Expectancy bias in trait anxiety.

The ability of anxious and control participants to learn about signals for danger and safety was tested within an autonomic conditioned inhibition (A+/AB-) procedure. Only participants who could verbalize the differential contingencies between the stimuli and shock (aware participants) showed discrimination on electrodermal and shock expectancy measures. In Experiment 1, aware high-anxious participants showed similar responding to control participants. However unaware high-anxious participants showed heightened shock expectancy to all stimuli. Experiment 2 replicated this expectancy bias in anxious unaware participants controlling for shock intensity. In both experiments, expectancy bias was associated primarily with anxiety rather than depression. Results support the notion of an interpretive bias in anxious participants under ambiguous threat, consistent with recent findings from information-processing research on linguistic stimuli.