.27 cm memorex videotape containing the sequences stimuli was played on a Panasonic Model 3020 VTR and a Model AN-69V monitor displayed the stimuli. With the equipment in operation, the illuminance at the subject’s face was 117 cd/m2 as measured by a Tektronix J-16 Photometer using a Model 56511 illuminance probe. To ensure a standard background, white noise produced by a GrasonStadler Model 901 B Noise Generator was delivered via a 10.16 cm KLH model 708 speaker within the chamber. The sound level at the site of the subject’s head measured 32 dB (A) on a General Radio 1551C sound level meter. 2.4. Design Five factors were manipulated in a 2 X 2 X 2 X 2 X 6 factorial design. The first three factors, Stimulus Sex, Subject Sex and Order were, respectively, the sex of the person portrayed on the monitor, the sex of the viewer, and the order in which each condition was viewed. All were between-subject factors. The fourth and fifth factors, Condition and Trials were, respectively, the Toward and Away conditions and the six episodes of each. These were within-subject factors. A Seconds (second by second change) was an additional within-subject factor for analysis of HR data. The design was selected to allow consideration of possible order effects, sex differences and differential rates of habituation in response to direct and averted gaze. The subjects were randomly assigned to one of the eight between-subjects conditions involved (N= 6 for each group) with the restriction that the Toward condition was viewed first by half the subjects and that the sex of the stimulus figure was the same as that of the viewer for half the subjects. 2.5. Procedure Subjects were tested individually in a sound-attenuated chamber. They were instructed to watch the monitor displaying the stimuli throughout the session. A camera was placed directly above the stimulus display monitor and by means of another monitor outside the chamber the experimenter was able to monitor the subjects’ gaze behavior. Subjects were aware of the camera but were also assured that a videorecording was not being made. The purpose of the study was described as an investigation of responses to social stimuli, especially responses to other persons. No mention was made, however, of our specific interest in the effects of eye contact upon response. A woman served as experimenter and attached the electrodes. After attaching electrodes a minimum of 5 min elapsed before the session commenced. Subjects were requested to relax and sit quietly during this period and throughout the session. All tapes had been constructed so that a blank screen preceded the first stimulus presentation by 45 sec. The duration of each stimuluswas
W.L. Donovan and L.A. Leavitt /Physiologic
correlates of direct and averted gaze
193
10 sec. The intertrial interval, measured from stimulus offset was randomly varied between 25 and 45 set with a mean of 3.5 sec. Following the session and the removal of the electrodes, subjects left the soundchamber and were seated in the adjoining room. Here they completed the Locus of Control Inventory (Rotter, 1966).
3. Results 3.1. Electrodermal responses The magnitude of the skin conductance response to each stimulus was determined by computing the difference between SC level in pus2 at stimulus onset and the maximum SC reached during the 10 set interval following stimulus onset. An analysis of variance was performed on log transformed scores. Analysis of SC responses revealed a significant effect of stimulus sex with the male (M) stimulus figures eliciting greater SC responses than females (F), [F (1,40) = 4.73, p < 0.051. Means for each group were TM = 0.136, s.d. = 0.16, and XF = 0.077, s.d. = 0.12. The Condition effect was marginally significant, [F (1,40) = 3.36, O.lO>p > 0.051 with the Toward condition eliciting a greater response than the Away condition. The respective means were x = 0.116, s.d. = 0.156 and x = 0.095, s.d. = 0.134. A general decrement in responding was observed as indicated by a significant Trials effect, [F (5,200) = 24.6, p < O.OOl] , accounted for by the approximate 50% response decrement (similar for both conditions) between trials 1 and 2. The significant Condition X Order interaction [F (1,40) = 12.86, p < O.OOl] reflects the increased responsiveness to the Toward condition being viewed first, x= 0.148, s.d. = 0.167, as compared to being viewed second, x = 0.086, s.d. = 0.139. Response difference to the Away condition viewed first and second was negligible,~=0.105,s.d.