Journal of Abnormal Psychology 1977, Vol. 86, No. 6, 66S-668

Anxiety and Perceived Catisation in Social Success and Failure Experiences: Disconfirmation of an Attribution Hypothesis in Two Experiments Hal Arkowitz University of Arizona

William R. Miller University of New Mexico

Two experiments were designed to test hypotheses derived from an attribution model of psychopathology as applied to social anxiety. In both studies, male subjects first interacted with a female confederate who behaved either warmly (success) or coldly (failure) toward them. All subjects then interacted with a second confederate who behaved warmly. In both experiments, it was predicted that high social anxiety would be associated with the internal attribution of social failure and the external attribution of social success. By contrast, it was predicted that low social anxiety would be associated with the internal attribution of social success and the external attribution of social failure. In Experiment 1, patterns of attribution were manipulated in normal subjects, and the effects of the manipulations were examined with respect to their subsequent social anxiety. In Experiment 2, the attributional patterns of high and low socially anxious men were examined in success and failure situations. Neither study provided any support for the hypothesis relating attributional patterns to social anxiety. Several papers have recently attempted speculations and extrapolations from basic attribution research to the areas of psychopathology and behavior change (Kopel & Arkowitz, 197S; Valins & Nisbett, 1971; Winett, 1970). The major conclusion emerging from these reviews is that behavior or behavior change that is attributed to an external agent (e.g., drugs or situational factors) is less likely to persist than is behavior attributed to a stable internal source (e.g., personality traits or dispositions). Valins and Nisbett (1971) reasoned from this that certain attribution patterns could be pathogenic. For example, the attribution of a negative outcome or behavior to stable personal causes (e.g., incompetence or ineptness) may result in the perpetuation of the behavior and associated emotional distress. Further, attributing positive behavior or outcomes to external, serendipitous causes may be equally pathogenic. This pattern of attribution is consistent The research described in this short report was conducted while the authors were at the University of Oregon. A more extensive report of the research is available from the first author. Special appreciation is expressed to Claudette Cummings for her energy and resourcefulness in serving as research coordinator for both studies. Requests for reprints should be sent to William R. Miller, Department of Psychology, University of New Mexico, Albuquerque, New Mexico 87131.

with common conceptualizations of thought processes among depressed individuals (Beck, 1976). The two experiments reported in the present article were designed to provide an empirical test of this attribution model of psychopathology and to extend it to the clinical problem of social anxiety. We were specifically interested in examining the relation between social anxiety and attributions of success and failure in heterosexual interactions. The attribution hypothesis described above would predict that high social anxiety should be associated with the internal attribution of social failures and the external attribution of social successes. By contrast, low social anxiety should be associated with the internal attribution of social successes and the external attribution of social failures. Experiment 1 Method Sixty-four male undergraduates were told that they were participating in a study of social competence in college women. Each was told that he would be interacting with a female subject (actually a confederate). Prior to this first conversation, half of the subjects were led to expect a female who was warm and accepting toward most people (the expect-acceptance condition), and the other half were led to expect a female who was cold and rejecting toward most people (the ex-

665

666

SHORT REPORTS

pect-rejcction condition). Within each of these two groups, half of the subjects actually received a "successful" social interaction (the receiveacceptance condition), and half received a social "failure 1 ' experience (the receive-rejection condition). Groups for whom expectation and outcome were congruent (the expect-acceptance-receiveacceptance condition and the expect-rejectionreceive-rejection condition) were expected to attribute the outcome of the interaction more externally than were subjects in discrepant conditions (the expect-acceptance-receive-rejection condition and the expect-rejection-receive-acceptance condition), where more Internal attributions were anticipated (cf. Weiner et al., 1971). After this instructional set, subjects were told to talk with the female for 5 minutes, as if trying to get to know her during an initial meeting. The female confederates were trained to remain neutral for 1 minute and then to become either increasingly warm (the receive-acceptance condition) or increasingly cold (the receive-rejection condition). Following this first conversation, subjects were told to interact with a second female for 5 minutes. No specific expectation set was given this time, and the second female confederate behaved warmly toward all subjects. The experimental assistants and confederates were unaware of the hypotheses of the study. Prior to receiving instructions, subjects completed the Social Avoidance and Distress (SAD) Scale (Watson & Friend, 1969) and Rotter's (1966) Internal-External (I-E) Scale. Both conversations were observed by trained undergraduates from behind a one-way mirror. The observers were trained in the use of a modified timed behavioral checklist of overt anxiety-related behaviors (Miller, 1973; Paul, 1966). Conversations were also tape recorded and later coded for speech disfiuencies indicative of anxiety (Mahl, 1956; Murray, 1971). All observational data were coded independently by pairs of observers who were blind to subjects' group membership and the main hypotheses of the study. Following each conversation, subjects rated their anxiety and indicated their attribution of the outcome on sevenpoint bipolar attribution scales (Feather, 1969; Miller, 1973). Results The four experimental groups did not differ in initial scores on the SAD and I-E scales. Observational checks on the behavior of confederates showed significant differences between accepting and rejecting females during the first conversa-

