Professional Documents
Culture Documents
Perceived Social Competence and Depression: Shelly L. Gable
Perceived Social Competence and Depression: Shelly L. Gable
Glenn D. Shean
College of William and Mary
ABSTRACT
Differences in impressions of social skills between depressed and non-depressed college student participants were investigated. Participants engaged in a 15-minute, face-to-face getting acquainted exercise with an assigned partner of the same sex. Based on Beck Depression Inventory scores, participants were classied as depressed or not and three types of dyads were studied: (i) depressed participant interacting with depressed participant; (ii) depressed participant interacting with non-depressed participant; and (iii) non-depressed participant interacting with non-depressed participant. Prior to the interaction, participants reported their mood. After the interaction, participants evaluated their own social competence and the competence of their partners, and completed a second measure of their mood. On all measures of social competence, depressed participants rated themselves as less competent than non-depressed participants. Depressed participants also rated their partners as less competent than nondepressed participants, regardless of the depression level of the partner. Participants mood prior to their interaction covaried with ratings of ones own skills but not ratings of partners skills. These results suggest that depressed individuals have a trait-like bias to perceive themselves and others in a negative manner, independent of the mood and manner of expression of their partners. Non-depressed participants did not rate depressed partners as less competent than non-depressed
Address correspondence to Shelly Gable, Department of Clinical and Social Sciences in Psychology, University of Rochester, Rochester, NY 14627, USA. [E-mail: [email protected]]. The Action Editor on this article was Susan Hendrick. The authors thank Harry Reis, Richard Gonzalez, John Nezlek, and an anonymous reviewer for advice regarding data analysis and comments on earlier versions of this manuscript. Data collection for this study was supported by a Minor Research Grant from the College of William & Mary, and preparation of the manuscript was supported by grant #F31 11766 from the National Institutes of Health, both awarded to the rst author. Journal of Social and Personal Relationships Copyright 2000 SAGE Publications (London, Thousand Oaks, CA and New Delhi), Vol. 17(1): 139150. [02654075 (200002) 17:1; 011285]
140
partners, suggesting that depression may be associated more closely with a cognitive set than a skills decit.
KEY WORDS
According to interpersonal theories of depression (Brown & Harris, 1978; Coyne, 1976, 1990; Hammen, 1997; Joiner, Alfano, & Metalsky, 1992; Joiner & Coyne, 1999; Segrin & Abramson, 1994), depressed/dysphoric individuals interactions are often dysfunctional and induce negative reactions in others. Research indicates that poor social skills and negative self-schemata contribute to problems with depression. A meta-analytic review of the literature on social skill decits has indicated, however, that depression is most strongly related to self-reports of social skill decits (Segrin, 1990). Dysphoric individuals rate themselves and are often judged by observers to be less competent in social interactions (Dykman, Horowitz, Abramson, & Usher, 1991). During interactions, dysphoric persons reportedly evidence a lack of responsiveness that is manifested, for example, in poor eye contact, long silences, and lack of acknowledgment of partners comments (Segrin & Abramson, 1994). However, observer ratings and detailed behavioral analyses of social skills during dyadic interactions reveal fewer consistent relationships with depression than do selfevaluations of depressives (Segrin & Dillard, 1992). Several theoretical interpretations have been proposed to explain the depressogenic aspects of dissatisfying or problematic interactions. Coyne (1976, 1990) has emphasized the role of an emergent pattern of depressive interactions, induced by the interpersonal behaviors of people who are depressed, with consequent ambivalence, negative mood, and withdrawal by friends and partners, culminating in covert rejection of the depressed person. This rejection serves, in turn, to increase the depressives sense of isolation and dysphoria. Lewinsohn, Mischel, Chaplin, and Barton (1980) maintain that depression can result from decits in such basic skills as expressions of warmth, friendliness, clear communication, interest in others, and reasonableness. In summary, current theories suggest that depressed individuals possess either a state dependent (Coyne) or trait-like (Lewinsohn) decit in social competence and that this decit has a negative impact on the quality of their social interactions. Studies examining the interactions of dysphoric students and their roommates over extended periods have indicated that their relationships tend to become more negative over time (Burchill & Stiles, 1988). Hokanson and Butler (1992), for example, reported that, over the course of an academic year, depressed students room-mates exhibited increased levels of hostility and a progressive decline in social contact with their dorm-mates. The relationship between depression and negative interactions is not limited to room-mates. Nezlek, Imbrie, and Shean (1994) studied the everyday social interactions of non-clinical depressives over a 3-week period. People who were classied as depressed reported that they had less satisfying and less intimate social interactions than those who were classied as not depressed, even though the frequency of their interactions did not differ. Despite evidence of dissatisfying and negative interactions associated with depression, the literature is inconsistent regarding the question of whether depressed people are actually less competent in terms of social skills (Dykman
