Emotion 2003, Vol. 3, No. 3, 303–309
Copyright 2003 by the American Psychological Association, Inc. 1528-3542/03/$12.00 DOI: 10.1037/1528-3542.3.3.303
BRIEF REPORT
Happy Affect and Smiling: Is Their Relation Moderated by Interpersonal Power? Judith A. Hall and Terrence G. Horgan Northeastern University Three studies examined the hypothesis that the relation between self-rated happy affect and overall smiling during dyadic interactions is moderated by one’s relative power. In all 3 studies, interpersonal power was experimentally manipulated, and smiling and self-reported happy affect were measured during and after the interaction, respectively. Happy affect was positively correlated with overall smiling for both low-power and high-power participants. Moreover, the correlations for both low- and high-power participants were comparable to those found when power roles were not unequal or made salient, based on analysis of the present studies as well as comparison with the previously published literature. Happy affect and overall smiling had a positive relation that was not moderated by interpersonal power.
The belief that a smiling person is experiencing happiness is well documented around the world (Biehl et al., 1997; Fridlund, Ekman, & Oster, 1987) and is held by very young children as well as adults (Markham & Adams, 1992). Smiling is not only believed by observers to reflect happiness, research also indicates that smiling is actually associated with happiness and related positive affects. Experimental studies show that people smile more when exposed to positive than to negative stimuli (e.g., Ekman, Friesen, & Ancoli, 1980) and when exposed to stronger compared with weaker positive stimuli (e.g., Hess,
Banse, & Kappas, 1995; Jakobs, Manstead, & Fischer, 1999a). People’s self-ratings of happy affect have also been investigated in relation to smiling. In studies that do not consider social power as a variable, overall smiling shows correlations with self-reported happy affect that range from weakly to substantially positive (Bonnano & Keltner, 1997; Ekman, Davidson, & Friesen, 1990; Ekman et al., 1980; equal-power group in Hecht & LaFrance, 1998; Jakobs et al., 1999a; Jakobs, Manstead, & Fischer, 1999b). Duchenne smiles typically show the strongest correlations with self-rated happy affect (e.g., Ekman et al., 1990; Ruch, 1993; Scherer & Ceschi, 2000), consistent with findings showing that Duchenne smiles are more strongly correlated with enjoyment than are other smiles. Type of smile is, therefore, likely to be a moderator of the relation between smiling and self-rated happy affect. Hecht and LaFrance (1998) suggested that interpersonal power is also a moderator of this relation, arguing that high- and low-power people follow different smiling rules: High-power people have the “license” to smile at low-power people when feeling happy, whereas low-power people have an “obligation” to smile at high-power people no matter how they feel. Because of these rules, the relation between
Judith A. Hall and Terrence G. Horgan, Department of Psychology, Northeastern University. Terrence G. Horgan is now at the Department of Psychology, The Ohio State University. This research was made possible by grants from the National Science Foundation awarded to Judith A. Hall. We thank Marianne Schmid Mast and Nora A. Murphy for their highly valuable advice. Correspondence concerning this article should be addressed to Judith A. Hall, Department of Psychology, Northeastern University, 125 NI, Boston, Massachusetts 02115. E-mail:
[email protected]
303
304
BRIEF REPORT
smiling and self-rated happy affect should be larger in magnitude for high- than for low-power individuals.1 Hecht and LaFrance’s (1998) argument was based on their study in which college students interacted in the roles of interviewer (high power) and interviewee (low power), or else in equal-power roles. A positive correlation was not found between self-rated positive affect and overall smiling among those in the lowpower condition, r(22) ⳱ −.16, ns; but in both the high-power and equal-power conditions there was a positive correlation, r(22) ⳱ .55, p < .01; and r(22) ⳱ .38, p < .10, respectively. The correlations for the low- and high-power groups were significantly different from each other.2 Hecht and LaFrance’s (1998) finding that interpersonal power moderated the happy affect–smiling relation was based on a single, rather small study and is therefore well deserving of replication. Replication is especially called for because the theoretical foundation for their interpretation is debatable. When considering role constraints that may impact emotional expression, Hecht and LaFrance (1998) focused on socially weak people who may feel that they cannot smile when they are happy or must smile when they do not feel happy. However, it is equally easy to think of situations in which the same constraints apply to socially powerful people. As examples, it seems likely that police officers often feel constrained to hide feelings of cheerfulness or amusement when dealing with those over whom they have power, and, conversely, parents often feel constrained to smile at their children when their own moods would dictate the opposite. Therefore, the intuitive logic underlying Hecht and LaFrance’s (1998) interpretation is questionable.3 A second concern involves not a priori theorizing but rather how Hecht and LaFrance (1998) interpreted their correlations. They interpreted the near-zero correlation for low-power participants as indicating an obligation to smile and the positive correlation for high-power participants as indicating a license to smile. However, one could argue that those interpretations should be switched. The concept of obligation may be better captured in a positive correlation than in no correlation, if the positive correlation means that one feels required to show one’s affective state (cf. Rosenthal & DePaulo, 1979). In contrast, no correlation between happy affect and smiling, as Hecht and LaFrance (1998) found in low-power participants, suggests that one is free to smile when one pleases (i.e., to convey other messages through the smile besides happiness). However, Hecht and LaFrance said
