Explaining sex differences in the evaluation of comforting messages: the mediating role of...

The current study hypothesized that sex differences in evaluations of comforting messages could be accounted for in terms of the extent to which people indicated they would pursue affectively oriented versus instrumentally oriented goals in support situations. Participants (N = 292; 165 men

and 127 women) completed (a) an interaction goals task that assessed the pursuit of affectivety oriented and instrumentally oriented goals in emotional support situations, and Co) a message evaluation task that assessed perceptions of the sensitivity and effectiveness of 27 comforting messages that varied in level of "person centeredness." Consistent with expectations, females evaluated the comforting messages in a manner significantly more consistent with Applegate's (1980) theoretical hierarchical ordering of person centeredness than did males. More important, the likelihood of pursuing affectively oriented goals mediated the connection between sex and perceptions of message sensitivity and effectiveness.

**********

* Research examining sex differences in emotional support has consistently found that there are some important ways that men and women differ in the performance of comforting behavior. For example, some research has found that women tend to produce more "person-centered" messages in comforting situations (e.g., Hale, Tighe, & Mongeau, 1997; Samter, this issue). Within the constructivist framework, person centeredness refers to "message behavior that reflects an awareness of and adaptation to the subjective, affective, and relational aspects of communicative contexts" (Burleson, 1987, p. 305). Highly person-centered comforting messages typically provide an explicit acknowledgment and elaboration of another's feelings and perspective, while less person-centered comforting messages typically ignore feelings, dismiss them, attempt to shift attention away from them, or challenge their legitimacy.

In addition to differences between men and women in the production of person-centered comforting messages, there are also sex differences in the evaluation of these messages (e.g., Burleson & Samter, 1985b; Jones & Burleson, 1997; Samter, Burleson, & Murphy, 1987; Samter, Whaley, Mortenson, & Burleson, 1997). Typically, it has been found that although both men and women view highly person-centered comforting messages as more sensitive and effective than messages exhibiting a low level of person centeredness, men (on average) do not rate low person-centered messages quite as negatively as do women (on average), while women rate highly person-centered messages somewhat more positively than do men. For example, in a detailed examination of sex differences in comforting message evaluations, Kunkel and Burleson (1999) asked participants to rate the sensitivity and effectiveness of 27 comforting messages varying in level of person centeredness (i.e., nine messages each for three separate situations). These researchers found that although men and women evaluated the comforting messages in substantially similar ways, women tended to rate highly person-centered messages somewhat more favorably than did men and rated less person-centered messages less favorably than did men. Thus, there are small, but reliable, sex differences in the evaluations of person-centered comforting messages.

Evaluations of comforting messages are theoretically and pragmatically important. People's judgments about what count as sensitive and effective comforting strategies (a) express their preferences for comforting efforts that others produce for them, (b) may indicate what kind of comforting messages they produce when seeking to help others, and (c) may indicate the kind of messages people direct to themselves when attempting to cope privately with some upset. It is likely that those comforting messages that people view as helpful to others in specific contexts of distress are the messages that individuals would want others to employ with them. Further, though not the subject of much empirical scrutiny to date, message evaluations may predict the kinds of comforting messages people produce when attempting to help distressed others. At the very least, people won't be motivated to use a particular message strategy with others if they don't see that strategy as effective and helpful. Finally, it is possible that the comforting strategies that people view as sensitive and effective with others may reflect those things that people tell themselves when seeking to cope privately with an upsetting event. In sum, evaluations of comforting messages are indicators of what people believe to be useful in support situations, and speak to the values and goals that may be enacted in comforting interactions by both producers and recipients of support messages.

Why do men and women differ (to the limited extent they do) in their evaluations of comforting messages exhibiting different levels of person centeredness? One plausible reason is that men and women typically develop somewhat different goals in emotional support situations, and these goals may serve as the basis for evaluating the sensitivity and effectiveness of different messages. As Burleson and Mortenson (2001, pp. 10-11) explain:

[T]he typical interaction goals people develop for a certain class of situation (such as providing emotional support) might well influence judgments of the quality and appropriateness of particular messages and behaviors targeted for use in those situations. Interaction goals reflect what people want to accomplish in a situation; they express people's pragmatic orientation to an interactional situation (Dillard, 1997). Goals are used, both projectively and retrospectively, to judge the appropriateness or suitability of a behavior (Gollwitzer & Brandstatter, 1997). Thus, it seems reasonable to view typical interaction goals as standards that people use to judge the quality and appropriateness of communicative behaviors.

