When Will Bulimics Be Depressed and When Not ...

Cognitive Therapy and Research, Vol. 21, No. 1, 1997, pp. 61-72

When Will Bulimics Be Depressed and When Not? The Moderating Role of Attributional Style1 Gerald I. Metalsky2 Lawrence University

Thomas E. Joiner, Jr.2 University of Texas Medical Branch at Galveston

Stephen A. Wonderlich University of North Dakota School of Medicine

William W. Beatty North Dakota State University

R. Dennis Staton University of North Dakota School of Medicine

Janice A. Blalock University of Texas Medical Branch at Galveston

We examined whether attributional style would moderate whether or not bulimic patients exhibited depressive symptoms. As predicted, clinical bulimics 1First

authorship is jointly shared by Thomas Joiner, Steve Wonderlich, and Jerry Metalsky, who contributed equally to the study. Preparation of this article was supported, in part, by a Public Health Serice Biomedical Research Support Grant (2 S07 RR05407-28) to S. A. Wonderlich; by a Young Investigator Award to Thomas Joiner from the National Alliance for Research on Schizophrenia and Affective Disorders (NARSAD); and by grants from the Hogg Foundation for Mental Health and Lawrence University to G. I. Metalsky. 2Address all correspondence concerning this article to Thomas Joiner, 3.102 Graves Building, D-2S, Department of Psychiatry and Behavioral Sciences, University of Texas Medical Branch at Galveston, Galveston, Texas 77555-0425, or to Jerry Metalsky, Department of Psychology, Lawrence University, Box 599, Appleton, Wisconsin 54912-0599. 61 0147-591697/0200-0061$12.50/0 C 1997 Plenum Publishing Corporation

62

Metalsky, Joiner, Wonderiich, Beatty, Staton, and Blalock

with a negative attributional style exhibited depressed symptoms whereas clinical bulimics without a negative attributional style did not. Negative atnibutional style bulimilcs fell in the severe range of symptom severity while bulimics without the negative style fell in the normal range of severity. Additionally, attributional style moderated severity of depressed symptoms just as much in clinical bulimics as in clinical depressives. Implications for future work are discussed, including how the cognitive theories of depression may further advance our understanding of depressed symptoms among bulimics. KEY WORDS: attributional style; bulimia; depressed symptoms.

An association between bulimia nervosa and depressed symptoms has been well established (Hudson, Laffer, & Pope, 1982; Strober & Katz, 1988). Although this has prompted investigation of cognitive variables that may help account for depressive symptoms among bulimics (e.g., negative body attitudes: Clark, Feldman, & Channon, 1989; irrational cognitions: Ruderman, 1986; cognitive errors: Strauss & Ryan, 1988), the vulnerability factors posited by the cognitive theories of depression (Abramson, Metalsky, & Alloy, 1989; Abramson, Seligman, & Teasdale, 1978; Beck, 1967; Beck, Rush, Shaw, & Emery, 1979) have received surprisingly little empirical attention in this line of research. Two recent studies represent important exceptions. First, Goebel, Spalthoff, Schulze, and Florin (1989) reported that a negative attributional style was highly prevalent among women with Diagnostic and Statistical Manual of Mental Disorders (3rd ed.) (DMS-III; American Psychiatric Association, 1980) diagnoses of bulimia. Second Joiner, Metalsky, and Wonderlich (1995) found that undergraduate women with bulimic symptoms and a negative attributional style were more vulnerable to depressed symptoms than women with bulimic symptoms but no negative attributional style. Hinz and Williamson (1987) reported that the incidence of a current mood disorder among bulimics is 24% to 33% [by Schedule for Affective Disorders and Schizophrenia (SADS; Endicott & Spitzer, 1978) and Research Diagnostic Criteria (RDC; Feighner et al., 1972) criteria], which suggests that many but not all bulimics experience depression. Consistent with Joiner et al. (1995), we suggest that helplessness/hopelessness theory (Abramson et al., 1978, 1989) may explain which bulimics experience depression and which do not. We propose that an internal, stable, global attributional style for negative events places bulimics at risk for depression, whereas an external, unstable, specific attributional style for negative events buffers them against depression.

Attributional Style, Bulimia, Depression

63

In line with this proposal, the purpose of the present investigation was to examine whether the negative attributional style featured in helplessness/hopelessness theory would moderate the relationship between clinical bulimia and depressed symptoms. We hypothesized that attributional style would predict which bulimics display depressed symptoms, such that clinical bulimics with a negative attributional style would report more depressive symptoms than clinical bulimics without a negative attributional style. Furthermore, we examined whether attributional style's modulating role operates similarly among clinical depressives as among clinical bulimics. If so, we would argue that attributional style not only modulates which bulimics experience depression, but further, that it does so to a similar degree as among depressives. It should be noted that our study represents the first to apply the hopelessness theory to a combined sample of clinical bulimics and clinical depressives.

