Explaining sex differences in interaction goals in support situations: some mediating effects...

The current study examined the extent to which personality orientations such as expressiveness and instrumentality mediate the effects of sex on goal orientations in support situations. Participants (N = 184; 89 males and 95 females) completed a questionnaire that provided assessments of biological

sex, instrumental and expressive personality orientations, and interaction goals in support contexts. Regression procedures indicated that the influence of sex on communication goals was indirect, with sex primarily associated with personality orientation, and personality orientation influencing two of the four communication goals. The personality trait of expressiveness also exerted a direct influence on the goal of discussing feelings. Interestingly, personality did not predict the goal of talking about the problem. The success of this model suggests that it is possible to develop more explanatory accounts of how sex influences varied forms of social behavior.

**********

* Recent research has revealed patterns of both similarity and difference in assessments of men's and women's social support orientations and behaviors (see the reviews by Burleson, 1997; Kunkel & Burleson, 1998; Reis, 1998). For example, growing research indicates that members of both sexes have very similar ideas about what counts as a supportive message, similar values concerning the importance of support in close relationships, and similar preferences regarding the gender of persons from whom they receive support (e.g., Burleson, Kunkel, Samter, & Werking, 1996; Kunkel & Burleson, 1999). Despite these similarities, sex-based differences continue to be reported in the character of the goals people pursue in support situations (e.g., Kunkel & Burleson, 1999; Samter, Whaley, Mortenson, & Burleson, 1997), as well as in the behaviors employed in the effort to realize those goals (e.g., Barbee et al., 1993; Burleson, 1982; Hale, Tighe, & Mongeau, 1997). In accord with prevailing stereotypes, most of the research has found that men take a somewhat more instrumental, action-focused approach to support situations while women take a somewhat more expressive, emotion-focused approach (see the reviews by Cutrona, 1996; Vaux, 1985). In addition, women appear more inclined to offer support to those experiencing emotional distress than do men (e.g., George, Carroll, Kersnick, & Calderon, 1998; Vaux, 1985), perhaps because men are less inclined than women to discuss others' emotions, as well as their own (Jansz, 2000; Saurer & Eisler, 1990). Although the magnitudes of these sex differences are often small, they do tend to be reliable.

The purpose of the present study was to develop and assess a mediational model that connects biological sex, gender-based personality traits, and interaction goals in emotional support situations. Although certain theorists have suggested how some of these constructs are connected with others, a comprehensive model relating all of them simultaneously has not been proposed and tested. Recent debates concerning the study of sex and gender differences in support contexts (e.g., Burleson, 1997; Vangelisti, 1997; Wood, 1997; Wood & Dindia, 1998) underscore the need to develop comprehensive conceptual models capable of integrating aspects of sex, gender, personality, and interaction.

In most contemporary treatments of the message production process, goals are viewed as highly proximal generative mechanisms for the messages people produce (see Dillard, 1997; Miller, Cody, & McLaughlin, 1994). Barbee and Cunningham (1995; Barbee, Lawrence, & Cunningham, 1998) have proposed a four-category typology of strategies actors can employ in support situations when confronting a distressed other. This typology was developed by crossing the dimensions of approach versus avoid and problem-focus versus emotion-focus. Thus, when individuals avoid dealing with a distressed other's emotions, they enact the strategy of escape; when they approach the other's emotions they enact the strategy of solace; when they avoid dealing with the other's problem they enact the strategy of dismiss; and when they approach the other's problem, they enact the strategy of solve. This typology of behavioral strategies rather directly suggests a complementary typology of interaction goals. That is, actors in support situations can be viewed as pursuing (to greater or lesser extents) the goals of escape, dismiss, solace, and solve.

Previous research has detected sex differences in some of the interaction goals suggested by Barbee and Cunningham's work (1995). For example, Kunkel and Burleson (1999) found that both men and women gave greater priority to approach-oriented, emotion-focused goals than to approach-oriented, problem-focused goals in support situations. Further, although men and women did not differ in the importance they assigned approach-oriented, problem-focused goals in such situations, women did view approach-oriented, emotion-focused goals as significantly more important than did men. Similar results were obtained by Samter et al. (1997).