Interaction goals are often conceptualized as states that "an individual wants or desires to attain" (Miller, Cody, & McLaughlin, 1994, p. 171) as a result of the exchanges they experience with others. In recent years, scholars of strategic communication (e.g., Dillard, 1997; Wilson, 1997) have emphasized that many communicative situations are open to alternative interpretations or definitions, giving rise to the pursuit of varied interaction goals. Because interaction goals reflect how social situations are interpreted and defined, the goals pursued in comforting situations ought to impact not just choices related to the production of supportive messages, but also the evaluation of the messages produced by others. For instance, an individual who focuses on the feelings of a distressed other in support situations, and therefore privileges assuaging the other's distressed emotions, is likely to also favor messages that are steeped in emotional content. Simply put, individual differences in goal orientations should wield predictive power with respect to the evaluation of comforting messages.

Recent research suggests that people may pursue one or more of several distinct goals in support situations, including avoidance, emotion management, and problem management (Goldsmith & Dun, 1997; MacGeorge, 2001). Some research indicates that people do occasionally seek to avoid helping others in support situations (see Barbee & Cunningham, 1995). However, perhaps because the preponderance of research on emotional support has examined this process in the context of close relationships, most studies have focused on the extent to which helpers pursue instrumental (problem-focused) and affective (emotion-focused) goals (e.g., Barbee, Gulley, & Cunningham, 1990; Derlega, Barbee, & Winstead, 1994). Affective or emotion-focused goals involve a focus on listening to and helping another to work through and talk about feelings, whereas instrumental or problem-focused goals focus on giving advice and helping another to solve problems so that they might move on with life.

Previous research has found that men and women tend to exhibit differences in goal orientations in support situations. This research shows that women tend to be more affectively oriented or emotion-focused, while men tend to be more instrumentally oriented or problem-focused (see the reviews by Cutrona, 1996; Reis, 1998). Women also are inclined to be more nurturant and prosocially oriented, while men are disposed to be more instrumental and project-focused (Caldwell & Peplau, 1982; Gilligan, 1982). Further, although Kunkel and Burleson (1999) found that both men and women viewed affective goals as more important than instrumental goals in comforting situations, women tended to rate affective goals as somewhat more important than did men whereas men rated instrumental goals as slightly more important than did women (see also Samter et al., 1997).

The present research seeks to determine whether sex differences in sensitivity and effectiveness evaluations of comforting messages stem from the goals that people find most relevant in emotional support situations. If so, a mediated-effects model in which goal orientations impact message evaluations, and are impacted by sex, would be evidenced by a substantial reduction of sex differences in evaluations when controlling for differences in goal orientations. As no research has yet attempted to explain the documented sex differences in message evaluations in terms of support goal orientation, this study will ascertain whether sex differences in the extent to which highly person-centered messages are evaluated as sensitive and effective (and less person-centered messages are evaluated as insensitive and ineffective) remain when controlling for the effects of goal orientations.

Thus, the purpose of the present study is to assess the validity of a mediational model that connects biological sex, interaction goal orientations in support situations, and sensitivity and effectiveness ratings of comforting messages that vary in person centeredness. This effort entails the reanalysis of data previously published in Human Communication Research by Kunkel and Burleson (1999). Kunkel and Burleson found that while sex did not strongly determine evaluations of comforting messages, nor which goals were considered more important in comforting situations, sex did influence message evaluations, as well as the extent to which both affective and instrumental goals were considered important. However, Kunkel and Burleson did not consider whether differences in goals might influence or mediate the effects of sex on evaluations. The rationale developed above suggests several hypotheses:

H1: The importance assigned to instrumental goals in support situations will mediate the effects of sex on evaluations of comforting messages.