tion, supporting the validity of our manipulation. No such differences were observed during the second conversation, again consistent with our intended manipulation. The effectiveness of our attribution manipulation was less clear. Instead of the predicted pattern of attribution, main effects of expectation and outcome, F(l, 60) =4.52, p < .05 (in both cases), were observed; so that during Conversation 1, expect-acceptance groups attributed more externally than did expect-rejection groups, and receive-rejection groups attributed more externally than did receive-acceptance groups. During Conversation 2, however, the expected interaction effect was obtained, falling just short of statistical significance, F(l, 60) =3.82, /> < .06, so that subjects in the expect-rejectionreceive-acceptance and the expect-acceptancereceive-rejection conditions attributed more internally than subjects in the other two groups. A principal-components analysis of all anxiety measures produced no recognizable configuration of factor score coefficients. Anxiety measures were thus analyzed separately by domains: behavioral measures, verbal measures, and selfreport measures. Two-way analyses of variance on these individual anxiety components failed to reveal the predicted pattern of effects of attribution on anxiety for either the first or second conversation. In addition, internal analyses were conducted in which the sample was subdivided on the basis of obtained rather than predicted attribution scores. Once again, no support for the original hypothesis was found. Experiment 2 Having failed to find a relation between attribution and anxiety within a normal population, we attempted a more direct test of the attribution hypothesis by repeating our study using extreme groups. We selected high and and low socially anxious males as measured by the SAD scale. We were interested in determining whether high and low socially anxious subjects would have different patterns of attribution for social success and failure and, if so, whether such patterns would be associated with different levels of anxiety. We predicted that high socially anxious subjects would attribute failure more internally and success more externally and that low socially anxious subjects would show the opposite pattern. Method The design of Experiment 2 was identical to that of Experiment 1 with the following exceptions: (a) Subjects were selected on the basis of

SHORT REPORTS social anxiety, (b) attribution was treated as a dependent variable, and (c) we abandoned the problematic behavioral measures of anxiety and instead monitored subjects' heart rate throughout both conversations. Sixty-four undergraduate males were used, 32 with high social anxiety (HSA) scores (SAD mean score =18.21) and 32 with low social anxiety (LSA) scores (SAD mean score =.71). Half of each group encountered an accepting female ( + ) during the first conversation, and the other half talked with a rejecting (—) female. This created four groups: HSA+, HSA—, LSA+, and LSA—. All groups again encountered an accepting female during the second conversation. No attempt was made to manipulate expectations of subjects about confederates. All assistants and confederates were again blind to the hypotheses. Subject's heart rate was recorded before, during, and after both conversations by means of a photoplethysmograph attached to the little finger of the left hand and monitored by a polygraph. Self-report measures of anxiety and attribution were again obtained after each conversation. Res-nits

As in Experiment 1, observations of confederate behaviors supported the effectiveness of the success/failure manipulation. Consistent with the selection rationale, HSA subjects reported greater anxiety than did LSA subjects, F(l, 60) = 10.18, p < .005, but only during the first conversation. There were, however, no differences between HSA and LSA subjects with regard to heart rate at any point during the experiment. Support for the main predicted effect would emerge from this design as an interaction effect between social anxiety and the success/failure condition, so that HSA+ and LSA— groups should attribute more externally than HSA— and LSA+ groups. A two-way analysis of variance failed to produce this significant interaction effect. Thus, once again, the attribution hypothesis was not supported. In one final attempt, we reanalyzed the data by splitting the groups on actual attributions of the conversations and examining anxiety measures. The main hypotheses would predict that subjects who attributed success externally or failure internally should be more anxious than subjects who showed the reverse pattern. Still no support for the hypothesis was found. Instead, external attribution of any event (success or failure) was found to be related to higher self-reports of anxiety and lower heart rates.