that this kind of license was the prerogative of the more powerful. The present article describes three studies that allowed interpersonal power or status to be examined as a moderator of the relation between happy affect and smiling.4 In all three studies, participants were made to have either more or less power in relation to an interaction partner by being assigned to be the owner of an art gallery or the owner’s actual or potential assistant. Participants interacted in these roles while being videotaped. In two of the studies, we also included a condition in which the two interactants had equal power. In all three of the studies, participants rated how they had felt during the interaction, and
1 Hecht and LaFrance (1998) described their results as showing a moderating effect of positive affect on the relation between power and smiling (pp. 1338, 1340). However, this is a misstatement, as their analysis actually examined the moderating effect of power on the relation between positive affect and smiling, and their interpretations uniformly reflected this approach. Therefore, this is the framework within which we discuss their data. 2 Hecht and LaFrance (1998) acknowledged the similarity between the correlations in the equal- and high-power conditions but proceeded to develop their theorizing only in terms of the privileges of holding more power than a partner. Though they can be challenged on this choice of emphasis, it is probably moot to do so considering the results of the present research. 3 Note that neither the Hecht and LaFrance (1998) findings nor those reported in the present article speak to the question of whether people high and low in power, status, or dominance differ in their overall amounts of smiling. This question has a research literature of its own, which shows essentially no overall difference (e.g., Friedman & MillerHerringer, 1991; Gifford, 1994; Hall, Carter, Jimenez, Frost, & Smith LeBeau, 2002; Hall & Friedman, 1999; Hall, Horgan, & Carter, 2002; Hall, Smith LeBeau, Gordon Reinoso, & Thayer, 2001; Hecht & LaFrance, 1998 [for either Duchenne or non-Duchenne smiles]; Johnson, 1994). Deutsch (1990) found that low-power individuals in an experiment smiled more than high-power individuals. However, in that study, participants in the low-power condition were instructed to make a favorable impression, whereas those in the high-power condition were not, thereby confounding power role with the motive to ingratiate. Ingratiation is known to increase smiling (e.g., Godfrey, Jones, & Lord, 1986). 4 Results from the present studies relating to participants’ accuracy of recalling each other’s nonverbal cues were published in Hall, Carter, and Horgan (2001), and results pertaining to owners’ versus assistants’ overall levels of smiling were published in Hall, Horgan, and Carter (2002).
BRIEF REPORT
their smiling was coded from the videotapes. The design of these studies is very similar to that of Hecht and LaFrance (1998). To facilitate the summary and comparison of results across studies, the methodology for all three studies is presented first, followed by an integrated Results section.
Method Participants Participants in all three studies (N ⳱ 480) were students recruited from the participant pool of the Northeastern University Department of Psychology in fulfillment of introductory psychology course requirements (Study 1: 52 men, 68 women; Study 2: 58 men, 122 women; Study 3: 54 men, 126 women). In addition to these participants, 180 students were recruited from the participant pool to rate smiling in Study 2.
Procedure In all three studies, participants could be randomly assigned to the owner (high power) or assistant (low power) role, and in Studies 2 and 3, they could also be assigned to an equal-power condition. After role assignment, participants received complementarily worded instructions. The instructions to the owner in Study 1 were as follows: In your first activity, you are the owner of an art gallery and the other participant is your assistant. You are considering giving the assistant a promotion, and to make your decision, you have decided to ask the assistant to help you choose some artwork for the gallery; that way you can evaluate the assistant’s taste and judgment about art and get a better feel for how it would be to work closely with the assistant. We will provide you with various pieces of art, and you will choose the best 3 for your gallery, using input from the assistant any way you wish. You will have 5 to 7 minutes in which to discuss the art and choose the best 3. You will also be given a form on which you will evaluate the assistant’s ideas and manner. In the second activity, you will still be owner of the art gallery. You and your assistant will be building a creative structure to be used as a window display. You will build this out of commonplace materials such as paper cups, straws, and toothpicks. I will give you more detailed instructions later, but the important thing to understand right now is that as art gallery owner, your job is to help guide the two of you toward making the best structure you can, while evaluating the assistant’s performance. Again, you will be given an evaluation form to complete.