What aspects of gender (understood as the social and cultural meaning associated with biological sex) might account for these sex differences in interaction goals in support situations? In contemporary Western societies, cultural conceptions of femininity are closely associated with concepts of emotionality and expressivity while cultural conceptions of masculinity are closely associated with rationality and instrumentality (e.g., Bern, 1974; Eagly, Wood, & Diekman, 2000; Spence & Buckner, 1995). Internalization of these cultural ideal types over the course of development gives rise to more-or-less stable individual differences in the personality traits of men and women, with men exhibiting greater instrumentality and women exhibiting greater expressiveness. Consistent with this view, a large body of research indicates that there are substantial differences in the personality traits of contemporary American men and women, with men being more instrumental than women, and women being more expressive than men (for a summary of research, see the meta-analysis by Feingold, 1994).

To the extent that males and females internalize our culturally celebrated conceptions of masculinity and femininity, men may be more inclined to avoid interacting with emotionally distressed persons than women. Jansz (2000) and others (e.g., Levant, 1995) argue that conceptions of masculinity in contemporary American society lead men to inhibit the expression and discussion of their most intimate feelings. Consistent with this, men report that interacting with emotional others is more stressful than do women (Saurer & Eisler, 1990), perhaps resulting in men being less likely than women to provide support to peers suffering from emotional upsets (George et al., 1998). In addition, when men do intervene in support situations, masculinity norms may lead them to manifest a more instrumental orientation while women manifest a more expressive orientation. As one example of this, MacGeorge (1998) found that a feminine, expressive personality orientation was associated with placing higher value on emotion-focused forms of communication (e.g., comforting, ego support) while a masculine, instrumental personality orientation was associated with valuing instrumental forms of communication (e.g., informing, persuading). Such communication values may influence the interaction goals people develop for specific communicative encounters. The expressive-instrumental aspect of personality has also been associated with the behaviors people employ in support situations. For example, Winters and Waltman (1997) found that women scoring high on a measure of expressivity generated more "person-centered" comforting messages (which explicitly acknowledge, elaborate, and legitimize feelings) than did women scoring low on the expressivity measure. In a similar vein, Shamblen, Cunningham, and Barbee (1999) found that expressivity was positively associated with the use of emotional support strategies in supportive interactions. Interestingly, Shamblen et al. also found that high-expressive individuals used more problem-solving strategies in these supportive interactions.

In sum, existing research indicates that (a) men and women exhibit some important differences in the interaction goals they prioritize and pursue in support situations, (b) men and women differ substantially with regard to the personality traits of instrumentality and expressivity, and (c) highly expressive individuals value comforting skills, utilize sensitive "person-centered" messages when seeking to provide comfort, and are more likely to employ both emotion-focused and problem-focused strategies in support situations. Thus far, however, no research has directly examined the association between the personality traits of instrumentality and expressivity and the pursuit of various interaction goals in support situations. More important, no research has sought to explain the documented sex differences in support goals in terms of the underlying personality traits of instrumentality and expressivity. Clearly, though, the available findings suggest that the effects of sex on interaction goals in support situations should be mediated by the personality traits of instrumentality and expressivity.

We sought to test the mediational model implied by the line of reasoning developed above. Specifically, we predicted that:

HI: Women will be more likely than men to pursue approach-oriented, emotion-focused goals (solace) in support situations, and less likely than men to pursue avoidance-oriented, emotion-focused goals (escape).

H2: Men will be more likely than women to pursue approach-oriented, problem-focused goals (solve) in support situations, and less likely than women to pursue avoidance-oriented, problem-focused goals (dismiss).

H3: Expressivity will mediate the effect of biological sex on the importance accorded emotion-focused (solace, escape) goals in support situations.

H4: Instrumentality will mediate the effect of biological sex on the importance accorded problem-focused (solve, dismiss) goals in support situations.

METHOD

Participants

Participants in the study were 184 undergraduate college students (89 males and 95 females) attending a large Midwestern university. These participants ranged in age from 18 to 26; 13.6% were 18-19, 66.3% were 20-21, 16.3% were 22-23, and 3.8% were 24-26. The participants were largely Caucasian (174 of 184, or 94.6%); there were 3 African-Americans, 3 Asian-Americans, 3 Hispanic Americans, and 1 participant who reported "other" as an ethnic affiliation. Participants received information about the study and were asked to volunteer to participate; no compensations were used to induce participation. All participants received and completed an informed consent statement regarding the study.