=0.138andX=0.087,s.d.=0.131. Similarly, response differences across trials to the two conditions were more pronounced upon initial viewing of each, whereas differences between the conditions viewed second were negligible, Condition X Order X Trials interaction [F (5,140) = 2.69, p < 0.051. Mean responses on Trial 1 for the Toward and Away condition were x = 0.3017 and x = 0.2290, respectively, with the percentage drop from Trial 1 to 6 being 66% for the Toward condition, and 75% for the Away condition. Thus two factors contributed to the Condition effect: greater response to the Toward condition viewed first and less response decrement across trials for this condition viewed first. The sex of the stimulus figure influenced response as reflected by the Condition X Order X Stimulus Sex X Trials interaction [F (5,200) = 2.64, p < 0.051. Post hoc analyses were conducted to detect those contrasts that accounted for this interaction. Response to the Away condition viewed second was a function of the sex of
194
W.L. Donovan and L.A. Leavitt 1 P~~_vsiologiccorrelates of direct and averted gaze
the stimulus figure. In the analysis of these A2 responses, the effect of Stimulus Sex was marginally significant, [F (1,22) = 3.1, 0.10 > p > 0.051. Mean responses of subjects viewing the male and female were _? = 0.118, s.d. = 0.145; x = 0.055, s.d. = 0.105, respectively. Since the Condition X Order X Stimulus Sex X Trials interaction suggested that the two sexes elicited differential responding across trials as a function of gaze, further post hoc analyses analyzed responses of subjects viewing the male vs. the female stimulus figures to the Toward condition viewed first (Tl) vs. second (T2), and to the Away condition viewed first (Al) vs. second (A2). Only the male stimulus figure elicited continued responding during T2 and A2. This was reflected in the Tl and T2 comparison with the male stimulus figure, and the Al and A2 comparison with the male stimulus figure: both comparisons revealed significant Order X Trials interactions, [F, (5,100) = 3SO,p< 0.051, and [F (5,100) = 2.76,~ < 0.051, respectively. In contrast, response decrement across trials was similar for all conditions involving a female viewing a female figure. Thus it appears that our initial finding of a main effect for Stimulus Sex is, in part, because the male stimulus figures continue to elicit responding especially across the latter trials (T2 and A2 responses). 3.2. Cardiac responses Analysis of second by second HR responses was made for the 10 set interval following stimulus onset. Responses were determined by expressing the mean HR during each of these 10 periods of 1 set each following stimulus onset as a deviation from HR during the prestimulus 1 set period. An analysis of variance performed on these deviations scores revealed a significant main effect for Seconds (second by second change) [F (9,360) = 7.32, p < O.OOl]. * Response was one of deceleration across the 10 set period with a peak deceleration of -2.3 bpm occurring at poststimulus second 6. The significant Stimulus Sex X Trials X Seconds interaction [F (45,180O) = 1.73, p < 0.0251, reflected trial differences in this deceleratory response to the male and female figures. The initial deceleratory response to male figures habituated across trials. In contrast, response to the female figures on Trials 1 and 6 was of greater magnitude than on intermediate trials (e.g., Trial 2). Other significant higher order interactions, Subject Sex X Condition X Trials X Seconds [F (45,180O) = 1.98, p < 0.0251, and the Condition X Order X Stimulus Sex X Subject Sex X Seconds, [F (9,360) = 2.75,~ < O.OOS], suggested the need to compute separate ANOVAs and trend components for each of the 16 groups. For three of the four groups viewing the Toward condition first, cardiac deceleration was consistent and did not habituate; F view F (Seconds = 2.20,~ < 0.05), F view M (linear Seconds X cubic Trials = 9.22,~ < 0.05) M view F (linear Seconds = * Due to the likelihood of heterogeneity of convariance being present in the data when rapid sampling of physiologic responses is involved, the degrees of freedom (df) for each significance test were reduced by one-half (Wilson, 1974).