The assistant’s instructions were modified to be consistent with the assistant role.
305
After reviewing their instructions, the participants were brought together in the laboratory, where the video camera was in plain view. The experimenter assigned seats, turned on the recording equipment, and left the participants to complete the art discussion task. Following this, participants were videotaped for 10 min while building an aesthetically pleasing structure out of everyday materials. Participants completed the postexperimental questionnaire in separate rooms. Studies 2 and 3 differed from Study 1 in the following ways: The assistant was described as applying for the assistant’s job (rather than hoping for a promotion), and accordingly, a 6-min job interview with the owner was the first task, and the art discussion task (described above) was the second task. The building task was not used in Studies 2 and 3. The wording of the role inductions was also strengthened by repeating key phrases at a later point in the instructions.5 In the equal-power condition in Studies 2 and 3, participants were assigned to be co-owners of the art gallery; during the art discussion task, they participated equally, and during the job interview task, they took turns being interviewer and interviewee as though practicing to interview an applicant.
Manipulation Check for Felt Power In Studies 1 and 2, the following five postexperimental questions were combined to form a power composite: I felt I had some authority/power over the other person, I felt the other person looked up to me while we were playing our roles, I felt that I was the dominant one in the interaction, I felt that the other person had some authority over me (reverse scored), and I felt as though I had less power in the interaction than the other person (reverse scored), ranging from –3 (strongly disagree) to +3 (strongly agree). In Study 3, the following six items were combined into a power composite: I feel I have some authority/ power over the other person, Maintaining a feeling of authority is important to me in these activities, I feel the other person looks up to me while we are playing our roles, It is important for me to show the other person that I am in control of the activities we perform, I feel that I am the dominant one in the inter-
5 In all three studies, additional experimental manipulations were designed to influence the role attitudes of the participants (see Hall, Carter, & Horgan, 2001; Hall, Horgan, & Carter, 2002, for details). These are not relevant to the present article and are not discussed further.
306
BRIEF REPORT
action, ranging from –3 (strongly disagree) to +3 (strongly agree); and During my interaction with my partner, I felt dominant, ranging from 1 (not at all) to 9 (a great deal). The mean responses of assistants and owners are shown in Table 1. Matched t tests were highly significant (p < .0001) in each study, indicating that owners reported feeling more power than assistants reported.
Self-Ratings of Happy Affect In all three studies, participants were asked in the postexperimental questionnaire how happy and cheerful they had felt in their interaction, with both items ranging from 1 (not at all) to 9 (a great deal). The two ratings were averaged to form a happy-affect composite for each dyad member (see Table 1). Cronbach’s alphas were .71 (assistants) and .73 (owners) in Study 1, .83 (assistants) and .81 (owners) in Study 2, and .92 (assistants) and .92 (owners) in Study 3. Assistants and owners did not differ in their happy affect (see Table 1).
Coding of Smiling From Videotape Only 1 participant in each dyad was scored at a time (the other participant’s image was covered by a piece of paper taped to the video monitor), and the videotape was scored with the sound turned off. In Study 1, a trained coder counted the number of smiles in Minutes 1, 2, and 3 of the art discussion task and Minutes 1, 2, 4, and 5 of the building task. Frequencies were summed across coded minutes and tasks. In Study 2, smiles were rated by 180 naive judges who spent 1 hr rating smiling every 30 s for 4 or 5 participants from different dyads during Minutes 1, 2, 4, and 5 of each task on a Likert scale ranging from 1 (hardly at all) to 9 (nearly all the time). Each participant was rated by four naive judges. Ratings were summed across 30-s time blocks and tasks. In Study 3, a trained coder counted the occurrence of smiles in Minutes 1, 2, 4, and 5 of each task. Fre-
quencies were summed across coded minutes and tasks. Intercoder reliability (assessed as correlations between the trained coder and an independent coder, or alphas among raters) was .99 in Study 1, .82 in Study 2, and .94 in Study 3.