Procedure

Participants received a copy of a study questionnaire and an answer sheet from a research assistant. They were directed to read the cover sheet of the questionnaire, which explained the general purpose of the study (finding out "what young adults thought about communication in close relationships"), and then follow detailed instructions for each portion of the questionnaire. The questionnaire was written to be self-administering. Participants recorded responses to all questions on a separate answer sheet and then returned the answer sheet and questionnaire booklet to the research assistant. The questionnaire booklet consisted of several different instruments. Three of these are relevant to the current study: a brief demographic questionnaire, an assessment of instrumentality and expressivity, and an assessment of interaction goals in support situations.

Instrumentation

Demographic background. Participants answered a brief series of questions concerning their sex, age, and ethnicity.

Instrumentality and expressivity. Individual differences in the personality traits of instrumentality and expressivity were assessed with the 16-item version of Spence and Helmreich's (1978) Personal Attributes Questionnaire (PAQ). The PAQ was utilized in the current study since it provides cleaner assessments of instrumentality and expressivity than alternatives such as Bem's (1974) Sex Role Inventory (see Spence & Buckner, 1995). Considerable evidence supports the reliability and validity of the PAQ as a personality trait measure of instrumentality and expressiveness (see Spence & Buckner, 1995).

The PAQ instructs participants to rate themselves on a series of bipolar personality trait terms (e.g., not at all independent--very independent; not at all emotional--very emotional); in the current study, a five-point scale separated the two polar terms. Participants were told to describe what kind of people they thought they were by indicating for each pair of terms where they fell on the five-point scale. Eight of the term pairs pertained to instrumental characteristics (e.g., competitive, makes decisions easily) and the other eight term pairs pertained to expressive characteristics (e.g., kind, understanding of others).

Consistent with previous work (e.g., Spence & Helmreich, 1978), the 16 items loaded on two distinct factors (principal axis extraction, oblimin rotation). Internal consistency (Cronbach's alpha) for the 8-item expressivity scale was .76; internal consistency for the 8-item instrumentality scale was .75.

Interaction goals for support situations. Participants' interaction goals for support situations were assessed with Burleson's (1999) Support Goals Inventory (SGI), a refined version of the support goals measure developed by Kunkel and Burleson (1999). The current version of the SGI provides a measure of participants' evaluations of the four goals relevant in support situations suggested by Barbee and Cunningham (1995): dismiss, escape, solace, and solve.

The version of the SGI employed in this study describes two situations to participants in which a good friend is depicted as distressed by an upsetting circumstance. In one situation the friend is upset about having received a speeding ticket; the friend is portrayed as worried that his or her parents might stop paying for car insurance or might take away the friend's car while he or she is in college. In the other situation, the friend is upset about having recently discovered that his or her boyfriend/girlfriend for the last year-and-a-half has been cheating on him or her; the friend is depicted as both angry and despondent. Accompanying each of these situations was a list of 12 statements describing goals that might be pursued when talking to the upset friend. There were three items for each of the four goals of escape, dismiss, solve, and solace (e.g., escape, "avoid embarrassing talk about emotions"; dismiss, "get his/her attention off the troubling situation"; solve, "give him/her the right advice about how to handle the situation"; solace, "help him/her talk about what he/she is feeling"). Using a 5-point scale anchored by "not at all important" and "very important," participants rated each goal for how important they thought it would be when talking to their distressed friends.

A principal axis factor analysis with a varimax rotation forcing a four-factor solution found that 22 of the 24 items loaded as intended (i.e., all solve items had primary loadings on the solve factor, etc.). However, one escape item for each situation loaded somewhat more strongly on the dismiss factor than the escape factor. Overall, however, the factor analysis indicated that the items clustered as intended.

Reliabilities (Cronbach's alphas) were computed for each of the four support goals by assessing the internal consistency of the six items for each goal (i.e., the three items for each goal for the two situations). Reliabilities were as follows: dismiss = .87, escape = .85, solace = .80, and solve = .76. Given the acceptable reliabilities, scores were averaged over the six items for each of the four goals.

