Fear- and disgust-related covariation bias and eating disorders symptoms in healthy young women Birgit Mayer * , Peter Muris, Maaike Wilschut Institute of Psychology, Erasmus University Rotterdam, The Netherlands article info Article history: Received 24 September 2009 Received in revised form 30 June 2010 Accepted 6 September 2010 Keywords: Eating disorders symptoms Obesity Fear Disgust Covariation bias abstract Covariation bias refers to the phenomenon of overestimating the contingency between certain stimuli and negative outcomes, which is considered as a heuristic playing a role in the maintenance of certain types of psychopathology. In the present study, an attempt was made to investigate covariation bias within the context of eating pathology. In a sample of 61 female undergraduates, a priori and a posteriori contingencies were measured between pictures of obese and slim bodies, on the one hand, and fear- or disgust-relevant outcomes, on the other hand. Results indicated that participants in general displayed an a priori and an a posteriori covariation bias reflecting an overestimation of the link between obese bodies and disgust-relevant outcomes. However, this bias was not related to eating disorder symptomatology. Meanwhile, eating pathology was positively associated with a priori covariation biases referring to the associations between obese bodies and fear-relevant outcomes, and between slim bodies and disgust- relevant outcomes. All in all, these findings suggest that covariation bias plays a role in eating pathology. Ó 2010 Elsevier Ltd. All rights reserved. 1. Introduction Since the cognitive revolution in clinical psychology, it is generally assumed that cognitive biases play an important role in the etiology and maintenance of psychopathology (Harvey, Watkins, Mansell, & Shafran, 2004). For instance, it has been demonstrated that people not always deploy their cognitive abili- ties for the purpose of logical reasoning but rather comply with personal goals and motives (Evans, Newstead, & Byrne, 1993). This means that everyday rationality is characterized by various reasoning errors, and this is particularly true for individuals who suffer from a psychological disorder. A nice example of such dis- torted reasoning can be observed in patients with anxiety disorders who tend to display a so-called covariation bias, which refers to the tendency to overestimate the contingency between certain stimuli and negative outcomes (Chapman & Chapman, 1967). The experi- mental demonstration of this bias in anxiety disordered individuals is straightforward (see Mineka & Tomarken, 1989). Fearful and non-fearful participants are shown a series of slides consisting of fear-relevant (e.g., spiders) and neutral (e.g., flowers) pictures. Each picture type is equally often followed by one of three outcomes, namely an aversive electric shock (i.e., negative outcome), a tone (i.e., neutral outcome), or nothing. After the series of slides, participants are asked to estimate the contingencies between slides and outcomes, and in general it is found that fearful participants genuinely believe that fear-relevant stimuli are more frequently accompanied by negative outcomes (e.g., De Jong, Merckelbach, Arntz, & Nijman, 1992; Pauli, Montaya, & Martz, 1996). In a series of experiments, Tomarken, Sutton, and Mineka (1995) have convincingly demonstrated that this overestimation of co-occur- rence of fear-relevant stimuli and negative outcomes is indeed primarily based on their affective linkages of shared (evolutionary- related) aversiveness and not on other stimulus-outcome similar- ities, like a strong semantic, associative or temporal linkage (e.g., pictures of damaged electrical outlets and shock). Thus, fear is found to be linked to a style of confirmatory processing of certain particular fear-relevant stimuli with certain specific affectively- linked outcomes, which makes that it serves the maintenance or enhancement of fear (see also Tomarken, Mineka, & Cook, 1989). This overestimation of the contingency of fear-relevant stimuli and negative outcomes has already been demonstrated prior to such an experiment. That is, fearful participants are found to already expect beforehand that fear-relevant stimuli will be more often followed by negative outcomes (e.g., Wiedemann, Pauli, & Dengler, 2001). Thus, fearful participants appear to exhibit both an a posteriori and an a priori covariation bias when estimating the co-occurrence of fear-relevant stimuli and negative (threat-relevant) outcomes. In certain types of anxiety disorders, however, the feared stimuli are found to be more clearly associated with disgust-relevant * Corresponding author. Faculty of Social Sciences, Institute of Psychology, Erasmus University Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands. E-mail address: mayer@fsw.eur.nl (B. Mayer). Contents lists available at ScienceDirect Journal of Behavior Therapy and Experimental Psychiatry journal homepage: www.elsevier.com/locate/jbtep 0005-7916/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbtep.2010.09.002 J. Behav. Ther. & Exp. Psychiat. 42 (2011) 19e25