W.L. Donovan and L.A. Leavitt /Physiologic
correlates of direct and averted gaze
195
13.9 1, p < 0.05), for M view M, the initial deceleratory response habituated by Trial 6 (linear Trials = 7.49, p < 0.05). A comparison of responses of same-sex dyads showed that reciprocated gaze from a female elicited a deceleratory response in female subjects which did not habituate, whereas an initial deceleratory response of comparable magnitude in males to the male figure habituated. This difference in response to Tl was significant as reflected in the Groups (same-sex dyads) X linear Trials interaction, [F (1 JO) = 5.38, p < 0.05. Although M view M was the only group that habituated to Tl, it was also the only group displaying a consistent response to Al (linear Trials = 19.62, p < 0.01). The deceleratory response on Trial 1 of males to another’s averted gaze was of considerable magnitude, with deviations at poststimulus seconds 5,6 and 7 being as large as 9 bpm. In addition, response to T2 and A2 differed as a function of the viewer’s sex and the sex of stimulus figure. A2 response was consistent only for M view M (Secons = 2.66, p < 0.05). In contrast, response to T2 was consistent for all groups, M view M (quad Seconds = 10.07, p < 0.05), M view F (Seconds = 2.42, p < 0.05), F view M (linear Seconds 9.96,~ < 0.05) except F view F (n.s.). In summary, data from both physiologic measures suggest that information concerning sex of stimulus figure and sex-of-viewer/sex-of-figure match increases our ability to predict response to direct versus averted gaze behavior. Analysis of skin conductance response revealed a significant effect of Stimulus Sex with the male stimulus figures eliciting the greater skin conductance response. Male stimulus figures elicited greater SC responses than females primarily because the male figure continued to elicit responding across the latter trials. HR responses to the two conditions varied not only as a function of Stimulus Sex but also as a function of the viewer’s sex. Only for males viewing another male, was responding consistent to unreciprocated gaze viewed first and viewed second; and for this group alone, cardiac deceleration to reciprocated gaze viewed first habituated. 3.3. Questionnaire data Scores on Rotter’s Locus of Control Inventory were used to classify subjects as Externals or Internals. Mean score for females was 11 .l (range 2-20) and for males 9.5 (range 2-21). Overall, scores were not systematically related to HR response.
4. Discussion
Data from the present study extend earlier findings by suggesting that the variables of sex of stimulus figure and sex-of-viewer/sex-of-stimulus-figure match must be considered when predicting physiologic response to gaze behavior. Although sex differences were not significant in the Nichols and Champness study (1971), reliable sex differences were obtained in our study. Consistently, the male
196
W.L. Donovan and L.A. Leavitt /Physiologic
correlates of direct and averted gaze
stimulus figures elicited a greater response. The male figures differed from female figures especially in their ability to elicit continued responding across the latter trials. Unfortunately, Nichols and Champness do not report the number of persons serving as confederates (i.e., stimulus figures). If a single confederate was used, it is possible that the finding of no sex differences was a function of the single confederate’s particular characteristics. To guard against this possibility, two adults of each sex served as the stimulus figures in the present study. Also, in contrast somewhat with the Nichols and Champness (1971) report of increased SC during eye contact compared to averted gaze, our data indicate that increases in SC during eye contact were only marginally significant. This finding is perhaps due to the use of filmed subjects in our study. Kleinke and Pohlen (1971) measured males’ cardiac change in response to eye contact with another male. Looking at tonic changes in HR level as a function of gaze behavior, they reported that males in their direct gaze condition had significantly higher mean HR (78.5 bpm) than males in the averted condition. Unfortunately, a comparison between resting heart rate in males for the two gaze groups prior to testing was not reported. * For the comparison, we looked for possible changes in tonic level in our males engaged in eye contact with another male but did not find similar increases in tonic level following six trials of eye contact. Our primary focus, however, was on possible phasic changes in response to variations in gaze behavior. Phasic changes following stimulus onset are frequently used as a