METHOD Participants and Procedure Overall, 36 participants between the ages of 18 and 61 took part in the study. Twenty-two met diagnostic criteria for bulimia, and 14 for depression. Bulimic Participants. Participants were consecutive adult admissions to the Fargo Clinic Eating Disorders Program (average age = 22.27; years SD = 5.70; all were female). All bulimic participants met Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.) (DSM-III-R; American Psychiatric Association, 1987) criteria for bulimia nervosa and were either starting treatment or in the initial phase of the treatment program. Consistent with previous work (e.g., Mitchell et al., 1990), participants were required to have an ideal body weight of 80% to 120% for inclusion in the study. Initial clinical diagnosis was made by Ph.D.-level clinical psychologist and a masters level psychiatric nurse at the Fargo Clinic Eating Disorders Program following a comprehensive psychological evaluation, which included assessment of attributional style and depression. Subjects who met DSM-III-R criteria for bulimia nervosa at initial evaluation, and who indicated an interest in volunteering for the study, were again interviewed by a trained research assistant (blind to diagnosis) with a structured diagnostic interview for bulimia, based on DSM-III-R criteria, to verify their diagnoses. In all cases, the initial diagnosis of bulimia was confirmed. Subjects were given $25.00 for their involvement in the project.

64

Metalsky, Joiner, Wonderlich, Beatty, Staton, and Blalock

Depressed Participants. Fourteen depressed participants were recruited from the outpatient clinics of a large academic medical center in the southwestern U. S. (average age = 42.77 years; SD = 11.43; 11 females, 3 males). It should be noted that the depressed participants were, on average, older than the bulimic participants. As we demonstrate below, however, when age was used as a covariate throughout our analyses, the pattern of findings was unaffected. The depressed participants met DSM-III-R criteria for chronic major depression (n = 8) or dysthymia (n =6), as determined by Structured Clinical Interview for DSM-III-R (SCID; Spitzer, Williams, Gibbon, & First, 1988). SCID interviews were conducted by masters-level psychological associates, who were rigorously trained in SCID administration, and who were supervised by a Ph.D. clinical psychologist. All diagnoses of depression were verified in second interviews by a psychiatrist. Assessment of attributional style and depression were conducted in conjunction with initial interviews. Laboratory tests were performed on all subjects to rule out medical explanations (e.g., hypothyroidism) for depressed symptoms. Participants were paid $50 for taking part in the study. Those who experienced current medical disorders or another primary psychiatric diagnosis were not included in the depressed group, nor were those with a history of bipolar disorder, obsessive-compulsive disorder, or organic mental syndrome or disorder. Measures Beck Depression Inventory (BDI). Depression level was assessed by the BDI (Beck et al., 1979), a 21-item self-report inventory of depressive symptoms. The BDI has yielded adequate reliability estimates, with a mean coefficient alpha of .86 for psychiatric patients and .81 for nonpsychiatric subjects (Beck, Steer, & Garbin, 1988). The BDI also has been well validated (see Beck et al., 1988, for a review). Several of the items of the BDI overlap in content with those often experienced by eating-disordered patients. To avoid tautology resulting from this overlap, we deleted the following items from the BDI: Item 5 (feel guilty), 14 (feel ugly or repulsive looking), 16 (sleep problems), 18 (appetite disturbance), and 19 (weight loss). It is important to note, however, that the same pattern of results emerged regardless of whether we used the "adjusted" or full BDI. Accordingly, we focus on the full BDI in reporting our results, to enhance the comparability of our findings with those of other studies. Scores on the BDI can range from 0 to 63. Extended Attributional Style Questionnaire (EASQ). The EASQ (Metalsky, Halberstadt, & Abramson, 1987) consists of 12 hypothetical negative

Attribudonal Style, Bulimia, Depression

65

life events. Similar to the original ASQ (Peterson et al., 1982; Seligman, Abramson, Semmel, & von Baeyer, 1979), subjects write down the one major cause of a given event, in an open-ended format, and then rate the cause on a 1- to 7-point scale separately for degree of internality, stability, and globality. We used the EASQ composite subscale (mean of internality, stability, and globality) for negative outcomes since we did not have differential predictions concerning individual dimensions or different symptoms (Peterson et al., 1982). The EASQ has adequate internal consistency reliability (typically falling between .80 and .90) and, along with the original scale, has been well validated (Metalsky & Joiner, 1992; Metalsky, Joiner, Hardin, & Abramson, 1993; Needless & Abramson, 1990; see Peterson & Seligman, 1984; Sweeney, Anderson, & Bailey, 1986, for reviews of the original scale). For simplicity, we will refer to the composite subscale as EASQ. Scores range from 1 to 7, with high scores reflecting a tendency to attribute negative events to more internal, stable, and global factors.3

RESULTS

Comparisons Between Bulimic and Depressed Participants Although our main focus was on the relation between EASQ and BDI within the bulimic and depressed groups, it is of interest to compare the groups on relevant variables. The groups differed widely in age [t(35) = 7.72,p