In the present study, two groups of 21 subjects with either high or low fear of both snakes (or spiders) and damaged electrical outlets/appliances participated in a two phase experiment. After reading a description of an illusory correlation experiment, subjects were asked to imagine themselves participating in it. They rated their expectations for the number of occasions on which slides of snakes, electrical outlets, and flowers would be followed by either a shock, tone, or nothing. As predicted, both high and low-fear subjects reported an expectancy bias for both phylogenetic and ontogenetic fear-relevant stimuli and shock. In the second phase subjects were actually exposed to the random slide/outcome presentation. Only high-fear subjects demonstrated a covariation bias which was specific to phylogenetic fear-relevant slides and shock, indicating all other biases were effectively attenuated.