Attentional biases in sub-clinical anxiety

The experiments presented here were designed to address experimental problems due to uncontrolled low level features by using highly controlled schematic stimuli. The emotional valance of these stimuli was changed by a conditioning procedure in which they became associated with either neutral or negative pictures. In order to test the effectiveness of this conditioning, subjects undertook a behavioural task (Implicit Association Test; IAT), which measured implicit valence to exclude the possibility of a null result simply reflecting ineffective conditioning. In Experiment 1, 32 participants searched for either threat-associated or neutral targets. There was not advantage for the threat-paired relative to the neutral-paired targets, although a non-significant result in the IAT meant that findings were inconclusive. To increase the effectiveness of the conditioning, Experiment 2 used a within-subjects design. 46 students searched for both treat-associated and neutral conditioned targets. Again, there were no differences in search for the neutral and threat-paired targets. Furthermore, search was unaffected by anxiety level. However, conditioning was significant as evidenced by a main effect of response congruity (subjects were faster to associate the negatively paired target with bad words than good words). As phylogenetic based theories of fear conditioning purport that snakes and spiders are particularly potent sources of threat, in Experiment 3, 63 undergraduates screened for high and low snake and spider fear searched for snake or spider and neutral-associated targets. Again, there were no differences in search efficiency across target type and fear level, although importantly, once more there was a significantly effect of conditioning. As studies 1-3 showed no evidence of preattentive biases towards threat-associated stimuli, Experiment 4 used a cueing paradigm to investigate whether biases in attention reflect difficulty in disengaging attention from threatening stimuli once identified.