141
et al., 1994; Lewinsohn, 1974), or simply perceive themselves as less competent (Beck, 1972). Segrins (1990) meta-analytic study of the relationship between depression and social skills indicates that negative self perceptions, rather than actual skills decits, are reported most consistently in the research literature. For example, Gotlib and Meltzers (1987) study of the accuracy of depressed peoples perceptions of their social competence is consistent with Segrins (1990) conclusions. Interactions between non-depressed participants and depressed or non-depressed targets were videotaped and rated by observers for social competence. Depressed participants rated themselves as less competent than did non-depressed individuals; however, neither the observers nor partners rated the depressed targets as less competent. Several methodological issues may have contributed to the inconsistency in results from previous studies regarding the degree to which social skills decits or negative self-evaluations contribute to depression. Studies examining competence impressions have, for example, often used outside observer ratings to determine social competence (Borden & Baum, 1987; Gotlib & Meltzer, 1987). Evaluation of social competence by someone who is not part of the interaction may not be analogous to in situ evaluations. Observations by an outsider may not reect the ongoing internal perceptions and subsequent reactions and adjustments of the interactants. In addition, the behavior of people who are being taped or observed may be quite different from their behavior in normal conversations. Therefore, the present study used only the perceptions of the participants as dependent measures. Further, studies of the interactions of depressed people have focused on ratings of communications skills or partners ratings of the impact of brief interactions with depressives on the partners mood. Few studies have evaluated ratings of depressed individuals in terms of more detailed and multidimensional indices of social competence. A broad index of competence provides a more comprehensive framework in which to evaluate behavior, including motivation, knowledge, skills, and context (Spitzberg & Cupach, 1984; Spitzberg & Hurt, 1987). Spitzberg and his associates have developed a model of social competence that includes motivation, knowledge, and skills (altercentrism, composure, expressiveness, and interaction management). A core prediction of the model is that enacting the following skills will increase the probability that a person will be viewed as competent by self and others: (i) showing attention to and concern for the other person, and the display of interest in the other person altercentrism; (ii) the ability to relate in a manner appropriate for the situation composure; (iii) behavioral animation expressiveness; and (iv) management of turns, interruptions, and topic management interaction management (Spitzberg, 1993; Spitzberg, Brookshire, & Brunner, 1990). The Conversational Skills Rating Scale (CSRS; Spitzberg & Hurt, 1987) measures these four indices and a fth index of global competence. Spitzbergs rating scales allow for a more comprehensive and empirically validated focus than ratings of specic skills decits. The present study investigated self and partners rating of social competence following face-to-face interactions between college students classied as depressed/dysphoric and non-depressed according to their scores on the Beck Depression Inventory. Both self and partner ratings of social competence were examined. It was predicted that depressed/dysphoric participants would rate themselves lower on all subscales of Spitzbergs social competence rating scale than non-depressed participants. Second, it was expected that depressed/dys-
142
phoric participants would be rated lower on all CSRS subscales than nondepressed participants by their partners. In addition, because characteristics such as at affect and a tendency to self-focus have been reported to be characteristic of depressed individuals (Ingram, 1990), we predicted that participants would rate their depressed partners lowest on the skill clusters of altercentrism and expressiveness. Finally, participants self-reported mood ratings obtained immediately prior to the interaction were expected to covary with ratings of their partners social skills (i.e., participants in better moods would rate their partners skills higher). Method Participants One hundred and two undergraduate students (58 women, 44 men; M 18.7 years of age, SD 2.3) participated in partial fulllment of an introductory psychology course requirement. Participants were recruited based on their scores on the Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) from an earlier testing session. Students who scored at or above a score of 10 were classied as depressed/dysphoric and students