The Role of Feature and Familiarity Justifications on the Interpretation of Eyewitness Confidence

I conducted four experiments that advance our knowledge of the interpretation of eyewitness confidence. These experiments focused on four key questions: (1) are eyewitness justifications—when combined with confidence and decision time—meaningful postdictors of identification accuracy?; (2) how accurately can observers interpret the intended meaning of an eyewitness’s confidence statement given a particular type of justification?; (3) what is the consequence of a particular kind of justification on an observer’s behavior?; and (4) are differences in perceived confidence across different types of justifications a result of expertise with faces or do these findings represent a more general memory phenomenon?

These experiments yielded several key findings: 1) familiarity justifications were significantly more likely to occur when not choosing within a lineup than any other justification type; 2) when choosing a face from a lineup, familiarity justifications were associated with lower accuracy and a poorer confidence/accuracy relationship at higher levels of confidence than other justification types; 3) differences in perceived confidence were minimized when justification type varied within-subjects, although multiple observable features were perceived as more confident than a single observable feature; 4) highly confident unobservable justifications were rated as stronger evidence than both observable featural justifications and confidence alone; and 5) the featural justification effect was not specific to faces, but also occurred for novel objects (e.g., greebles) and other crime-relevant stimuli (e.g., cars and weapons), suggesting that expertise alone does not drive the featural justification effect (i.e., the finding that observable justifications are perceived as less confident than unobservable justifications and confidence alone; Dodson & Dobolyi, 2015)—rather it may rely on judgments of the perceived memorability of features made "on the fly."