RESULTS

Assessment of Sex Differences in Support Goals Hypotheses 1 and 2, which predicted sex differences in the four support goals, were assessed initially with a 2 (sex) x 4 (goal type) mixed-model ANOVA. Sex was a between-groups factor, goal type was a repeated measures factor, and the dependent variable was the rated importance of the goal. Cell means and standard deviations for this analysis are presented in Table 1.

The ANOVA indicated that the main effect for subject sex was not statistically significant, F(1,182) = 0.82, ns. However, the ANOVA detected a very strong effect for the factor of goal type, F (3, 546) = 262.71, p < .001, [eta.sup.2] = .59. In addition, the ANOVA detected a significant interaction effect between the factors of sex and goal type, F (3, 546) = 15.52, p < .001, [eta.sup.2] = .08.

The significant interaction between sex and goal type was decomposed by evaluating the effects of sex for each goal type. Women (M = 2.12) were significantly less likely than men (M = 2.47) to purse the goal of escape, t (182) = 2.66, p < .01. In addition, women (M = 4.37) were significantly more likely than men (M = 3.80) to pursue the goal of solace, t (182) = - 6.26, p < .001. These results strongly support Hypothesis 1.

Men (M = 3.58), as predicted, were more likely than women (M = 3.31) to pursue the goal of solve, t (182) = 2.45, p < .05. However, for the goal of dismiss, men (M = 2.41) and women (M = 2.21) did not differ significantly. These results provide only partial support for Hypothesis 2.

Assessment of the Mediating Effects of Expressivity and Instrumentality

To assess the extent to which there were dependencies among participants' evaluations of the four interaction goals (dismiss, escape, solace, and solve) a correlation analysis was conducted on the mean importance ratings assigned these goals. The results of this analysis are summarized in Table 2. Importance ratings for the two avoidance goals (dismiss and escape) were strongly associated, r = .57, p < .001. In contrast, the ratings for the two approach goals (solace and solve) were unassociated, r = - .05. Rating for the two emotion-focused goals (escape and solve) were negatively associated, as would be expected, r = - .21, p < .01. Surprisingly, however, ratings for the two problem-focused goals (dismiss and solve) were positively associated, r = .33, p < .01. Overall, these findings suggest that those inclined to avoid the provision of support do so both with respect to problem-focused and emotion-focused assistance. In contrast, having the goal of providing problem-focused support is completely unrelated with the goal of providing emotion-focused support.

Hypotheses 3 and 4 predicted that the personality traits of expressivity and instrumentality would mediate the effects of biological sex on the pursuit of interactional goals in support situations. To assess this mediational model, it is necessary that three conditions be met: (a) the independent variable of interest (here, sex) must be associated with the dependent variables (here, the four support goals), (b) the mediational variables (here, expressivity and instrumentality) must be associated with the dependent variables, and (c) the independent variable must be associated with the mediational variables. The ANOVAs reported previously indicate that the first of these assumptions was not met for the goal of dismiss; thus, this goal was excluded from further analyses. To assess the remaining assumptions, correlations among the variables included in the study were calculated. These correlations are reported in Table 2. The correlational analyses indicated that instrumentality was not associated with any of the interaction goals; thus, the potential mediating variable of instrumentality was excluded from further analyses. Further, expressivity was not associated with the solve goal, so this goal was dropped from the mediational analysis. Expressivity was significantly associated with both sex and the two goals of escape and solace. Thus, the conditions for the test of mediation were met only by the personality trait of expressivity, and only for support goals of escape and solace.

Regression analyses were used to assess the mediating effects of expressivity for the association between sex and both escape and solace goals. Two regression equations were calculated for each goal. In the first equation, sex was entered at the initial step and expressivity was entered subsequently at a second step. The second equation reversed the order of entry for the two predictors. Comparison of the increments in explained variance for each predictor in the two equations permits a specification of whether, and how much, variance in a dependent variable explained by sex is mediated by expressivity (for details on this approach to the assessment of mediation, see Biddle & Marlin, 1987).

