The emotional Stroop interference effect in anxiety: attentional bias or cognitive avoidance?

Interference effects on threat words in anxious subjects on the emotional Stroop task have generally been interpreted as evidence for mood-congruent attentional bias in anxiety states. However, several recent studies have yielded results that run contrary to this attentional bias explanation. The most important of these conflicting findings show that: (1) panic disorder patients displayed interference on threat words, but also on other emotional words, including positively valenced words, and (2) 'repressors' showed even greater interference than high trait anxious subjects. We propose an alternative explanation for these findings, in which both attentional bias and cognitive avoidance are assumed to operate in the emotional Stroop task, but in which cognitive avoidance is hypothesized to be chiefly responsible for the greater interference effects found in anxious subjects and 'repressors'. We suggest that future research into cognitive processes associated with anxiety states should employ a variety of experimental paradigms on the same subjects and include measures of 'defensiveness'.