physiologic index of stimulus receptivity. Direction of HR change has been used to distinguish between processes which facilitate stimulus intake as opposed to those processes which reduce the effects of stimulation. Current theory suggests that a protective-defensive system is associated with cardiac acceleration (the cardiac component of the DR, Graham and Clifton, 1966) presumed to limit the effects of stimulation (Lacey, l967), whereas cardiac deceleration, a component of the OR, appears to be associated with an orienting-attentional system presumably operating to enhance stimulus intake. Although the analysis of cardiac response did not reveal a significant main effect for Conditions, the Condition X Trials X Subject Sex X Seconds interaction suggests that direct and averted gaze may differentially affect HR response. Results of the separate trend analyses revealed that seven of the possible eight Toward groups yielded significant effects, whereas only two Away groups exhibited response consistency. Perhaps the most interesting data come from those males who viewed another male. This group alone responded consistently to Al and A2. Apparently, a male’s averted gaze is potent in attracting attention, at least from another male. The sig-
* In the present study, we found a significant difference in mean resting HR prior to any testing (x = 87.67) and (x = 70.78) for the Toward and Away groups respectively. Thus following the viewing of six trials of Toward the mean HR was higher, x = 85.11, than for those viewing six trials of Away, x= 66.35. However, the difference between resting HR after eye contact, x= 85.11, was not significantly different from that group’s prestimulus resting HR, x= 87.67.
W.L.D5~5~~nand LA
Leavitt
~~~ysiolo~.c
correlates
of direct and averted gaze
197
nificant linear Trials reflects habituation to Al, whereas the signifcant seconds effect to A2 reflects consistent responding and suggests that engaging in mutual gaze with another male (as during Tl) enhances the potency of that stimulus figure in its ability to attract attention (reflected in A2 responses) after mutual gaze has been broken. Possibly, information acquired during mutual gaze triggers the male observer to continue processing information about the other male. Similarly, response during T2 was both large and persistent for males viewing another male. In contrast to its T2 response, this group showed habituation during Tl . Perhaps fam~iarization with the male during episodes not involving eye contact (i.e., Al) allows for initial processing of information which in turn enables the male viewer to prolong responding to (tolerate) eye contact during T2. Males viewing another male also differed from other subjects in their response to Tl . Only for these subjects did response habituated by Trial 6. Based on our earlier data (Leavitt and Donovan, 1979) we speculate that, unlike typical habituation effects with repeated stimulation, response to continued encounters with eye contact may shift from an initial OR to a DR under certain conditions. Our earlier research (Leavitt and Donovan, 1979) has shown that maternal personality variables (Locus of Control, Rotter, 1966) predicted direction of HR response to continued encounters with infant eye contact. For those mothers who perceived little or no contingency between outcomes and their behavior (labeled Externals), and initial deceleratory response during El was replaced by an acceleratory response on Trial 6 and was interpreted by us as reflecting aversion to continued mutual gaze (Argyle and Cook, 1976). In the present study, a similar acceleratory response to continued encounters with eye contact was looked for only for males viewing another male since responses of all other groups continued to be deceleratory. Interestingly enough, the single male labeled as an External (scoring above the median score for males) showed a similar acceleratory response on Trial 6. Eye contact with another male may become aversive for those males who do not feel in control of events. Of course, in a more natural setting, a male may avert his gaze as a means of avoiding eye contact and possible confrontation. This bit of data is intriguing in light of data from the animal literature suggesting that mutual gaze between male primates initiates mutual threat displays which reaffirm or reorder one’s place in the dominance hierarchy, and that averted gaze is a sign of submission and a means of avoiding confrontation (Hall and Devore, 1965). The difference in cardiac response of males viewing another male from the response of the other same-sex or mixed-sex dyads may be due in part to social pressures toward adopting sex appropriate behaviors involving eye contact. Since one of the functions of eye contact is the gathering of information; the observed sex difference may reflect differential emphasis upon dependency and autonomy behaviors for the two sexes in our culture. For example, the consistent finding that women engage in more mutual glances regardless of the partner’s sex’(Exline et al., 1965) may reflect the greater acceptance of this behavior in females, and the extent to which one engages in mutual and even one-way glances may be determined in