Results Table 2 presents correlations between self-rated happy affect and smiling for low-power, high-power, and equal-power participants, showing first the samegender dyads (along with the Hecht & LaFrance, 1998, results, which were only for same-gender dyads), followed by opposite-gender dyads and both genders combined. For same-gender dyads, a metaanalytic summary (Rosenthal, 1991) is provided for our three studies and also for our three studies combined with the Hecht and LaFrance (1998) study. Both summaries show that the correlations for lowand high-power groups were positive and similar. The table shows that the mean correlations were essentially identical whether calculated with or without weighting by sample size, and the p values indicate a very credible overall effect in both groups. The middle section of Table 2 presents correlations between happy affect and smiling for opposite-gender dyads. These correlations were somewhat smaller than those for same-gender dyads, although again, the correlations for low- and high-power participants were all positive, and both combined p values were significant. When the results for same-gender dyads were compared with those for opposite-gender dyads (within power conditions), using a planned contrast between independent groups of correlations (Rosenthal, 1991), neither contrast was significant. The bottom portion of Table 2, showing all participants in our three studies, also shows that low- and high-power correlations were similar. Table 2 also presents the correlations for equalpower individuals, both in our studies and in Hecht and LaFrance (1998). Using Hecht and LaFrance’s
Table 1 Means for Power Manipulation Check and Self-Rated Happy Affect for Studies 1–3 Study 1
Study 2
Study 3
Variable
Assistants
Owners
Assistants
Owners
Assistants
Owners
Power manipulation check Happy affect
−0.68 7.00
1.19a 7.48d
−0.77 6.77
1.25b 6.84e
−0.02 6.63
1.69c 6.28f
Note. The t values represent comparisons between owners and assistants. a t(58) ⳱ 8.29, p < .0001. b t(59) ⳱ 8.70, p < .0001. c t(59) ⳱ 6.08, p < .0001. d t(58) ⳱ 1.52, p < .14. e t(59) ⳱ 0.27, p < .79. f t(59) ⳱ 1.26, p < .22.
307
BRIEF REPORT Table 2 Correlations Between Self-Rated Happy Affect and Smiling for Low-, High-, and Equal-Power Participants in Studies 1–3 and in Hecht and LaFrance (1998) Study
Low power
High power
Equal power
Study 1 Study 2 Study 3 Mean r (unweighted)a Mean r (weighted) Combined Z Combined p (one-tailed) Hecht and LaFrance (1998) Mean r (unweighted)b Mean r (weighted) Combined Z Combined p (one-tailed)
Same-gender .20 (31) .35* (33) .56*** (34) .37 .38 3.68 .001 −.16 (24) .25 .29 2.80 .01
.12 (32) .27 (33) .30† (35) .24 .24 2.31 .01 .55** (24) .32 .30 3.34 .001
— .11 (38) .42** (38) .27 .27 2.31 .01 .38† (24) .31 .29 2.96 .001
Study 1 Study 2 Study 3 Mean r (unweighted)a Mean r (weighted) Combined Z Combined p (one-tailed)
Opposite-gender .16 (28) .13 (27) .21 (24) .17 .16 2.55 .005
.18 (28) .42* (27) .15 (24) .25 .25 2.23 .01
— −.04 (22) .17 (20) .06 .06 0.40 .34
Study 1 Study 2 Study 3 Mean r (unweighted)a Mean r (weighted) Combined Z Combined p (one-tailed)
Genders combined .19 (59) .23† (60) .38** (58) .26 .26 3.54 .001
.10 (60) .33** (60) .22† (59) .22 .22 2.89 .01
— .13 (60) .20 (58) .16 .16 1.79 .05
Note. Sample n is given in parentheses. Dashes mean this condition was not run. a Meta-analytic summary is based on Studies 1–3. b Meta-analytic summary is based on Studies 1–3 and Hecht and LaFrance (1998). † p < .10. *p < .05. **p < .01. ***p < .001.
(1998) results along with our three studies (both genders combined), we calculated a contrast between the three equal-power correlations and the eight unequalpower correlations (Rosenthal, 1991). There was no significant difference (Z ⳱ 0.01; weighted mean correlation for equal power ⳱ .20, weighted mean correlation for unequal power ⳱ .24). Furthermore, a homogeneity test (Rosenthal, 1991) conducted across these 11 correlations was not significant, indicating that Hecht and LaFrance’s correlations can be viewed as reflecting normal sampling variation within the larger distribution of effect sizes, 2(10) ⳱ 10.95, p > .30.