H2: The importance assigned to instrumental goals in support situations will be associated (a) negatively with evaluations of comforting messages exhibiting low levels of person centeredness, and (b) positively with evaluations of comforting messages exhibiting high levels of person centeredness.

H3: The importance assigned to affective goals in support situations will mediate the effects of sex on evaluations of comforting messages.

H4: The importance assigned to affective goals in support situations will be associated (a) negatively with evaluations of comforting messages exhibiting low levels of person centeredness, and (b) positively with evaluations of comforting messages exhibiting high levels of person centeredness.

METHOD

This study involves the reanalysis of data previously published by Kunkel and Burleson (1999), so details regarding the methods of the study can be found there. Essential features of the study are briefly summarized here.

Participants

165 men and 127 women (N = 292) from a large midwestern university participated in this study (92.1% were Caucasian). Participants ranged in age from 17 to 50 with a mean age of 21.29.

Procedure

Participants came to a research session that lasted approximately one hour. Upon arrival to the research session, each participant received a packet of questionnaires that included a background information questionnaire, as well as instructions and appropriate forms for a comforting message evaluation task and a supportive goal priority task.

Message evaluation task. To elicit evaluations of the perceived sensitivity and effectiveness of comforting messages varying in degree of person centeredness, participants completed a message evaluation task derived from the work of Burleson and Samter (1985a). This task contained a list of nine preformulated comforting messages that might be used with a distressed other in three different situations (i.e., a parental separation situation, a lost scholarship situation, and a romantic relationship break-up situation). The messages used with each situation were written to reflect each of the nine levels of person centeredness in Applegate's (1980) hierarchical coding system for comforting messages (see Kunkel & Burleson, 1999, as well as Burleson & Samter, 1985a for additional details regarding this task and sample messages). Participants were asked to rate separately the sensitivity and effectiveness of each preformulated message on five-point rating scales. Table 1 (which is based on Table 3 from Kunkel & Burleson, 1999) presents the means for sensitivity and effectiveness ratings for each of the strategy levels across the three stimulus situations. Internal consistencies for ratings of message sensitivity and effectiveness over the three situations were .95 for the sensitivity ratings and .92 for the effectiveness ratings. (1)

Supportive goal priority task. To determine the priority participants assigned to affective and instrumental goals when dealing with distressed others in emotional situations, participants completed a supportive goal priority task. Participants responded to the question, "In general, when someone is emotionally upset about something, how important is it to: (a) work through their feelings, (b) solve their problems, (c) talk about what they are feeling, (d) provide the right advice about how to handle the situation, (e) help them blow off steam by allowing them the freedom to express their thoughts and feelings, and (f) suggest ways of putting the problem behind them so they can get on with life?" For each of the six items, participants were asked to rate on five-point scales the importance of the implied communicative goal. Three of the items (a, c, and e) tapped the priority assigned to affective goals, with an emphasis on the acknowledgment and expression of feelings. The other three items (b, d, and f) tapped the priority assigned to instrumental goals, with an emphasis on suggestions and action. These emphases were selected to represent the critical features of each goal orientation as defined within the extant research (e.g., Barbee & Cunningham, 1995; Barbee et al., 1990). Table 2 (which is based on Table 5 from Kunkel & Burleson, 1999) presents the means for males and females on the supportive goal priority task. Cronbach's alpha for affective goal priority was .67; for instrumental goal priority, it was .64. The moderate internal consistencies for these variables may be a function of the small number of items used to tap these constructs (N = 3).

RESULTS

To assess Hypotheses 1, 2, 3, and 4 and determine whether a mediated-effects model fits the data, three necessary assumptions must be met: (a) the independent variable must be associated with the mediating variable, (b) the mediating variable must be associated with the dependent variable, and (c) the independent variable must be associated with the dependent variable (Pettit, Harrist, Bates, & Dodge, 1991; see also Biddle & Marlin, 1987). For the current analysis, sex is treated as the independent variable, goal orientations (i.e., affective and instrumental) serve as the potentially mediating variables, and ratings of message sensitivity and effectiveness are the dependent variables.

