Negative self-imagery blocks inferences

Previous research has shown that, unlike non-anxious individuals, people with social phobia fail to generate non-threatening inferences when ambiguous social information is first encountered (i.e. ‘on-line’; Hirsch and Mathews Journal of Abnormal Psychology, 109 (2000) 705–712). Patients with social phobia also report negative self-imagery in social situations, while individuals without high social anxiety do not. The negative self-imagery in social phobia may prevent the generation of non-threatening inferences. If so, then training non-anxious individuals to hold in mind a negative self-image should remove the ‘on-line’ non-threat inferential bias normally evident in this population. In the present study, low anxious volunteers were allocated to negative image training or a control task that did not manipulate self-imagery. Following negative image training, or the control task, volunteers read descriptions of job interviews and at certain points during the text performed lexical decisions. Some decisions were made after ambiguous text that could have been interpreted in both a threatening and a non-threatening manner. In a baseline condition, decisions were made following the text for which there was only one possible inference (either threat or non-threat). The results for the control group replicated earlier findings of a non-threat inferential bias for non-anxious individuals. In contrast, and as predicted, non-anxious volunteers who were trained to hold a negative image in mind lacked any non-threatening inferential bias, and also experienced higher levels of state anxiety.