interest is that the current effects of threat are expressed in terms of intercept differences rather than slope difference. As an additional check, the RT scores were analyzed in the same way as the efficiency scores and exactly the same pattern of effects arose. That is, when response speed was considered, the corresponding search functions for the different target trials revealed intercept and not slope differences. This suggests that the effects reflect non-search processes. A suggestion is that the time to find the target did not vary according to threat content but that the time to respond to the content did. Participants were simply quicker to respond to the threatening than the non-threatening images. This in turn may be linked to the fact that the threat images were more arousing than the non-threatening images (cf. Lundqvist et al., 2014). Aside from this, the most striking patterns of performance relate to the data from the classification task. In this case, participants were less efficient in responding to the threatening targets than the non-threatening targets. The data revealed robust reverse
effects of threat. Indeed such a reverse pattern contrasts with the threat advantage found in the corresponding detection task. Whereas there was a threat advantage when participants responded to the presence of a distinctive target, there was a reverse threat effect when participants were asked to search for and classify the distinctive target. This reverse threat effect was unexpected but was consistent across all display sizes and for both the cat and dog images. This effect
was also expressed as an intercept and not a slope effect. Indeed when the corresponding RT data were analyzed in the same manner as the efficiency scores then the same pattern of statistical significance arose. What this again suggests is that the time to search for the target did not differ across the different target types but the time to respond to them did. In order to test this simple idea a final experiment was carried out in which the classification task was repeated but in a non-search version of the paradigm. In this case only a single image was presented on each trial. The image occupied one of the previous peripheral image locations used in Experiments 1 and 2 and this was chosen at random prior to the start of the trial. If the reverse threat effect is a reflection of a decisional, non-search process, then it Ponatinib manufacturer
should recur in a non-search version of the task. Experiment 3 In this final experiment a new sample of participants was tested in a partial replication of the classification task in Experiment 2. The images used in Experiment 2 were used here. Participants Twenty-four na?ve participants (mean age = 221, 20 female) were recruited from students of The University of York. There were three left-handed individuals. They received either a course credit or ?4.