UCL Psychology and Language Sciences


Experimental Psychology Seminar - Are eyewitnesses untrustworthy?

01 November 2016, 4:00 pm–5:00 pm

Event Information


room 305, 26 Bedford Way, WC1H 0AP

Speaker: Dr Laura Mickes, Royal Holloway University of London

To inform a criminal investigation, police may ask an eyewitness to that crime to try to identify the perpetrator from a lineup. Eyewitnesses, however, have a bad reputation for being untrustworthy. Are they? To answer that question, two types of accuracy must be separately considered: discriminability and reliability. To assess discriminability, false identifications (incorrectly identifying the innocent suspect) and correct identifications (correctly identifying the guilty suspect) are measured together. The lineup procedure with the best discriminability is the one that yields a low false identification rate and a high correct identification rate. Reliability has to do with the probability that the identified lineup member is the perpetrator. To assess reliability, confidence expressed at first identification should be taken into consideration. An eyewitness who identifies the suspect with high confidence is likely to be accurate. On the other hand, an eyewitness who identifies the suspect with low confidence is likely to be making an error. Either way, the confidence expressed during initial identification is often diagnostic of accuracy. Once the distinction between the types of accuracy is made, and the relevant analyses are conducted, the results challenge the notion that eyewitnesses are untrustworthy.

Time: 1st November 2016, 4pm
Venue: Room 305, 26 Bedford Way, WC1H 0AP