Abstract
We tested direct, indirect, and behavioural measures of fear of spiders under neutral instructions, and when participants were asked to fake high and low fear of spiders. Our findings indicate that the Approach Avoidance Task (AAT) was the only measure that could be faked in one of the faking conditions only. We also assessed how easily faked results could be detected on each measure for different diagnostic criteria. The direct and behavioural measures showed good performance for all criteria. The AAT performed comparably only for a conservative criterion, when detecting fakers is less important than correctly labelling non-fakers.
Acknowledgements
This study was funded by a research grant of the Behavioural Science Institute of the Radboud University Nijmegen.
Notes
1By indirect measures we refer mainly to reaction time tasks, as those are widely used. But indirect measures of other variables than latencies exist.
2Usually AAT data are analysed as AAT effects: For each condition, median pull RTs are subtracted from the corresponding push RTs. Positive (negative) AAT effects indicate stronger approach (avoidance) tendencies towards the stimulus. As faking appeared systematically on RT level, the analysis of AAT effects is omitted.
3We used the arcsine transformation of the form , which is standardly used to correct for the inherent dependence of means and variances in proportion data (Winer, Citation1971).