Table 3 summarizes the results of the regression analyses for the two interaction goals. Sex and expressivity collectively explained a small, but significant, portion of the variance in the importance assigned to the goal of escape, R2 = .051, F (2, 181) = 4.90, p < .01. When entered in the regression equation first, sex explained a statistically significant 3.7% of the variance in escape; however, when entered in the equation after expressivity, sex explained only 1.8% of the variance in escape. Thus, approximately half of the sex-related variance in the importance assigned the escape goal was mediated by expressivity. Sex had a marginally significant (p < .07) direct effect on escape even after controlling for expressivity. Expressivity uniquely explained a nonsignificant 1.4% of the variance in escape. As predicted, all betas for expressivity were negative, indicating that as expressivity increased, people were less likely to pursue the goal of escape.

Sex and expressivity collectively explained ar substantial, and significant, portion of the variance in the importance assigned to the goal of solace, [R.sup.2] = .305, F (2, 181) = 39.80, p < .001. When entered in the regression equation first, sex explained 17.7% of the variance in solace; however, when entered in the equation following expressivity, sex explained only 6.4% of the variance in solace (see Table 3). Thus, approximately two-thirds of the sex-related variance in the importance assigned the solace goal was mediated by expressivity. Sex continued to exert a significant (p < .001) direct effect on solace even when controlling for expressivity. Expressivity uniquely accounted for 12.8% of the variance in solace. All betas for expressivity were positive, indicating that participants were more likely to pursue the goal of solace as their level of expressivity increased.

DISCUSSION

Consistent with our expectations, men and women differed in the degree of importance assigned interaction goals in support situations. More important, we found, as predicted, that for at least some of these goals, sex differences were mediated by the personality variable of expressivity.

Specifically, men and women differed in expected ways for three of the four interaction goals we examined with respect to support situations. Women were more likely to pursue the goal of "solace," and less likely to pursue the goal of "escape" than were men. In contrast, men were more likely to pursue the goal of "solve" than were women. The sexes did not differ for the "dismiss" goal. These patterns of difference are largely consistent with literature suggesting that women are more emotion focused in support situations while men are more problem focused (e.g., Cutrona, 1996; Vaux, 1985).

Although the sex differences we detected in interaction goals are important, our results also indicated that there were very substantial similarities in men's and women's interaction goals. Both men and women indicated that the goal of "solve" would be more important in support situations than the goals of "escape" or "dismiss." And both men and women indicated that the goal of "solace" was more important in these situations than the goal of "solve." Thus, while we detected some noteworthy sex differences in interaction goals, these differences exist within a much stronger pattern of similarity. This pattern of small sex differences existing within larger similarities has been observed for several other facets of cognition, affect, and behavior in support situations (see Kunkel & Burleson, 1998, 1999). Caution must be exercised in interpreting these results, however, due to our use of self-report methods in assessing interaction goals. Research assessing supportive behaviors with observational methods (see Barbee et al., 1998; Shamblen et al., 1999) suggests that avoidance-oriented behaviors (dismiss and escape) are considerably more common than the self-reports obtained in the present study appear to indicate. It is also possible that self-report methods such as those employed in the current study attenuate the magnitude of sex differences manifest in actual support interactions.

Contrary to our expectations, instrumentality was not associated with the importance accorded any interaction goal in support situations. This is surprising. Moreover, our results are not consistent with results of Shamblen et al. (1999), who found that high instrumentals provided more emotional support to distressed targets than did low instrumentals. Shamblen et al. collected diary data from participants over a two-week period regarding their supportive interactions, and thus may well have obtained a more reliable assessment of supportiveness than that obtained in the current study. Perhaps the lack of an effect for instrumentality in our study had something to do with the stimulus situations utilized in the current version of the SGI. Neither of these situations (involving a friend who had received a speeding ticket and a friend who had discovered that his/her dating partner was cheating) afforded much opportunity for instrumental intervention. Future versions of the SGI should, therefore, include a broader array of stimulus situations that permit instrumental as well as affective responses.

In contrast to the results for instrumentality, expressivity was associated in expected ways with interaction goals in support situations. Expressivity was negatively associated with escape and dismiss, and positively associated with solace. These results are consistent with previous research finding expressivity contributing to emotion-focused behavior in support situations (e.g., MacGeorge, 1998; Shamblen et al., 1999; Winters & Waltman, 1997).

