When do eyewitness behaviors reflect the presence or absence of the culprit in a lineup?

Decades of research demonstrates that confidence and decision time are useful for distinguishing correct from incorrect suspect identifications and that other eyewitness lineup behaviors might also prove useful (e.g., automaticity, deliberative strategies, featural references). However, these same “postdictors” have proven less effective at distinguishing correct from incorrect rejection decisions. Moreover, except for confidence and decision time, postdictor variables have typically been studied in isolation rather than in combination. In this thesis I address both problems. I theorized that the failure of postdictors to distinguish correct from incorrect rejections was attributable to standard laboratory methodologies restricting eyewitness memory strength for the culprit. Accordingly, I exposed participant-eyewitnesses (N = 790) to a wide range of viewing conditions with the goal of increasing variability in both memory strength and postdictor variables. After viewing a simulated crime, participants completed a short filler task, and then completed either a culprit-present or culprit-absent lineup in the presence of a “blind” administrator. Eyewitnesses were video recorded as they completed the lineup procedure. Raters, who were blind to the presence or absence of the culprit in the lineup, then evaluated these recordings and scored each eyewitness on 58 behavioral variables. Logistic regression analyses with a LASSO constraint, revealed that confidence, decision time, and numerous other behaviors, worked in combination to distinguish not only correct from incorrect suspect identifications, but also correct from incorrect rejections. Many of these useful postdictors seemed to reflect how difficult eyewitnesses found the lineup task (eyewitness fluency). As a general rule, one might put more faith in the decision of a fluent eyewitness and less faith in the decision of a disfluent eyewitness