Thus, the first obligatory procedure for assessing whether the mediated-effects model fits the data is to assess simple zero-order correlations among all of the variables. Table 3 displays the zero-order correlations among sex, affective goals, and instrumental goals. Table 4 displays the zero-order correlations of instrumental goal orientation and sensitivity and effectiveness ratings for each message level. Tables 5 and 6 display the zero-order correlations between affective goal orientation and sensitivity and effectiveness ratings for each message level, as well as the zero-order correlations between sex and sensitivity and effectiveness ratings for each message level.

Table 3 shows that the independent variable, sex, is not correlated with instrumental goal orientation. Thus, instrumental goal orientation, as a potential mediating variable, was dropped from subsequent analysis and there was no support for Hypothesis 1. Further, as Table 4 indicates, instrumental goal orientation was not correlated with any of the comforting messages exhibiting low levels of person centeredness, but was correlated with sensitivity and effectiveness ratings of messages exhibiting high levels of person centeredness. Thus, there is only limited support for Hypothesis 2.

In contrast, Table 3 shows that the independent variable, sex, was substantially correlated with affective goal orientation (r = .41). Moreover, the mediating variable, affective goal orientation, was correlated with six of the nine message sensitivity ratings (Table 5) and six of the nine message effectiveness ratings (Table 6). Finally, the independent variable, sex, was correlated with six of the nine message sensitivity ratings (Table 5) and five of the nine message effectiveness ratings (Table 6).

To test whether affective goal orientation mediated the effects of sex on sensitivity and effectiveness ratings (Hypothesis 3), partial correlational analyses were employed. Tables 5 and 6 present these partial associations and the percent of sex-related variance in message ratings mediated by affective goal orientation. (2) Overall, the results of these analyses indicate that 26.26% to 75.36% of the sex-related variance in sensitivity ratings is mediated by affective goal rating, and 33.59% to 96.06% of the sex-related variance in effectiveness ratings is mediated by affective goal rating. Clearly, then, Hypothesis 3 was supported. That is, the import accorded to affective goals in support situations substantially mediated the relationship between sex and perceptions of message sensitivity and effectiveness. Interestingly, the percentage of sex-related variance mediated by the importance assigned to affective goals was particularly large for messages either very low or very high in person centeredness.

Finally, the importance assigned to affective goals in support situations was associated negatively with sensitivity and effectiveness evaluations of comforting messages exhibiting low levels of person centeredness, and positively with sensitivity and effectiveness evaluations of comforting messages exhibiting high levels of person centeredness (see Tables 5 and 6). These results provide support for Hypothesis 4.

DISCUSSION

Two major results were uncovered by the present study. The first was the status of instrumental goal orientation as an unproductive mediating variable of the relationships between sex and the evaluation of comforting messages. The second, and perhaps most noteworthy, result was the mediation by affective goals of large shares of sex-related variance in evaluations of messages exhibiting high, or low, levels of person centeredness. On average, affective goal orientation mediated nearly half of the sex-related variance in sensitivity ratings and two-thirds of the sex-related variance in effectiveness ratings.

The finding that instrumental goal orientation failed to mediate the effect of sex on message evaluations is consistent with several recent studies suggesting that an instrumental orientation plays little role in the comforting process. For example, Burleson and Gilstrap (this issue) found no mediating effect for the personality trait of instrumentality with respect to sex-related differences in specific support-related goals. It is also worth noting that both Burleson and Gilstrap (this issue) and the present study found no association between sex and instrumental goal orientation in support situations.

The three items utilized to assess instrumental goal orientation focused on the importance of solving problems, providing the right advice, and putting the problem behind you and getting on with life (see Kunkel & Burleson, 1999 for a full description). It may be that, in the context of trying to help someone who is emotionally upset, an instrumental focus is not all that useful with respect to evaluating the relative sensitivity and effectiveness of comforting messages (especially those that vary in degrees of person centeredness).

