Overprediction of anxiety, and disconfirmatory processes, in anxiety disorders.

The overprediction of anxiety phenomenon and its relationships with fear, dysfunctional and functional beliefs, and emotional experiences during confrontations with feared stimuli were investigated in two studies. Study 1 investigated exposure in vivo exercises executed by anxiety patients during treatment (n = 37). Study 2 investigated behavioural experiments executed by anxiety patients (n = 11) during cognitive treatment. In both studies patients rated various variables just before and immediately after their exercises. The results indicate that anxiety patients tend to overpredict the level of anxiety they are going to experience. There is no evidence that this phenomenon is a statistical artefact, caused for instance by a pre-test response style. There is also no evidence that the observed adjustment of incorrect anxiety predictions is a statistical artefact. A global negative emotional evaluation of the experience appears to have an adverse influence on the reduction of anxiety predictions and on the reduction of fear. Fear, and its reduction after exposure in vivo or after behavioural tests, appears to be influenced by both anxiety predictions and dysfunctional beliefs. Positive emotions and functional beliefs did not appear to play an important role. The interrelationships between these factors are comprised in a path-model which describes how emotional and cognitive information yielded by disconfirmatory experiences influences changes in anxiety predictions and changes in fear. Unexpectedly, change in dysfunctional belief did not relate directly to change in fear, but only indirectly, via change in anxiety predictions. Theoretical and therapeutical implications are discussed.