Using a Dual-Target Rapid Serial Visual Presentation Task (RSVP) as an Attention-Based Measurement Procedure of Sexual Preference: Is it Possible to Fake?

Abstract

As faking threatens the validity of sexual preference measurements in forensic contexts, we investigated in an analogue study the fakeability of the dual-target Rapid Serial Visual Presentation (RSVP), an attention-based measurement procedure. In the dual-target RSVP, when observers attempt to detect two targets (T1 and T2) in a stream of stimuli presented in rapid succession, identification of T2 is impaired when it follows T1 within approximately 500 ms, a phenomenon called Attentional Blink (AB). Emotional stimuli result in an “emotion-induced blindness” or “attentional rubbernecking” effect: AB increases (detection of T2 decreases) when T1 is salient and decreases (detection of T2 íncreases) if T2 is salient or when participants are concurrently engaged in distracting mental activity. The participants were 9 gay men, 8 straight men, and 12 straight men with the last group instructed to fake their T2 responses assuming the expected response style of gay men when the T1 and T2 stimuli were pictures of nude and clothed men and women. We found differences in the reporting of both T1 and T2 between the gay men and the faking straight men as a function of the type of T1 resulting in good differentiation between the two groups (AUC = 0.78–0.92). The results suggest that the different patterns of reporting are in most explainable by a combination of the increase in cognitive load in the faking group and of the subtlety of the attentional effects of whether the stimuli were intrinsically of sexual interest to the participants. We conclude that the RSVP as an attention-based measurement procedure of sexual preference has a moderate resilience to faking. The generalizability of these findings to forensic contexts should be further explored.

Key words:

• attention-based measurement
• attentional blink
• fakeability
• rapid serial visual presentation
• sexual preference

Acknowledgement

This research was supported by an Academy of Finland Grant (121232).