who scored at or below a maximum score of two were classied as nondepressed. These cut points were adopted in accordance with previous recommendations (Gotlib & Meltzer, 1987; Kendall, Hollon, Beck, Hammen, & Ingram, 1987). The BDI is a 21-item questionnaire that assesses cognitive, affective, motivational, and somatic symptoms of depression. Testretest reliability of the BDI is adequate, as is its internal consistency and validity with both clinical and non-clinical samples (Beck, Steer, & Garbin, 1988). The term depressed/dysphoric refers to people who scored above the cutpoint on the depression inventory and is used for ease of discussion. The term is not intended to imply that our participants had been diagnosed as clinically depressed. Prior to arriving at the study, the experimenter randomly assigned participants to one of three dyads based on their pre-session BDI scores: (i) depressed/dysphoric participant with another depressed/dysphoric participant; (ii) depressed/dysphoric participant with a non-depressed participant; or (iii) non-depressed participant with another non-depressed participants. All dyads were same sex. There were 13 depressed participant with another depressed participant dyads (10 female and 3 male); 24 depressed participant with nondepressed participant dyads (10 female and 14 male); and 14 non-depressed participant with non-depressed participant dyads (9 female and 5 male). Procedure Participants were scheduled in groups of six or eight. When they arrived at the lab, they were told that the experiment was about social interaction impressions, a general description of the requirements of the study was given, any questions were answered, and consent forms were distributed and signed. Next, participants rated their present mood on a scale of 0 to 99, using a visual analogue mood barometer rating scale (Pangalila-Ratu-Langi & Jansen, 1988). High scores represented positive mood (labeled happy, elated), whereas lower scores represented negative mood (labeled sad, blue). Participants were next instructed to pair themselves with their pre-determined partner. Dyads were seated in chairs facing each other. Each dyad was
143
seated in separate corners of a large room (approximately 25 40 feet) so that each pairs conversation would not interfere with others. Once dyad partners were assigned, the experimenter questioned the participants in each pair to make sure that they were not previously acquainted with each other. Participants were told to engage in a 15-minute conversation with their partner. They were told to imagine they have just met this person in a social situation and are getting acquainted. The experimenter then left the room while the conversations ensued. Following the 15-minute interaction, one participant from each dyad was led into an adjacent room to complete the follow-up questionnaires and ratings. Participants were told they would have no further contact with their partners during the study and their partners would not be made aware of their responses. Participants then completed another mood barometer identical to the one completed prior to their interactions. Participants next completed the Conversation Skills Rating Scale (CSRS; Spitzberg & Hurt, 1987), rst rating their own skills and then rating their partners skills. Participants were instructed to report only on the previous conversation when completing the CSRS. Finally, the BDI was readministered. The CSRS consists of 25 molecular questions and ve global ratings. The molecular questions comprise the four subscales (altercentrism, composure, expressiveness, and interaction management). Cronbachs alphas were computed for the subscales and each showed adequate reliability: altercentrismself, .78, altercentrismpartner, .84; composureself, .81, composurepartner, .81; expressivenessself, .84, expressivenesspartner, .87; interaction managementself, .74, interaction managementpartner, .75. The global ratings scale also had adequate reliability: globalself, .81, and globalpartner, .82. All items were rated on a scale ranging from 1 to 5, using the following verbal anchors: 1 needs much development, 2 needs some development, 3 satisfactory, 4 good, and 5 excellent. Results Pretest BDI scores were used to classify participants as depressed or nondepressed. Fifty participants scoring 10 or above were labeled depressed/dysphoric, and 52 scoring 2 or below were classied as non-depressed. Of the 50 depressed participants, 30 were women and 20 were men. Dichotomous codes of depressed or not depressed were used in subsequent analyses instead of the continuous scale BDI score because mid-range scores of 39 on the BDI were not represented in the present sample. The mean initial BDI score for the depressed group was 14.26, SD 3.75. Of the 52 non-depressed participants, 28 were female and 24 were male. The mean initial BDI score for the nondepressed group was 0.61, SD 0.75. Initial BDI scores correlated signicantly with the follow-up BDI scores obtained approximately 3 weeks later during the experimental session (r .74, p .001). Although the second BDI score obtained at the end of the experiment reclassied some of the participants, the original classications were used in subsequent analyses. The decision to use pretest BDI scores for analyses was made for two reasons. First, regardless of which BDI score (pretest or experimental) was utilized to determine depression status in data analyses, the direc-