Interestingly, the importance assigned the approach-oriented, problem-focused goal of "solve" was significantly associated only with sex; men rated this goal as more important than did women. This highly "instrumental" goal was not itself predicted by the personality trait of instrumentality. Perhaps an effect for instrumentality would be found if the situations used in the SGI provided more of an opportunity for instrumental action. In any event, there is a sex difference with respect to pursuit of this goal--a difference that remains to be accounted for.

We proposed a mediational model in which personality traits are viewed as mediating the effects of biological sex on interaction goals in support situations. This model could be evaluated in the current study only for the emotion-focused goals of escape and solace. For both of these goals, expressivity mediated the effects of sex on goal pursuit in the predicted manner (negatively for escape, and positively for solace). Expressivity mediated about half the sex-related variance in escape, although the total variance explained for escape was small. In contrast, sex and expressivity explained a very substantial 30% of the variance in ratings for the solace goal. Expressivity mediated about two-thirds of the sex-related variance in solace.

These results clearly indicate that sex differences in support-related processes are amenable to theoretical explanations featuring constructs such as personality traits. Our results are encouraging, and suggest that researchers interested in sex differences should strive to identify proximal influences on behavior (such as interaction goals) and then specify the mechanisms through which sex may affect these proximal influences through the mediation of more explanatory constructs, like personality traits.

But just how is it that personality traits (such as expressivity) influence interaction goals (such as those pursued in support situations)? There appear to be at least two possible, nonexclusive mechanisms. First, personality traits such as expressivity may influence how social situations are interpreted and defined. That is, traits may "chronically prime" certain interpretive schemes, leading persons to notice and define as relevant particular features of situations. The interpretations arising from the application of these primed cognitive schemes thus orient the individual to the situation in characteristic ways, thereby influencing the goals the individual develops and pursues. To assess this particular account, research needs to examine whether individuals with certain personality traits interpret social situations in a characteristic manner.

A second explanatory mechanism linking traits and interaction goals exploits recent developments in emotion theory. Cognitive theories of emotion (e.g., Lazarus, 1991) maintain that interpretations and appraisals of situations may arouse particular emotional states. Emotions, in turn, serve as "action tendencies," orienting and preparing the individual to act in situationally relevant ways (see Parkinson, 1995). Such action tendencies inform (in conjunction with other factors) the generation of specific communicative goals relevant to that particular situation. It is possible that certain personality traits may be associated with distinct emotional experiences (see Frijda, 1993) that, in turn, generate characteristic goals. For example, people scoring high on the trait of expressiveness may be more inclined to experience the emotions of sympathy and compassion in support situations, which then motivate them to provide forms of support such as solace. Viewing personality traits as influencing communicative goals through the emotions these traits characteristically arouse addresses some of the objections that have been raised to personality-based explanations of message behavior (for a detailed discussion of this mechanism, see Burleson & Planalp, 2000).

Currently, however, we do not know whether personality traits affect interaction goals through their influence on the interpretive process, the emotions aroused by particular interpretations, the goals activated by particular emotions, or some combination of these. Moreover, it is quite possible that different traits affect interaction goals through different mechanisms. These issues constitute fertile areas for future research.

TABLE 1
Cell Means and Standard Deviations
for the Support Goals Inventory

Goal Type          Men              Women        Marginal

Escape       2.47 (a) (.89)    2.12 (b) (.92)      2.29
Dismiss      2.41     (.88)    2.21     (.85)      2.31
Solve        3.58 (a) (.71)    3.31 (b) (.77)      3.44
Solace       3.80 (a) (.72)    4.37 (b) (.51)      4.10

NOTE: Figures in parentheses are cell standard deviations.
Within each row, means having different superscripts differ
significantly (p < .05).
TABLE 2
Intercorrelations Among Variables Included in the Study

       Variables             (1)         (2)       (3)

(1) Sex                      --
(2) Expressivity            .38 **        --
(3) Instrumentality        -.23 **      -.12         --
(4) Escape Goal            -.19 **      -.18 *      .02
(5) Dismiss Goal           -.12         -.16 *     -.10
(6) Solve Goal             -.18 *       -.01        .06
(7) Solace Goal             .42 ***     .49 ***     .02

       Variables             (4)         (5)       (6)