Unlike instrumental goal orientation, and consistent with Hypothesis 3, affective goal orientation did largely mediate the relationship between sex and perceptions of message sensitivity and effectiveness. The results indicate that affective goal orientation mediated, on average, half of the sex-related variance in message evaluations (approximately 47% for sensitivity ratings and 67% for effectiveness ratings). It appears that sex differences in evaluations of comforting messages stem from the kinds of goals that men and women are likely to define as most relevant (and likely to pursue) in emotional support situations. In particular, if people view affectively oriented goals in such situations as highly relevant, then they are likely to view highly person-centered messages as more sensitive and effective (and less person-centered messages as more insensitive and ineffective) compared to those who are less affective in their orientations.

Affective goals typically involve a focus on listening and helping another to work through and talk about feelings. Thus, it follows that appreciation of these behaviors influences the extent to which individuals perceive person-centered messages as more (or less) sensitive and effective in the provision of support. Indeed, closer inspection of both the instrument utilized to assess affective goal orientation and the hierarchy consulted for the construction of messages varying in person centeredness suggests that an emphasis on working through the feelings associated with distress is crucial to perceptions of effective comforting. The three items that measured affective goal orientation addressed the importance of working through feelings, talking about what the distressed is feeling, and allowing the distressed freedom to express his or her thoughts and feelings (see Kunkel & Burleson, 1999 for a full description). In Applegate's (1980) hierarchy of comforting messages, a low-level message "condemns or ignores specific feelings," a mid-level message "provides some implicit acceptance of feelings," and a high-level message "explicitly acknowledges, elaborates, and legitimizes the feelings" of the distressed (see Burleson & Samter, 1985a).

The percentage of sex-related variance mediated by the importance assigned to affective goals was especially large for messages that were either low or extremely high in person centeredness. Further, and supportive of Hypothesis 4, for both sensitivity and effectiveness ratings, the importance assigned to affective goals in support situations was associated negatively with comforting messages exhibiting low levels of person centeredness, and positively with comforting messages exhibiting high levels of person centeredness. Thus, possessing a stronger affective goal orientation enables an individual to more finely discriminate among comforting messages and to keenly recognize what is both most helpful and most unhelpful in support situations.

The existence of affective goal orientation as a mediator of sex-related differences in message evaluations may help serve to explain sex-related differences in comforting message preference and production. As noted earlier, the sexes tend to vary less with respect to comforting message preference than with respect to comforting message production. While both sexes prefer highly person-centered comforting, they also recognize that females are more likely to be the source of such messages (e.g., Kunkel & Burleson, 1999). Women view affectively oriented goals as somewhat more important than do men in the context of comforting. Apparently, in the current study, affective goal orientations were expressed in the rating of messages that vary in level of person centeredness so that the extreme high and low levels are more appreciated and more scorned, respectively. Therefore, if the message evaluations reported in this study do signal preferences for, and predict production of, comforting messages, the extant findings regarding sex differences in preference and production might be conceptualized as the results of different degrees of affective goal orientations adopted by the members of each sex.

Overall, the findings of this study suggest that sex differences in comforting message evaluations stem from differences in interaction goal orientations. Thus, affective goal orientations may create preferences for, and tendencies toward the production of, person-centered messages. Of course, other important variables that mediate the relationship between sex and evaluations of emotionally supportive communication may exist as well. Future research should explore how personality traits, cognitive competencies, cultural orientations, and comforting expertise explain sex differences in comforting messages.

TABLE 1
Means and Standard Deviations for Sensitivity
and Effectiveness Ratings

                      Sensitivity Ratings

                  Males               Females         Marginal

Strategy
 Level         M         SD         M        SD       M      SD

   1       1.38 (a)     0.52    1.11 (b)    0.21    1.26    0.44
   2       1.59 (a)     0.58    1.22 (b)    0.34    1.43    0.53
   3       2.11 (a)     0.56    1.69 (b)    0.51    1.93    0.58
   4       2.56 (a)     0.75    2.38 (b)    0.73    2.48    0.75
   5       3.22         0.62    3.18        0.65    3.21    0.63
   6       3.41         0.59    3.45        0.51    3.43    0.56
   7       3.29 (a)     0.55    3.44 (b)    0.58    3.36    0.57
   8       3.95         0.66    4.01        0.58    3.98    0.62
   9 *     4.14 (a)     0.69    4.32 (b)    0.62    4.22    0.66
Marginal   2.85 (a)     0.30    2.76 (b)    0.31    2.81    0.31