144
tion of the effects was the same and the signicance levels were similar. Second, the experimental BDI was obtained following the social interaction and completion of the other dependent measures. Thus, it was unclear whether the experimental BDI score may have been inuenced by the interaction, suggestion effects from the context of participation in the study, and/or present mood. Preliminary analyses revealed that sex composition of the dyad did not account for variance in social skill rating or moods ratings, neither as a main effect nor in an interaction with any other independent variable. Therefore, subsequent analyses were collapsed across sex. Participants were paired with another participant, and this created the potential for dependencies among scores within dyads. Our dyads were considered mixed (Kenny, Kashy, & Bolger, 1998) because depression status was the same in some dyads and different in other dyads. Therefore, the rst step was to examine potential dependence among the ratings from participants in the same dyad according to techniques discussed by Grifn and Gonzalez (1995). They recommend computing intraclass correlations for both X and Y variables, and they provide signicance tests for these correlations. Within this framework, depression status of self and partner and the mood ratings were considered predictor variables (X) and ratings of self and partner social skills were treated as dependent variables (Y). Within-dyad intraclass correlations for depression status, mood ratings, and ratings of social skill were calculated. The intraclass correlations for the self and partner ratings of social skills were moderate, r .28, t(102) 2.79, and r .37, t(100) 3.66, ps .05, respectively. However, the intraclass correlations for each of the predictor variables were non-signicant (depression status, r .04; mood time 1, r .05; and mood time 2, r .17). Kenny et al. (1998) suggest that, if intraclass correlations in the mixed variables case are less then .50, it is safe to use the person as the unit of analysis. Although none of the intraclass correlations exceeded .50, some of them approached it. Therefore, taking a conservative route, we assumed that our data were non-independent and analyzed them using a multilevel modeling approach. Specically, the data were analyzed using the hierarchical linear modeling (HLM) approach outlined by Bryk and Raudenbush (1992; see also Bryk, Raudenbush, & Congdon, 1996). These techniques were developed to analyze nested data using maximum likelihood estimation procedures. Our sample was represented by a two-level model; participants at the lower level were nested within dyads at the upper-level. A persons outcome variable (e.g., self-ratings of expressiveness) was predicted by equations that accounted for both within dyad variation and between-dyad variation (i.e., the non-independence). For example, the equations used for predicting ratings of ones own expressiveness were as follows: Lower-level: yij b0j b1j(depression) Upper-level: b0j g00 + u0 b1j g10 b2j g20 b2j(partners depression) r
[1]
In the lower-level equation, yij is the expressiveness of the ith person in the jth dyad; b0j is the intercept for the jth dyad (the average score of the dyad); b1j is
145
the effect of depression (a 0 or 1 dummy code); b2j was the effect of partners depression (also a 0 or 1 dummy code); and r was error. The upper-level equations treated b0j as a random effect; it was calculated from the grand mean (g00) plus the unique error (u0) associated with the upper-level grouping variable, the dyad. The effects of depression and partner depression were also calculated based on the sample-wide effects (g10 and g20), but these effects were treated as xed (i.e., there is no error term in these equations). In sum, each persons rating of her own expressiveness was predicted from her depression, her partners depression, and the variance associated with her dyad. Conveniently, because non-depressed was given a 0 dummy code and depressed was given a 1 dummy code, b1 is really the mean difference between the non-depressed participants and depressed participants. Similarly, b2 is the mean difference between having a non-depressed partner and depressed partner. Ratings of own social skills Self-ratings of social skill were examined using the equations described earlier. Each of the ve subscales of the CSRS (altercentrism, composure, expressiveness, management, and global) were predicted from depression status of participant and depression status of partner (depressed or not depressed). Thus, the relationship between depressive status of participant and participants partner with self-ratings of social skills were examined. As can be seen in Table 1, on each of the ve subscales, depression status of the participant contributed to the variance in social skill ratings of the self. Depression status of the partner did not affect social skill ratings of the self. The bs reported on Table 1 are unstandardized coefcients and, because depression status of participant and depression status of partner were coded as 0 and 1, bs are the mean differences between non-depressed and depressed participants. The b of .50 listed in the row marked Participants Depression and in the column marked Global indicates that Depressed participants rated their Global social skill, a mean of .50 