(1) Sex
(2) Expressivity
(3) Instrumentality
(4) Escape Goal              --
(5) Dismiss Goal            .57 ***       --
(6) Solve Goal              .21 **       .33 **      --
(7) Solace Goal            -.21 **      -.39 **    -.05

NOTE: N = 184. * p < .05; ** p < .01; *** p < .001.
Sex was coded males = 0 and females = 1.
TABLE 3
Summary of Regression Analyses Assessing the Meditational
Effects of Expressivity on the Association Between Sex and
Support Interaction Goals of Escape and Solace

                               Sex Entered First
Dependent                    [R.sup.2]
 Variable    Predictor         Change      F Change

Escape       Sex                .037       7.07 **
             Expressivity       .014       2.67
Solace       Sex                .177      39.13 ***
             Expressivity       .128      33.49 ***

                  Sex Entered Second
Dependent    [R.sup.2]
 Variable      Change      F Change

Escape          .018       3.43 (+)
                .033       6.28 *
Solace          .064      16.73 ***
                .241      57.86 ***

NOTE: N = 184. (+) p < .10; * p < .05; ** p < .01; *** p < .001.

REFERENCES

Barbee, A. P., & Cunningham, M. R. (1995). An experimental approach to social support communications: Interactive coping in close relationships. In B. R. Burleson (Ed.), Communication yearbook (Vol. 18, pp. 381-413). Thousand Oaks, CA: Sage.

Barbee, A. P., Cunningham, M. R., Winstead, B. A., Derlega, V. J., Gulley, M. R., Yankeelov, P. A., & Druen, P. B. (1993). Effects of gender role expectations on the social support process. Journal of Social Issues, 49, 175-190.

Barbee, A. P., Lawrence, T., & Cunningham, M. R. (1998). When a friend is in need: Feelings about seeking, giving, and receiving social support. In P. A. Anderson & L. K. Guerrero (Eds.), Handbook of communication and emotion: Research, theory, applications, and contexts (pp. 281-301). San Diego, CA: Academic Press.

Bem, S. L. (1974). The measurement of psychological androgyny. Journal of Consulting & Clinical Psychology, 42, 155-162.

Biddle, B. J., & Marlin, M. M. (1987). Causality, confirmation, credulity, and structural equation modeling. Child Development, 58, 4-17.

Burleson, B. R. (1982). The development of comforting communication skills in childhood and adolescence. Child Development, 53, 1578-1588.

Burleson, B. R. (1997). A different voice on different cultures: Illusion and reality in the study of sex differences in personal relationships. Personal Relationships, 4, 229-241.

Burleson, B. R. (1999). Sex, gender, personality, values, and goals in emotional support contexts: Development and test of a model. Paper presented at the International Network on Personal Relationships annual meeting, Louisville, KY.

Burleson, B. R., Kunkel, A. W., Samter, W., & Werking, K. J. (1996). Men's and women's evaluations of communication skills in personal relationships: When sex differences make a difference--and when they don't. Journal of Social and Personal Relationships, 13, 201-224.

Burleson, B. R., & Planalp, S. (2000). Producing emotion(al) messages. Communication Theory, 10, 221-250.

Cutrona, C. E. (1996). Social support in couples. Thousand Oaks, CA: Sage.

Dillard, J. P. (1997). Explicating the goal construct: Tools for theorists. In J. O. Greene (Ed.), Message production: Advances in communication theory (pp. 47-70). Mahwah, NJ: Erlbaum.

Eagly, A. H., Wood, W., & Diekman, A. B. (2000). Social role theory of sex differences and similarities: A current appraisal. In T. Eckes & H. M. Trautner (Eds.), The developmental social psychology of gender(pp. 123-174). Mahwah, NJ: Erlbaum.

Feingold, A. (1994). Gender differences in personality: A meta-analysis. Psychological Bulletin, 116, 429-456.

Frijda, N. H. (1993). Moods, emotion episodes, and emotions. In M. Lewis & J. M. Haviland (Eds.), Handbook of emotions (pp. 381-403). New York: Guilford.

George, D., Carroll, P., Kersnick, R., & Calderon, K. (1998). Gender-related patterns of helping among friends. Psychology of Women Quarterly, 22, 685-704.

Hale, J. L., Tighe, M. R., & Mongeau, P. A. (1997). Effects of event type and sex on comforting messages. Communication Research Reports, 14, 214-220.