                     Effectiveness Ratings

                  Males               Females         Marginal

Strategy
 Level         M         SD         M        SD       M      SD

   1       1.61 (a)     0.60    1.35 (b)    0.45    1.50    0.56
   2       1.78 (a)     0.62    1.48 (b)    0.59    1.65    0.62
   3       2.31 (a)     0.62    1.88 (b)    0.58    2.12    0.64
   4       2.68         0.77    2.61        0.78    2.65    0.78
   5       2.72         0.69    2.69        0.72    2.71    0.71
   6       3.46         0.52    3.51        0.54    3.48    0.53
   7       3.25         0.59    3.30        0.64    3.27    0.61
   8       3.69 (a)     0.68    3.87 (b)    0.74    3.77    0.71
   9 *     4.00 (a)     0.72    4.19 (b)    0.71    4.08    0.72
Marginal   2.83         0.32    2.76        0.34    2.80    0.31

NOTE: N = 292. Within each row, means with different superscripts
significantly differ from one another (p < .05, two-tailed). Ratings
were on five-point scales with 1 = very insensitive or ineffective) and
5 = very sensitive (or effective). * Most person-centered comforting
strategy.
TABLE 2
Means and Standard Deviations for Males and Females
on the Supportive Goal Priority Task

                             Sex of Subject

                         Male            Female           Marginal

Goal Type            M        SD       M        SD       M        SD

Affective         4.17 (a)   0.55   4.61 (b)   0.43     4.36     0.55
Instrumental      3.68       0.67   3.79       0.75     3.73     0.70
Marginal          3.92       0.49   4.20       0.45     4.04     0.50

NOTE: N = 292. Within each row, means with different superscripts
significantly differ from one another (p < .05, one-tailed). Ratings
for goal priority were on five-point scales with 1 = not at all
important and 5 = very important.
TABLE 3
Intercorrelations Among Sex, Affective Goals,
and Instrumental Goals

        Variables             (1)        (2)         (3)

(1) Sex                       1.0         --          --
(2) Affective Goals           .41 ***    1.0          --
(3) Instrumental Goals        .08         .24 ***    1.0

NOTE: N = 292. * p < .05; ** p < .01; *** p < .001
(two-tailed tests); males = 1; females = 2.
TABLE 4
Summary of Associations between Sensitivity and
Effectiveness Ratings and Instrumental Goal Rating

              Association between      Association between
               Sensitivity Rating      Effectiveness Rating
 Message       and Instrumental         and Instrumental
  Level           Goal Rating              Goal Rating

    1                -.042                    -.023
    2                -.050                    -.006
    3                 .018                     .051
    4                 .058                     .103
    5                -.027                     .062
    6                -.032                     .086
    7                 .080                     .077
    8                 .133 *                   .164 **
    9                 .122 *                   .143 *

NOTE: * p < .05; ** p < .01; *** p < .001 (two-tailed tests).
TABLE 5
Summary of Associations between Sensitivity Rating and Affective
Goal Rating, Zero-Order Associations between Sex and Sensitivity
Rating, Partial Associations of Sex and Sensitivity Rating
Controlling for Affective Goal Rating, and Percent of Sex-Related
Variance in Sensitivity Rating Mediated by Affective Goal Rating

             Association
               between          Zero-Order
             Sensitivity        Association
              Rating and        between Sex
 Message      Affective       and Sensitivity
  Level      Goal Rating          Rating

    1         -.313 ***          -.305 ***
    2         -.313 ***          -.351 ***
    3         -.214 ***          -.368 ***
    4         -.119 *            -.122 *
    5          .045              -.032
    6          .059               .032
    7          .072               .132 *
    8          .161 **            .049
    9          .186 **            .137 *