points (on the 5-point scale) below nondepressed participants. None of the coefcients representing the effects of partners depression was signicant. Non-depressed participants rated their social skills higher than depressed participants on all ve subscales on the CSRS, regardless of the depression status of their partners. Ratings of partners social skills Ratings of partners social skills were also analyzed with HLM procedures identical to those used in examining self-ratings of social skill described above (see Equation 1). Depression status of the participant (the rater) and depression status of the interaction partner (the one being rated) predicted ratings of the partners social skill. Again, each of the subscales of the CSRS was analyzed in separate procedures. Results are presented in Table 2. Depression status of the participant uniquely accounted for variance in ratings of their partners social skills on each of the ve subscales of the CSRS. Depressed participants rated their partners lower than non-depressed participants rated their partners. However, depression status of the partner (i.e., the person being rated) did not account for a signicant amount of the variance on any of the social skills measures. Pre-interaction mood effects Transitory mood was also measured prior to the 15-minute social interaction
TABLE 1 Self ratings of social skills as a function of depression of participant and partners depression
146
Predictor Altercentrism b Participants depression Partners depression .26 .13 t 2.39* 1.21 non-depressed and 1 Composure b .56 .00 t 4.94*** 1 depressed. b Expressiveness b .52 .14 t 4.85*** 1.29 Management b .45 .11 t 4.15*** 1.05 Global b .50 .03 t 4.57*** 1
TABLE 2 Ratings of partners social skills as a function of depression of participant and partners depression
Predictor Altercentrism b Participants depression Partners depression .25 .18 t 2.14* 1.58 non-depressed and 1 Composure b .30 .13 t 2.76** 1.20 depressed. b Expressiveness b .30 .17 t 2.62* 1.50 Management b .32 .07 t 3.11** 1 Global b .23 .16 t 2.21* 1.48
TABLE 3 Predicting self and partner ratings of social skills from pre-interaction mood of participant
Predictor Altercentrism b Ratings of self Participants mood Ratings of partner Participants mood .010 .003 t 3.43** 1.07 Composure b .017 .008 t 5.25*** 2.50* Expressiveness b .015 .008 t 5.25*** 2.67* Management b .014 .005 t 4.81*** 1.60 Global b .015 .004 t 4.65*** 1.30
147
task. The HLM procedures were repeated, only this time self-ratings of social skills and partner ratings of social skills were predicted from pre-interaction mood. The results of these analyses are presented in Table 3. Mood signicantly predicted self-ratings of each of the ve subscales of social skill. People in a better mood rated their skills higher than those in a worse mood. To determine if mood had an effect on self-ratings of social competence, above and beyond depression status, the models were reanalyzed, including the depression status of participant in the equation. For each subscale on the selfratings, the mood effect remained signicant. This indicates that mood of the participant inuenced ratings of their own social skills, above and beyond depression status effects. Pre-interaction mood also predicted ratings of partners composure and partners expressiveness. People in a better mood rated their partners composure and expressiveness higher than people in a worse mood. However, when depression status was also entered into the equation as a predictor, only the pre-interaction mood effect on expressiveness remained marginally signicant. Thus, pre-interaction mood had some association with ratings of partners social skill (composure and expressiveness), but these effects accounted for little variance above and beyond the effect of depression status. Post-interaction mood change Participants also reported their mood after the interaction. Traditional hierarchical multiple regression was used to analyze the relationship between depression scores and pre- to post-interaction mood change. Post-interaction mood scores were regressed onto pre-interaction mood scores at step 1 of the regression procedure. Mood at time 1 accounted for a large portion of the variance of mood at time 2, R2 .69, p .001. Being in a better mood prior to the interaction was associated with being in a better mood following the interaction. Depression status of the participant and depression status of partner were entered as a set in step 2 of the regression procedure. These two variables did not signicantly account for any additional variance in mood at time 2. Overall, participants reported an increase in mood from pre- to post-interaction (means 69.6 (SD 17.9) and 74.21 (SD 15.8), respectively). However, depression status of self or partner did not relate to mood change. Total effects of depression, depression of partner, and pre-interaction mood Finally, we were interested in determining how much of the variance in ratings of social skill that depression status of participant, depression status of partner, and mood accounted for. We used traditional OLS multiple regression for this, regressing social skill rating onto depression status, partners depression status, and pre-interaction mood of the participant. In order to calculate percentage of variance accounted for by our predictor variables (R 2), it was necessary to treat each participant as an independent case. For the self-ratings of social skill, the three predictor variables accounted for 16% of the variance in the altercentrism scale, 30% of the variance in the composure subscale, 32% of the variance in the expressiveness subscale, 28% of the variance in the management subscale, and 26% of the variance in the global skills subscale. For the ratings of partners social skill, the three predictor variables accounted for 9 percent of the variance in altercentrism, 11% of the variance in composure, 14% of the variance in expressiveness, 12% of the variance in management,
