Eyewitness identifications based on biased or unbiased line-up instructions after a realistic and violent hostage simulation

Few earlier studies have investigated the effects of highly stressful, realistic, violent, and threatening scenarios on eyewitness identification accuracy in an ecologically valid setting. The majority of studies have relied on laboratory-based simulated (videos/images) experiments. The present study investigated line-up accuracy approximately 1 week after a hostage simulation event. We administered biased line-up instructions to 50% of participants to investigate how this impacted choosing behaviour and accuracy. Based on 1030 line-up decisions (N = 122), we found that average accuracy was 38% in target present (TP) and 54% in target absent (TA) line-ups and that biased line-up instructions decreased overall accuracy (vs. unbiased). The hit rate for TP line-ups with biased instructions was 0.43 (unbiased instructions: 0.33), while the false alarm rate for TA line-ups with biased instructions was 0.60 (unbiased instructions: 0.32). We found that high confidence was associated with correct identifications and that shorter response times were indicative of correct rejections. Our findings demonstrate, in a more realistic scenario than the majority of eyewitness identification studies, the effect that biased line-up instructions lead to increased choosing and decreased accuracy.