Discussion In this article, we examined the relation between overall smiling and self-ratings of happy affect in
three studies in which power roles were experimentally assigned. We found the relation to be of moderate magnitude and fairly consistent across studies and not different between individuals assigned to lowversus high-power roles. Hecht and LaFrance (1998), in a similarly designed study, found a different result: The correlation with overall smiling was substantially stronger among those assigned to a high-power role than those assigned to a low-power role. When our results were combined with Hecht and LaFrance’s using meta-analytic methods, the average correlations did not differ between power roles, and, furthermore, the entire set of results (including three equal-power groups—two from our studies and one from Hecht & LaFrance, 1998) was statistically homogeneous, indicating a generally positive, moderately sized correla-
308
BRIEF REPORT
tion between self-rated happy affect and overall smiling. Many factors can, of course, account for why the relation between happy affect and smiling is not larger. People smile for many reasons besides feeling cheerful or happy (e.g., Brunner, 1979; Chovil, 1991/ 1992; Keltner, 1995; Kraut & Johnston, 1979; Milgram, 1974). As noted earlier, the type of smile moderates the relation, with Duchenne smiles showing a stronger relation with self-rated happy affect. And methodological factors could also be moderators, such as the timing of the measurement of happy affect (Rosenberg & Ekman, 1994). It is also likely that social psychological factors moderate the happy affect–smiling association. Individuals could find themselves in situations or in roles that increase or decrease this association, owing to the operation of display rules. Similarly, interpersonal power could moderate the association if role-related constraints produced a disjunction between a person’s current mood and his or her desire or need to smile. To observe such a possibility in future studies, relative power differences may need to be actual as opposed to experimentally created, and positive affect may need to be manipulated as opposed to measured via self-reports. However, we suspect that, even under the most ideal experimental setups, it would be important to know more about the motives of the individuals whose smiling is being observed. For example, a motive to ingratiate could attenuate the correlation, whereas a motive to behave sincerely may increase it. But one should not assume that any specific motive is the prerogative of those in a particular power role; for instance, the motive to be pleasing could be held by either low- or high-power individuals (Hall, Horgan, & Carter, 2002). The foregoing suggests that measuring different kinds of smiles may also be important. Hecht and LaFrance (1998) did, in fact, measure Duchenne smiles separately from non-Duchenne smiles. The presumed moderating effect of power should be stronger for Duchenne smiles (because the Duchenne smile is expected to be more closely related to felt affect and because Hecht and LaFrance assumed that highpower individuals have license to show felt affect). However, their data did not show this; the moderating effect of power was stronger for overall and nonDuchenne smiling. The correlations could be affected by differential amounts of Duchenne and nonDuchenne smiling shown by the two groups. However, this was not the case because Hecht and LaFrance did not find a difference between power
groups in the type of smile. In our research, we focused on an attempt to replicate their strongest pattern of results concerning overall smiling.6 The presumed moderating effect for overall smiling could not be replicated. Further research is needed to examine the potential role of power as a moderator. Such research should be guided by theoretically based predictions about the regulation of felt affect and the use of strategic affect expression in different contexts, making use of operational hypotheses concerning the functional role of different types of smiles and other facial actions as suggested by the literature. 6 In the present studies, Duchenne smiling could not be measured because of the distance of the camera from the participants.
References Biehl, M., Matsumoto, D., Ekman, P., Hearn, V., Heider, K., Kudoh, T., & Ton, V. (1997). Matsumoto and Ekman’s Japanese and Caucasian Facial Expressions of Emotion (JACFEE): Reliability data and cross-national differences. Journal of Nonverbal Behavior, 21, 3–21. Bonnano, G. A., & Keltner, D. (1997). Facial expressions of emotion and the course of conjugal bereavement. Journal of Abnormal Psychology, 106, 126–137. Brunner, L. J. (1979). Smiles can be back channels. Journal of Personality and Social Psychology, 37, 728–734. Chovil, N. (1991/1992). Discourse-oriented facial displays in conversation. Research on Language and Social Interaction, 25, 163–194. Deutsch, F. M. (1990). Status, sex, and smiling: The effect of role on smiling in men and women. Personality and Social Psychology Bulletin, 16, 531–540. Ekman, P., Davidson, R. J., & Friesen, W. V. (1990). The Duchenne smile: Emotional expression and brain physiology II. Journal of Personality and Social Psychology, 58, 342–353. Ekman, P., Friesen, W. V., & Ancoli, S. (1980). Facial signs of emotional experience. Journal of Personality and Social Psychology, 39, 1125–1134. Fridlund, A. J., Ekman, P., & Oster, H. (1987). Facial expressions of emotion: Review of literature, 1970–1983. In A. W. Siegman & S. Feldstein (Eds.), Nonverbal behavior and communication (2nd ed.). Hillsdale, NJ: Erlbaum. Friedman, H. S., & Miller-Herringer, T. (1991). Nonverbal display of emotion in public and in private: Selfmonitoring, personality, and expressive cues. Journal of Personality and Social Psychology, 61, 766–775.