148
and 9% of the variance in global skill. Each of these R 2s was signicant, ps .05. Based on the previous sets of analyses, it should be noted that depression status of the participant and mood accounted for the variance in self-ratings, while depression status of the participant accounted for the variance in ratings of partner. Depression status of the partner did not account for a signicant portion of variance in either ratings of own social skill or ratings of partners social skill. Discussion Depressed/dysphoric participants rated their own and their partners competence signicantly lower than non-depressed participants on all the subscales of the social skills measure: Altercentrism, Composure, Expressiveness, Interaction Management, and Global Skills. This negative bias was not inuenced by the depression status of the partners. That is, depression level of partners did not affect peoples ratings of themselves or of their partners. These results are consistent with previous ndings indicating that negative self-perceptions rather than actual skills decits are most characteristic of depressed students (Gotlib & Meltzer, 1987). These ndings expand on previous research by using an index of social competence that includes components of motivation, knowledge, skills, and context (Spitzberg & Cupach, 1984; Spitzberg & Hurt, 1987). In this study, the depressive symptoms of the rater were more powerful determinants of ratings of self and partner social competence than the affect and behaviors exhibited by the person being rated. This nding is consistent with those of Dobson (1989), who studied perceptions of interpersonal impact among depressed, anxious, and normal participants. Results indicated that, although the individuals in three groups were responded to by partners in equal ways, individuals in the anxious and depressed groups believed that they were rejected more by the partners. Depression/dysphoria appears to generate a bias that extends to perceptions of both self and others. Perhaps depression contributes to a failure to recognize the positive responses and competencies of others in social situations because interactions with others are viewed as unrewarding. Likewise, positive mood may result in a bias toward positive attributions about competencies, independent of what actually is experienced. Transitory mood state at the time of the interaction also covaried with selfratings of social skill. People in a bad mood rated their own skills during the interaction less positively than people who were in a good mood. However, when evaluating the social skills of their partners, peoples transitory moods did not inuence their ratings, except for those skills associated with expressiveness. These results indicate that, regardless of risk for depression, transitory moods inuence perceptions of the self, but not others; by contrast, depression inuences evaluations of both self and other. This nding points to the uniqueness of depression/dysphoria (as opposed to mood) in inuencing perceptions of others. This evidence may relate to the nding that depressed/dysphoric students interact as often as non-depressed students, but report less enjoyment and satisfaction with their social lives (Nezlek et al., 1994). Interestingly, both depressed and non-depressed participants reported an improvement in mood after their interactions. Further, interacting with a depressed participant did not mediate mood change. That is, people who inter-
149
acted with a depressed participant reported the same mood increases as those who interacted with non-depressed participants. Future research is necessary to explain why, despite rating themselves and others as lower in social skills, depressed/dysphoric participants experienced positive mood changes following the interaction. Perhaps, as Rook (1998) suggests, positive aspects of social exchange predict different outcomes than negative aspects of social exchange. That is, depression may predict evaluations of competence, while other variables may be inuencing post-interaction mood. There are several possible limitations to the generalizability of this study. First, participants were drawn from a college population, and the generalizability of these ndings to a clinically depressed population remains in question. Second, participants engaged in 15-minute conversations with strangers. This interaction may not have been long enough for perceptual biases in the assessment of partners social competence to be evidenced. Hokanson and Butler (1992) reported that room-mates of depressed college students avoided them and exhibited increasing levels of hostility over the course of a semester. It may be that extended obligatory interactions allow for the expression of cognitive biases that are not evidenced after brief interactions.
