TY - JOUR
AU1 - Wells, Gary L.
AU2 - Lindsay, R. C.
AB - Questions the criminal justice system's practice of treating eyewitness lineup identifications (IDs) of suspects (SUs) as highly informative while treating nonidentifications (NIDs) (i.e., no choice responses or choices of foils) as uninformative. A Bayesian model of information gain is used to mathematically prove that (a) if an eyewitness ID of an SU increases the probability that the SU is the criminal, then an NID must decrease the probability that the SU is the criminal; and (b) the  relative  diagnosticity of IDs vs NIDs is determined by the probability of obtaining an ID vs NID, with NIDs being more diagnostic if they are relatively less frequent than IDs. An application of the model to previously published data shows NIDs to be more than 150% as diagnostic as IDs regarding the probability that the SU is the criminal. A breakdown of NIDs into 2 types, eyewitness choices of a lineup foil vs no-choice decisions, suggests that the latter is more informative than the former regarding the probability that the SU is innocent. Cognitive mechanisms that may be responsible for criminal justice investigators discounting of NIDs are discussed in relation to research on human judgment. (22 ref)
TI - On estimating the diagnosticity of eyewitness nonidentifications
JF - Psychological Bulletin
DO - 10.1037/0033-2909.88.3.776
DA - 1980-11-01
UR - https://www.deepdyve.com/lp/american-psychological-association/on-estimating-the-diagnosticity-of-eyewitness-nonidentifications-GeCQ551gSA
SP - 776
EP - 784
VL - 88
IS - 3
DP - DeepDyve
ER -