Confidence and accuracy, while often considered to tap the same memory representation, are often found to be only weakly correlated (e.g. Bothwell, Deffenbacher, & Brigham, 1987;Deffenbacher, 1980). There are at least two possible (nonexclusive) reasons for this weak relation. First, it may be simply due to noise of one sort or another; that is, it may come about because of both within-and betweensubjects statistical variations that are partially uncorrelated for confidence measures on the one hand and accuracy measures on the other. Second, confidence and accuracy may be uncorrelated because they are based, at least in part, on different memory representations that are affected in different ways by different independent variables. We propose a general theory that is designed to encompass both of these possibilities and, within the context of this theory, we evaluate effects of four variables--degree of rehearsal, study duration, study luminance, and test luminance-in three face recognition experiments. Inconjunction with our theory, the results allow us to begin to identify the circumstances under which confidence and accuracy are based on the same versus different sources of information in memory, The results demonstrate the conditions under which subjects are quite poor at monitoring their memory performance, and are used to extend cue utilization theories to the domain of face recognition.Of interest in numerous circumstances is the ability to assess the degree to which a person's reported memory faithfully reflects the original, objective reality that gave rise to the memory. One such circumstance, for example, is the common legal scenario wherein a witness to a crime identifies a suspect as the person who committed the crime. Another is a laboratory setting wherein a subject claims to recognize a test stimulus in a recognition experiment.In a controlled laboratory setting, the researcher has various tools available to assess memory. Two of the most commonly used are accuracy and confidence. Thus, to each recognition test stimulus, a subject can respond "old" or "new" and can also provide a confidence rating (say on a scale from 1 to 5) indicating his/her subjective assessment that the just-made recognition response is correct. Often, these two kinds of responses are assumed, either implicitly or explicitly, to be two measures of the same underlying psychological dimension. Thus experimenters often report both confidence and accuracy as parallel measures, or combine them into a single measure (e.g, multiplying a 1-5 point confidence rating by

The false recognition of distractor faces created from combinations of studied faces has been attributed to the creation of novel traces in memory, although familiarity accounts are also plausible. In 3 experiments, participants studied parent faces and then were tested with a distractor that was created by morphing 2 parents. These produced high false-alarm rates but no effects of a temporal separation manipulation. In a forced-choice version, participants chose the distractor over the parents. R. M. Nosofsky's (1986) Generalized Context Model and variants could account for some but not all aspects of the data. A new model, SimSample, can account for the effects of typicality and distinctiveness, but not for the morph false alarms unless explicit prototypes are included. The conclusions are consistent with an account of memory in which novel traces are created in memory; alternative explanations are also explored.

Experimental research on eyewitness identification follows a standard principle of experimental design. Perpetrator-present and perpetrator-absent lineups are constructed with the same foils, so that the two conditions are identical except for the presence or absence of the true perpetrator of the crime. However, this aspect of the design simulates conditions that do not correspond to those of real criminal investigations. Specifically, these conditions can create perp-absent lineups in which the foils are selected based on their similarity to an unknown person-the real perpetrator. Analysis of the similarity relations predicts that when foils for perp-absent lineups are selected based on their match to the perpetrator the false identification rate will be lower than if the foils are selected based on their match to the innocent suspect. This prediction was confirmed in an experiment that compared these two perp-absent lineup conditions. These results suggest that false identification rates in previous experiments would have been higher if the foils had been selected based on their match to the innocent suspect, rather than the absent perpetrator.

Two experiments directly compare two methods of selecting foils for identification lineups. The suspect-matched method selects foils based on their match to the suspect, whereas the description-matched method selects foils based on their match to the witness's description of the perpetrator. Theoretical analyses and previous results predict an advantage for description-matched lineups both in terms of correctly identifying the perpetrator and minimizing false identification of innocent suspects. The advantage for description-matched lineups should be particularly pronounced if the foils selected in suspect-matched lineups are too similar to the suspect. In Experiment 1, the lineups were created by trained police officers, and in Experiment 2, the lineups were constructed by undergraduate college students. The results of both experiments showed higher suspect-to-foil similarity for suspect-matched lineups than for description-matched lineups. However, neither experiment showed a difference in correct or false identification rates. Both experiments did, however, show that there may be an advantage for suspect-matched lineups in terms of no-pick and rejection responses. From these results, the endorsement of one method over the other seems premature.

No abstract