BRIEF REPORT Gifford, R. (1994). A lens-mapping framework for understanding the encoding and decoding of interpersonal dispositions in nonverbal behavior. Journal of Personality and Social Psychology, 66, 398–412. Godfrey, D. K., Jones, E. E., & Lord, C. G. (1986). Selfpromotion is not ingratiating. Journal of Personality and Social Psychology, 50, 106–115. Hall, J. A., Carter, J. D., & Horgan, T. G. (2001). Status roles and recall of nonverbal cues. Journal of Nonverbal Behavior, 25, 79–100. Hall, J. A., Carter, J. D., Jimenez, M. C., Frost, N. A., & Smith LeBeau, L. (2002). Smiling and relative status in news photographs. Journal of Social Psychology, 142, 500–510. Hall, J. A., & Friedman, G. B. (1999). Status, gender, and nonverbal behavior: A study of structured interactions between employees of a company. Personality and Social Psychology Bulletin, 25, 1082–1091. Hall, J. A., Horgan, T. G., & Carter, J. D. (2002). Assigned and felt status in relation to observer-coded and participant-reported smiling. Journal of Nonverbal Behavior, 26, 63–81. Hall, J. A., Smith LeBeau, L., Gordon Reinoso, J., & Thayer, F. (2001). Status, gender, and nonverbal behavior in candid and posed photographs: A study of conversations between university employees. Sex Roles, 44, 677– 691. Hecht, M. A., & LaFrance, M. (1998). License or obligation to smile: The effect of power and sex on amount and type of smiling. Personality and Social Psychology Bulletin, 24, 1332–1342. Hess, U., Banse, R., & Kappas, A. (1995). The intensity of facial expression is determined by underlying affective state and social situation. Journal of Personality and Social Psychology, 69, 280–288. Jakobs, E., Manstead, A. S. R., & Fischer, A. H. (1999a). Social motives and emotional feelings as determinants of facial displays: The case of smiling. Personality and Social Psychology Bulletin, 25, 424–435.
309
Jakobs, E., Manstead, A. S. R., & Fischer, A. H. (1999b). Social motives, emotional feelings, and smiling. Cognition and Emotion, 13, 321–345. Johnson, C. (1994). Gender, legitimate authority, and leader–subordinate conversations. American Sociological Review, 59, 122–135. Keltner, D. (1995). Signs of appeasement: Evidence for the distinct displays of embarrassment, amusement, and shame. Journal of Personality and Social Psychology, 68, 441–454. Kraut, R. E., & Johnston, R. E. (1979). Social and emotional messages of smiling: An ethological approach. Journal of Personality and Social Psychology, 37, 1539– 1553. Markham, R., & Adams, K. (1992). The effect of type of task on children’s identification of facial expressions. Journal of Nonverbal Behavior, 16, 21–39. Milgram, S. (1974). Obedience to authority. New York: Harper & Row. Rosenberg, E. L., & Ekman, P. (1994). Coherence between expressive and experiential systems in emotion. Cognition and Emotion, 8, 201–229. Rosenthal, R. (1991). Meta-analytic procedures for social research (Rev. ed.). Newbury Park, CA: Sage. Rosenthal, R., & DePaulo, B. M. (1979). Sex differences in eavesdropping on nonverbal cues. Journal of Personality and Social Psychology, 37, 273–285. Ruch, W. (1993). Extraversion, alcohol, and enjoyment. Personality and Individual Differences, 16, 89–102. Scherer, K. R., & Ceschi, G. (2000). Criteria for emotion recognition from verbal and nonverbal expression: Studying baggage loss in the airport. Personality and Social Psychology Bulletin, 26, 327–339.
Received October 31, 2002 Revision received May 15, 2003 Accepted May 15, 2003 ■
