An interactional test of the reformulated helplessness theory of depression in women receiving clinical treatment for eating disorders Ken J. Rotenberg a, ⁎, Paula Costa a , Mark Trueman a , Paul Lattimore b a School of Psychology, Keele University, Keele, Newcastle-Under-Lyme, Staffordshire, ST5 5BG UK b School of Natural Science; Psychology, Liverpool John Moores University, Liverpool L3 2ET UK abstract article info Article history: Received 28 September 2011 Accepted 1 March 2012 Available online 8 March 2012 Keywords: Helplessness Attributional style Negative life events Interactions Eating disorder treatment The study tested the Reformulated Helplessness model that individuals who show combined internal locus of control, high stability and high globality attributions for negative life events are prone to depression. Thirty- six women (M= 29 years -8 months of age) receiving clinical treatment for eating disorders completed: the Attribution Style Questionnaire, the Beck Depression Inventory, and the Stirling Eating Disorder Scales. An HRA yielded a three-way interaction among the attributional dimensions on depressive symptoms. Plotting of the slopes showed that the attribution of negative life events to the combination of internal locus of con- trol, high stability, and a high globality, was associated with the optimal level of depressive symptoms. The findings supported the Reformulated Helplessness as a model of depression. © 2012 Elsevier Ltd. All rights reserved. According to the reformulated learned helplessness (RHT) model (Abramson, Seligman, & Tesdale, 1978; Peterson & Seligman, 1984), individuals are prone to depression if they characteristically adopt a pessimistic attribution style primarily by attributing negative life events to internal, stable, and global causes. The substantial body of research on RHT has summed across the three attributional dimen- sions (locus of control, stability, and globality) for negative life events to yield a composite measure of pessimistic attributional style. As sup- port for the RHT, researchers have found that that composite measure is associated with depressive symptoms as a medium effect in meta- analyses (Sweeney, Anderson, & Bailey, 1986; for a recent individual study see Reilly-Harrington et al., 2010). Also as full or partial support for RHT, researchers (e.g., Golin, Sweeney, & Shaeffer, 1981) have found that all or some of the attributional dimensions (locus of control, stability, and globality) for negative life events uniquely pre- dict depressive symptoms. The relations with the composite scores and individual attribution al dimensions have been found in samples with women with eating problems (e.g., Dalgleish et al., 2001; Mansfield & Wade, 2000) and provide some insight into the attribution- al styles that potentially contribute to their depressive symptoms. The research findings yield equivocal support for the HFT model of depression, however. Regarding the composite measure of attribu- tional styles, it is the case that highly elevated score on any of the three attributional dimensions could be responsible for the observed relation between the composite measure and depressive symptoms. Similar problems prevail regarding the observation that individual attributional styles uniquely predict depressive symptoms. For exam- ple, elevated stability attributions for negative life events by a sample of individuals could contribute to their depressive symptoms but they might be disinclined to attribute those events to global causes: thus not confirming the HFT model of depression. Based on the aforementioned reasons, the current study was guid- ed by the premise that an adequate test the HFT of depression re- quires finding that a three-way interaction among the attributional dimensions statistically predicts depression symptoms. The interac- tion would comprise one in which optimal depressive symptoms are shown by individuals who attribute negative life events to the combination of internal locus of control, high stability and high glob- ality. The current study was designed to examine whether or not women receiving clinical treatment for eating disorders would show the HFT expected interactions among the attributional dimensions and thus show the combinations of attributional dimensions that sta- tistically account for their depressive symptoms. 1. Method 1.1. Participants The participants were 36 women with a mean age of 29 years -8 months (SD = 10 years and 9 months, and ranging from 17 to 66 years of age). They were receiving clinical treatment for eating dis- orders in a major city in the United Kingdom. The study was carried out in accordance with the British Psychological Society and NHS eth- ical guidelines. Eating Behaviors 13 (2012) 264–266 ⁎ Corresponding author. Tel.: + 44 1782 733046; fax: + 44 1782 583387. E-mail addresses: k.j.rotenberg@keele.ac.uk (K.J. Rotenberg), paulacosta78@hotmail.com (P. Costa), m.trueman@psy.keele.ac.uk (M. Trueman), P.J.Lattimore@ljmu.ac.uk (P. Lattimore). 1471-0153/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.eatbeh.2012.03.001 Contents lists available at SciVerse ScienceDirect Eating Behaviors