                  Partial           Percent of Sex-
               Association of       Related Variance
            Sex and Sensitivity      in Sensitivity
             Rating Controlling     Rating Mediated
 Message       for Affective          by Affective
  Level         Goal Rating           Goal Rating

    1            -.205 ***               54.82%
    2            -.258 ***               45.97%
    3            -.316 ***               26.26%
    4            -.082                   54.82%
    5            -.055                     --
    6             .009                     --
    7             .113                   26.72%
    8            -.019                     --
    9             .068                   75.36%

NOTE: * p < .05; ** p < .01; *** p < .001 (two-tailed tests);
males = 1; females = 2.
TABLE 6
Summary of Associations between Effectiveness Rating and
Affective Goal Rating, Zero-Order Associations between Sex and
Effectiveness Rating, Partial Associations of Sex and Effectiveness
Rating Controlling for Affective Goal Rating, and Percent
of Sex-Related Variance in Effectiveness Rating
Mediated by Affective Goal Rating

              Association
                between           Zero-Order
             Effectiveness        Association
              Rating and          between Sex
 Message       Affective       and Effectiveness
  Level       Goal Rating           Rating

    1          -.258 ***           -.235 ***
    2          -.233 ***           -.239 ***
    3          -.226 ***           -.335 ***
    4          -.091               -.046
    5          -.078               -.025
    6           .137 *              .049
    7           .012                .045
    8          .267 ***             .131 *
    9          .247 ***             .134 *

             Partial Association     Percent of Sex-
                 of Sex and          Related Variance
                Effectiveness        in Effectiveness
             Rating Controlling      Rating Mediated
 Message        for Affective          by Affective
  Level          Goal Rating           Goal Rating

    1             -.148 *                 60.34%
    2             -.163 **                53.49%
    3             -.273 ***               33.59%
    4             -.011                     --
    5              .007                     --
    6             -.007                     --
    7              .044                     --
    8              .026                   96.06%
    9              .039                   91.53%

NOTE: * p < .05; ** p < .01; *** p < .001 (two-tailed tests);
males = 1; females = 2.

ENDNOTES

(1.) Previous research (e.g., Goldsmith, McDermott, & Alexander, 2000; Jones & Burleson, 1997) has suggested that ratings of message effectiveness, sensitivity, and helpfulness tend to be highly correlated. To assess whether an overall index of perceived strategy quality could be created, the extent to which sensitivity and effectiveness ratings were correlated was examined (also see Kunkel, 1994). Three different analyses were performed. First, the correlation between sensitivity and effectiveness ratings (averaged over situation) was computed for each message level. Second, averaged ratings for the nine strategies over the three situations were calculated and then correlated (collapsing across strategy level). Third, mean sensitivity and effectiveness ratings for each strategy type were computed for each of the three situations; these means were then correlated with one another (collapsing across individual). While the third method yielded a near perfect correlation between sensitivity and effectiveness ratings (r = .98), the first two methods yielded no correlations above the .70 level for any strategy level or across strategy levels. Thus, and consistent with previous research (e.g., Burleson & Samter, 1985a), the analyses in this study were performed separately for the sensitivity and effectiveness measures.

(2.) In their discussion of various modeling procedures, Biddle and Marlin (1987) argue for the use of the "simplest available techniques that will enable us to display and test results plausibly" (p. 16). Essentially, the procedure used in the testing of Hypotheses 3 and 4 implemented a simple form of path analysis. These analyses were performed for the purpose of determining the percentage of sex-related variance in both sensitivity and effectiveness ratings that is mediated by affective goal ratings. This analysis is similar to commonly recognized regression approaches for assessing mediation effects (e.g., Biddle & Marlin, 1987; Cohen & Cohen, 1983), but it is more compact and economical.

REFERENCES

Applegate, J. L. (1980). Adaptive communication in educational contexts: A study of teacher's communicative strategies. Communication Education, 29, 158-170.

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., Gulley, M. R., & Cunningham, M. R. (1990). Support seeking in personal relationships. Journal of Social and Personal Relationships, 7, 531-540.

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

Burleson, B. R. (1987). Cognitive complexity. In J. C. McCroskey & J. A. Daly (Eds.), Personality and interpersonal communication (pp. 305-349). Newbury Park, CA: Sage.

