Generalizability of the match/mismatch model of fear

Abstract

An experiment was carried out with 19 normals to investigate the generalizability of the match/mismatch model of fear towards a general model of human information processing of experiences with aversive stimuli. The Ss received 10 painful electric shocks and rated experienced pain, predicted pain at the next trial, certainty of prediction and fear of shock. The time used to fill out the rating scales was also measured. An experimental mismatch was induced in 10 Ss. Since the Rachman et al. studies used an unsatisfactory statistical procedure, alternative analyses were tried. Results of the experimental mismatch and of spontaneous matches/mismatches strongly support the core of the model: underpredictions were followed by increases in predicted pain, overpredictions by decreases. The experimental mismatch showed long-term effects on fear and certainty. After underpredictions, acceleration tendencies were found. The results indicate that the Ss' information processing is characterized by a strong recency effect and an assymetry between the two types of mismatch that is not self-evident. The model seems to offer a powerful paradigm for aversive information processing.