ABSTRACT
The goal of eyewitness research is to find the procedures most likely to result in identifications of guilty suspects without causing mistaken identifications of innocent suspects. This is not always easy. There are sometimes tradeoffs where increasing the likelihood of identifying a guilty suspect also increases the likelihood of a mistaken identification or vice versa. There is debate regarding the best way to make decisions when presented with tradeoffs, as different analyses will sometimes lead to different results as to which is the “best” procedure. In this chapter we discuss ways of using different types of measurement to deal with these trade-off examples, including the current debate between ROC analysis and utility approaches. We conclude that the best way of resolving tradeoffs is to directly measure expected utility under a number of different plausible assumptions concerning base rates and cost structures. We also use simulated and published data to examine how often there is disagreement between measures as to the “best” paradigm.
