Covariation bias for food-related control is associated with eating disorders symptoms in normal adolescents

Background and ObjectivesCovariation 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, covariation bias was investigated within the context of eating pathology.MethodsIn a sample of 148 adolescents (101 girls, 47 boys; mean age 15.3 years), a priori and a posteriori contingencies were measured between words referring to control and loss of control over eating behavior, on the one hand, and fear, disgust, positive and neutral outcomes, on the other hand.ResultsResults indicated that all adolescents displayed an a priori covariation bias reflecting an overestimation of the contingency of words referring to loss of control over eating behavior and fear- and disgust-relevant outcomes, while words referring to control over eating behavior were more often associated with positive and neutral outcomes. This bias was unrelated to level of eating disorder symptoms. In the case of a posteriori contingency estimates no overall bias could be observed, but some evidence was found indicating that girls with higher levels of eating disorder symptoms displayed a stronger covariation bias.ConclusionThese findings provide further support for the notion that covariation bias is involved in eating pathology, and also demonstrate that this type of cognitive distortion is already present in adolescents.

► Covariation bias in boys and girls was measured in the context of eating pathology.
► A priori overestimations were found of loss of control and negative outcomes.
► A priori overestimations were found of control and positive outcomes.
► This was more explicitly present in girls but unrelated to eating disorder symptoms.