Burleson, B. R., & Gilstrap, C. M. (this issue). Explaining sex differences in interaction goals in support situations: Some mediating effects of expressivity and instrumentality. Communication Reports.

Burleson, B. R., & Mortenson, S. R. (2001). Explaining cultural differences in evaluations of emotional support behaviors: Mediating influences of value systems and interaction goals. Paper presented at the 3rd Annual INPR-ISSPR joint conference, Prescott, AZ.

Burleson, B. R., & Samter, W. (1985a). Consistencies in theoretical and naive evaluations of comforting messages. Communication Monographs, 52, 103-123.

Burleson, B. R., & Samter, W. (1985b). Individual differences in the perception of comforting messages: An exploratory investigation. Central States Speech Journal, 36, 39-50.

Caldwell, M. A., & Peplau, L. A. (1982). Sex differences in same-sex friendship. Sex Roles, 8, 721-732.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.

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

Derlega, V. J., Barbee, A. P., & Winstead, B. A. (1994). Friendship, gender, and social support: Laboratory studies of supportive interactions. In B. R. Burleson, T. L. Albrecht, & I. G. Sarason (Eds.), Communication of social support.. Messages, interactions, relationships, and community (pp. 136-151). 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.

Gilligan, C. (1982). In a different voice.. Psychological theory and women's development. Cambridge, MA: Harvard University Press.

Goldsmith, D. J., & Dun, S. A. (1997). Sex differences and similarities in the communication of social support. Journal of Social and Personal Relationships, 14, 317-337.

Goldsmith, D.J., McDermott, V. M., & Alexander, S. C. (2000). Helpful, supportive and sensitive: Measuring the evaluation of enacted support in personal relationships. Journal of Social and Personal Relationships, 17, 369-391.

Gollwitzer, P. M., & Brandstatter, V. (1997). Implementation intentions and effective goal pursuit. Journal of Personality and Social Psychology, 73, 186-199.

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.

Jones, S. M., & Burleson, B. R. (1997). The impact of situational variables on helpers' perceptions of comforting messages: An attributional analysis. Communication Research, 24, 530555.

Kunkel, A. W. (1994). Assessing the adequacy of explanations for gender differences in emotional support: An experimental test of the different cultures and skill deficit accounts. Unpublished master's thesis, Purdue University, West Lafayette, IN.

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.

MacGeorge, E. L. (2001). Support providers' interaction goals: The influence of attributions and emotions. Communication Monographs, 68, 72-97.

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.

Pettit, G. S., Harrist, A. W., Bates, J. E., & Dodge, K. A. (1991). Family interaction, social cognition and children's subsequent relations with peers at kindergarten. Journal of Social and Personal Relationships, 8, 383-402.

Reis, H. T. (1998). Gender differences in intimacy and related behaviors: Context and process. In D. Canary & K. Dindia (Eds.), Sex differences and similarities in communication: Critical essays and empirical investigations of sex and gender in interaction (pp. 203-231). Mahwah, NJ: Erlbaum.

Samter, W. (this issue). How gender and cognitive complexity influence the provision of emotional support: A study of indirect effects. Communication Reports.

Samter, W., Burleson, B. R., & Murphy, L. B. (1987). Comforting conversations: The effects of strategy type on evaluations of messages and message producers. The Southern Speech Communication Journal, 52, 263-284.

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

Wilson, S. R. (1997). Developing theories of persuasive message production: The next generation. In J. O. Greene (Ed.), Message production.. Advances in communication theory (pp. 15-43). Mahwah, NJ: Erlbaum.

Adrianne Kunkel (Ph.D., Purdue University, 2000) is an Assistant Professor in the Department of Communication Studies, University of Kansas, 102 Bailey Hall, 1440 Jayhawk Blvd., Lawrence, KS 66045. Phone: 785-864-9884; FAX: 765-864-5203; E-mail: adkunkel@ku.edu. A version of this manuscript was presented at the annual meeting of the International Communication Association, Washington, DC, May 25-28, 2001.