The moderating effects of contamination sensitivity on state affect and information processing: Examination of disgust specificity

Abstract

Although contamination sensitivity has been implicated in several disorders, there is a paucity of research examining the influence of this trait on various outcomes. Accordingly, the present study examined the extent to which individual differences in contamination sensitivity moderated state affect in response to a mood induction and subsequent information processing biases, as assessed by a lexical decision task (LDT). It was hypothesised that the moderating effects of contamination sensitivity would be specific to disgust responding to a negative but not positive mood induction, and to reaction times to disgust and fear compared to happy words on the LDT. The findings were largely consistent with this hypothesis, as contamination sensitivity predicted increased disgust and arousal to the negative mood induction. Contamination sensitivity was also a better predictor of reaction times to disgust and fear words than happy words. However, the moderating effect of contamination sensitivity on reaction times on the LDT was not mediated by its effects on response to the negative mood induction. Implications of these findings for conceptualising the role of contamination sensitivity and its association with disgust in specific disorders are discussed.

Keywords:

• Disgust
• Contamination
• OCD
• Bias
• Emotion

Notes

¹IAPS pictures included in the positive mood induction were: 2352, 1710, 5626, 2150, 8500, 2260, 8200, 4700, 2352, 1710, 5626, 2150, 8500, 2260, 8200, 4700 (between vignettes); 1920, 2000, 2091, 2030 (between LDT blocks). IAPS pictures included in the negative mood induction were: 5950, 6150, 2120, 1930, 6250, 6800, 9630, 9600, 5950, 6150, 2120, 1930, 6250, 6800, 9630, 9600 (between vignettes); 1302, 9611, 5970, 6940 (between LDT blocks).

²Because the three-way Condition × Word Type × PI interaction was non-significant (p>.70), to aid interpretation of the two-way interaction we deleted this term from the model.

³In light of the significant differences between the fear and happy conditions in regression slopes, the narrow region of significance for happy versus fear yielded by the JN procedure might appear surprising. The primary reason for this disparity is the fact that the standard error for the JN procedure is a weighted sum of the standard errors of the Word Type ×PI interaction and of the effect of Word Type. Because items were considered a random effect, the latter standard error was relatively high in the present context.