Abstract stimuli associated with threat through conditioning cannot be detected preattentively

Studies of anxiety suggest that threat stimuli can be identified preattentively, but this conclusion is questionable because of possible low-level perceptual confounds. Two experiments used visual search tasks in which abstract shapes were conditioned to carry neutral or negative valence. Experiment 1 found generally faster responses to threat-associated abstract stimuli but no evidence that they were detected preattentively, irrespective of trait anxiety level. A similar pattern was found in Experiment 2, in which individuals high in snake or spider fear showed no evidence of preattentive detection of abstract stimuli associated with their feared object. In contrast, implicit behavioral measures showed significant effects of conditioning, demonstrating that targets associated with threat were negatively evaluated in these experiments.