REFERENCES Beck, A. T. (1972). Depression: Causes and treatment. Philadelphia: University of Pennsylvania Press. Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-ve years of evaluation. Clinical Psychology Review, 8, 77100. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561571. Borden, J. W., & Baum, C. G. (1987). Investigation of a social-interactional model of depression with mildly depressed males and females. Sex Roles, 17, 449-465. Brown, G. W., & Harris, T. (1978). Social origins of depression: A study of psychiatric disorder in women. New York: Macmillan. Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models. Newbury Park, CA: Sage. Bryk, A. S., Raudenbush, S. W., & Congdon, R. T. (1996) HLM: Hierarchical linear and nonlinear modeling with the HLM/2L and HLM3/L Programs. Chicago: Scientic Software International. Burchill, S. A. L., & Stiles, W. B. (1988). Interaction of depressed college students with their roommates: Not necessarily negative. Journal of Personality and Social Psychology, 55, 410-419. Coyne, J. C. (1976). Depression and the response of others. Journal of Abnormal Psychology, 85, 186-193. Coyne, J. C. (1990). Interpersonal processes in depression. In G. I. Keitner (Ed.), Depression and families: Impact and treatment. (pp. 3154). Washington, DC: American Psychiatric Press. Dobson, K. S. (1989). Real and perceived responses to subclinically anxious and depressed targets. Cognitive Therapy and Research, 13, 37-47. Dykman, B. M., Horowitz, L., Abramson, L. Y., & Usher, M. (1991). Schematic and situational determinants of depressed and non-depressed students interpretation of feedback. Journal of Abnormal Psychology, 100, 4555. Gotlib, I. H., & Meltzer, S. J. (1987). Depression and the perception of social skill in dyadic interaction. Cognitive Therapy and Research, 11, 41-54.
150
Grifn, D., & Gonzalez, R. (1995). Correlational analysis of dyad-level data in the exchangeable case. Psychological Bulletin, 118, 430-439. Hammen, C. (1997). Depression. Bristol, PA: Taylor and Francis/Psychology Press. Hokanson, J. E., & Butler, A. C. (1992). Cluster analysis of depressed college students social behaviors. Journal of Personality and Social Psychology, 62, 273-280. Ingram, R. E. (1990). Self-focused attention in clinical disorders: Review and a conceptual model. Psychological Bulletin, 107, 156-176. Joiner, T. E., Alfano, M., & Metalsky, G. I. (1992). When depression breeds contempt: Reassurance seeking, self esteem, and rejection of depressed college students by their roommates. Journal of Abnormal Psychology, 101, 165173. Joiner, T., & Coyne, J. C. (Eds.). (1999). The interactional nature of depression. Washington, DC: American Psychological Association. Kendall, P. C., Hollon, S., Beck, A. T., Hammen, C., & Ingram, R. (1987). Issues and recommendations regarding the use of the Beck Depression Inventory. Cognitive Therapy and Research, 11, 289299. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (Vol. 1, 4th ed., pp. 233-265). New York: McGraw-Hill. Lewinsohn, P. M. (1974). A behavioral approach. In R. M. Friedman & M. M. Katz (Eds.), The psychology of depression: Contemporary theory and research. New York: Wiley. Lewinsohn, P. M., Mischel, W., Chaplin, W., & Barton, R. (1980). Social competence and depression: The role of illusory self-perceptions. Journal of Abnormal Psychology, 89, 203212. Nezlek, J. B., Imbrie, M., & Shean, G. D. (1994). Depression and everyday social interaction. Journal of Personality and Social Psychology, 67, 1101-1111. Pangalila-Ratu-Langi, E. A., & Jansen, A. A. (1988). Ritanserin in the treatment of generalized anxiety disorders: A placebo controlled trial. Human Psychopharmacology: Clinical and Experimental, 3, 207-212. Rook, K. S. (1998). Investigating the positive and negative sides of personal relationships: Through a lens darkly? In B. H. Spitzberg & W. R. Cupach (Eds.), The dark side of close relationships (pp. 369393). Mahwah, NJ: Erlbaum. Segrin, C. (1990). A meta-analytic review of social skill decits in depression. Communication Monographs, 57, 292-308. Segrin, C., & Abramson, L.Y. (1994). Negative reactions to depressive behaviors: A communication theories analysis. Journal of Abnormal Psychology, 103, 655668. Segrin, C., & Dillard, J. P. (1992). The interactional theory of depression: A meta-analysis of the research literature. Journal of Social and Clinical Psychology, 11, 43-70. Spitzberg, B. H. (1993). The dialectics of (in)competence. Journal of Social and Personal Relationships, 10, 137-158. Spitzberg, B. H., Brookshire, R. G., & Brunner, C. (1990). The factorial domain of interpersonal skills. Social Behavior and Personality, 18, 137-150. Spitzberg, B. H., & Cupach, W. R. (1984). Interpersonal communication competence. Beverly Hills, CA: Sage. Spitzberg, B. H., & Hurt, T. H. (1987). The measurement of interpersonal skills in instructional contexts. Communication Education, 36, 29-45.