198
W.L. Donovan and L.A. Leavitt /Physiologic correlates of direct and averted gaze
part by the adherence to norms concerning the appropriateness of such behaviors. Also, Ellsworth and Ross (1975) report the perception of intimacy in response to direct and averted gaze to be a function of viewer’s sex. Specifically, direct gaze, whether constant or contingent upon partner’s response, appears to promote reports of intimacy between females and reticence between males, while gaze avoidance had the opposite effect. A final comment with regard to the data from earlier studies reporting a differential effect of direct and averted gaze upon arousal. Increases in SC (Nichols and Champness, 1971) and EEG-decreased alpha activity (Gale et al., 1975) were greatest under conditions of direct gaze. It may be noted that in these studies and in the present one, the stimulus figures were silent and trial length varied from 10 set (Nichols and Champness and the present study) to 18 set (Gale et al. study). Although our data do not lend strong support for differences in SC under the two conditions of gaze behavior, the cardiac data do suggest differential responding under direct and unreciprocated gaze. A duration of 18 set (and possibly even of 10 set) is uncommonly long for eye contact to be maintained in silence (see Exline, 1971). This may be a contributing factor accounting for increases in physiologic response to direct gaze. In summary, our data suggest that physiologic response to gaze behavior cannot best be predicted on the basis of a direct-averted gaze dichotomy. Rather the data indicate that variables such as sex of stimulus figure and sex-of-viewer/sex-of-figure match must be considered when assessing physiologic responses of subjects to gaze behavior. Especially interesting were the distinctive cardiac responses to those males viewing another male.
References Argyle, M. and Cook, M. (1976). Gaze and Mutual Gaze. Cambridge University Press: Cambridge. Ellsworth, P. and Ross, L. (1975). Intimacy in response to direct gaze. Journal of Experimental Social Psychology, 11, 592-613. Exline, R.V. (1971). Visual interaction: The glances of power and preference. In: Cole, J.K. (Ed.). Nebraska Symposium on Motivation, 163-206. Exline, R., Gray, D. and Schuette, D. (1965). Visual behavior in a dyad as affected by interview content and sex of respondent. Journal of Personality and Social Psychology, 1,201-209. Gale, A., Kingsley, E., Brookes, S. and Smith, D. (1978). Cortical arousal and social intimacy in the human female under different conditions of eye contact. Behavioral Processes, 3, 271 275. Gale, A., Spratt, G., Chapman, A.J. and Smallbone, A. (1975). EEG correlates of eye contact and interpersonal distance. Biological Psychology, 3, 237-245. Graham, I:. and Clifton, R. (1966). Heart-rate change as a component of the orienting response. Psychological Bulletin, 65, 305320. Hall, K.R.L. and Devore, I. (1965). Baboon social behavior. In: Devore, I. (I’d.). Primate Behavior. Holt Rinehart and Winston: New York.
W.L. Donovan and L.A. Leavitt /Physiologic
correlates of direct and averted gaze
199
Kleineke, C.L. and Pohlen, P.D. (1971). Affective and emotional responses as a function of other person’s gaze and cooperativeness in a two-person game. Journal of Personality and Social Psychology, 17, 3088313. Lacey, J. (1967). Somatic response patterning and stress: Some revisions of activation theory. In: Appley, M. and Trumbell, R. (Eds). Psychological Stress: Issues in Research. AppletonCrofts: New York, 14442. Leavitt, L.A. and Donovan, W.L. (1979). Perceived infant temperament, locus of control and maternal physiological response to infant gaze. Journal of Research In Personality, 13, 2677 278. Nichols, K.A. and Champness, B.G. (1971). Eye gaze and the GSR. Journal of Experimental Social Psychology, 7,623-626. Rotter, J.B. (1966). Generalized expectancies of internal versus external control of reinforcement. Psychological Monographs, 80, whole No. 609. Wilson, R.S. (1974). CARDIVAR: The statistical analysis of heart rate data. Psychophysiology, 11,76-85.