667

General Discussion Neither of the two experiments provided any support for the hypothesis relating attribution for success and failure to social anxiety. In both experiments, the effectiveness of the success/ failure manipulation was strongly supported. Our findings are consistent with those of other recent studies indicating that people in general tend to attribute success internally and failure externally (e.g., Luginbuhl, Crowe, & Kahan, 1975; Sobel, 1974). We examined our original hypothesis from several sides, treating attribution as both an independent and a dependent variable and using both normal and extreme populations. Not even indirect post hoc analyses provided any support for the main hypothesis. The theoretical evidence usually cited in relation to an attributional model of psychopathology is primarily derived from studies on the attribution (or misattribution) of emotional arousal to internal or external sources. Direct tests of the effects of attribution on measured arousal have been few. Neither of our studies support this notion. In addition, a recent study by Singerman, Borkovec, and Baron (1976) found precisely the opposite pattern from that predicted by attribution theory. One critical issue raised by our two experiments is the striking lack of intercorrelation found among anxiety measures. Self-report measures were found to be negatively correlated with behavioral measures. Intercorrelations across domains (e.g., self-report, verbal, and physiological) were consistently low. These findings are consistent with previous research (Lang, 1969; Martin, 1961) and further question the utility of any unitary construct of "anxiety." This problem is likely to plague future researchers in this area who use more than one "anxiety" measure. Failures to support or replicate are never conclusive. We do wish, however, to caution against premature acceptance of attributional formulations of psychopathology and behavior change. Two experiments, carefully conducted by two believers in an attributional model, failed to provide supporting evidence. We choose to attribute this failure externally (to the limitations of the hypothesis) rather than internally (to our design or procedures). References Beck, A. T. Cognitive therapy and the emotional disorders. New York: International Universities Press, 1976. Feather, N. T. Attribution of responsibility and

668

SHORT REPORTS

valence of success and failure in relation to initial confidence and task performance. Journal of Personality and Social Psychology, 1969, 13, 129-144. Kopel, S., & Arkowitz, H. The role of attribution and self-perception in behavior change: Implications for behavior therapy. Genetic, Psychology Monographs, 1975, 92, 175-212. Lang, P. J. The mechanics of desensitization and the laboratory study of human fear. In C. M. Franks (Ed,), Behavior therapy: Appraisal and status. New York: McGraw-Hill, 1969. Luginbuhl, J., Crowe, D., & Kahan, J. Causal attributions for success and failure. Journal of Personality and Social Psychology, 1975, 31, 86-93. Mahl, G. F. Disturbances and silences in the patient's speech in psychotherapy. Journal of Abnormal and Social Psychology, 1956, 53, 1-15. Martin, B. The assessment of anxiety by physiological behavioral measures. Psychological Bulletin, 1961, 58, 234-255. Miller, W. R. Social anxiety and attributions of success and failure in heterosexual interactions. Unpublished master's thesis, University of Oregon, 1973. Murray, D. C. Talk, silence, and anxiety. Psychological Bulletin, 1971, 75, 244-260. Paul, G. L. Insight vs. desensitization in psycho-

therapy: An experiment in anxiety reduction, Stanford, Calif.: Stanford University Press, 1966. Rotter, J. B. Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 1966, 80(1, Whole No. 609). Singerman, K. J., Borkovec, T. D., & Baron, R. S. Failure of a "misattribution therapy" manipulation with a clinically relevant target behavior. Behavior Therapy, 1976, 7, 306-313. Sobel, R. S. The effects of success, failure, and locus of control on postperformance attribution of responsibility. Journal of General Psychology, 1974, 91, 29-34. Valins, S., & Nisbett, R. E. Attribution processes in the development and treatment of emotional disorders. New York: General Learning Press, 1971. Watson, D., & Friend, R. Measurement of socialevaluative anxiety. Journal of Cons-idling and Clinical Psychology, 1969, 33, 448-457. Weincr, B., Frieze, I., Kukla, A., Reed, L., Rest, S., & Rosenbaum, R. Perceiving the causes of success and failure. New York: General Learning Press, 1971. Winett, R. A. Attribution of attitude and behavior change and its relevance to behavior therapy. Psychological Record, 1970, 20, 17-32. Received April 14, 1977 H

Anxiety and perceived causation in social success and failure experiences: disconfirmation of an attribution hypothesis in two experiments.

Journal of Abnormal Psychology 1977, Vol. 86, No. 6, 66S-668 Anxiety and Perceived Catisation in Social Success and Failure Experiences: Disconfirmat...
300KB Sizes 0 Downloads 0 Views