Remembering discrepant information in eyewitness testimony. Experiments and paired-associate learning tasks - an integrative model

Abstract

Part I of responses and (b) a local test procedure is used that asks for responses from only one of the lists at a time. Conversely, with a global test procedure that simultaneously asks for both responses associated with a stimulus term, no such effects are predicted. The predictions were confirmed in two experiments, using a two-alternative forced-choice test (experiment 1) or a multiple choice test with a 'don't know' option (experiment 2) as the local test procedure. The results were further analyzed in terms of the model and revealed a reasonably good fit of post-hoc simulations to the actual data. Implications of these findings for research on paired-associate learning and for eyewitness interrogation procedures are discussed.