Jansz, J. (2000). Masculine identity and restrictive emotionality. In A. H. Fischer (Ed.), Gender and emotion: Social psychological perspectives (pp. 166-186). New York: Cambridge University Press.

Kunkel, A. W., & Burleson, B. R. (1998). Social support and the emotional lives of men and women: An assessment of the different cultures perspective. In D. J. Canary & K. Dindia (Eds.), Sex differences and similarities in communication (pp. 101-125). Mahwah, NJ: Erlbaum.

Kunkel, A. W., & Burleson, B. R. (1999). Assessing explanations for sex differences in emotional support: A test of the different cultures and skill specialization accounts. Human Communication Research, 25, 307-340.

Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press.

Levant, R. F. (1995). Toward the reconstruction of masculinity. In R. F. Levant & W. S. Pollack (Eds.), A new psychology of men (pp. 229-252). New York: Harper Collins.

MacGeorge, E. L. (1998, July). The influence of sex and psychological gender on the communication skills mature adults value in their same-sex friendships. Paper presented at the International Conference on Personal Relationships, Saratoga Springs, NY.

Miller, L. C., Cody, M. J., & McLaughlin, M. L. (1994). Situations and goals as fundamental constructs in interpersonal communication research. In M. L. Knapp & G. R. Miller (Eds.), Handbook of interpersonal communication (2nd ed., pp. 162-198). Thousand Oaks, CA: Sage.

Parkinson, B. (1995). Ideas and realities of emotion. London: Routledge.

Reis, H. T. (1998). Gender differences in intimacy and related behaviors: Context and process. In D. J. Canary & K. Dindia (Eds.), Sex differences and similarities in communication (pp. 203-232). Mahwah, NJ: Erlbaum.

Samter, W., Whaley, B. B., Mortenson, S. R., & Burleson, B. R. (1997). Ethnicity and emotional support in same-sex friendship: A comparison of Asian-Americans, African-Americans, and Euro-Americans. Personal Relationships, 4, 413-430.

Saurer, M. K., & Eisler, R. M. (1990). The role of masculine gender role stress in expressivity and social support network factors. Sex Roles, 23, 261-271.

Shamblen, S. R., Cunningham, M. R., & Barbee, A. P. (1999). The effects of instrumentality and expressivity on communication: Cheerful versus supportive interactions. Paper presented at the International Network on Personal Relationships annual conference, Louisville, KY.

Spence, J. T., & Buckner, C. (1995). Masculinity and femininity: Defining the undefinable. In P. J. Kalbfleisch & M. J. Cody (Eds.), Gender, power, and communication in human relationships (pp. 105-140). Hillsdale, NJ: Erlbaum.

Spence, J. T., & Helmreich, R. L. (1978). Masculinity and femininity.. Their psychological dimensions, correlates, and antecedents. Austin: University of Texas Press.

Vangelisti, A. L. (1997). Gender differences, similarities, and interdependencies: Some problems with the different cultures perspective. Personal Relationships, 4, 243-253.

Vaux, A. (1985). Variations in social support associated with gender, ethnicity, and age. Journal of Social Issues, 41, 89-110.

Winters, M., & Waltman, M. S. (1997). Feminine gender identity and interpersonal cognitive differentiation as correlates of person-centered comforting. Communication Reports, 10, 123-132.

Wood, J. T. (1997). Clarifying the issues. Personal Relationships, 4, 221-228.

Wood, J. T., & Dindia, K. (1998). What's the difference? A dialogue about differences and similarities between women and men. In D. J. Canary & K. Dindia (Eds.), Sex differences and similarities in communication (pp. 19-40). Mahwah, NJ: Erlbaum.

Brant R. Burleson (Ph.D., University of Illinois at Urbana-Champaign, 1982) is a professor in the Department of Communication, Purdue University, 1366 LAEB 2114, West Lafayette, IN 47907-1366. Cristina M. Gilstrap (M. A., Southwest Missouri State University, 1999) is a doctoral candidate in Communication at Purdue University. An earlier version of this paper was presented at the annual meeting of the International Communication Association, Washington, DC, May 25-28, 2001. Correspondence concerning this paper may be addressed to Brant Burleson, email brantb@purdue.edu.