Source Monitoring in Eyewitness Memory: Implicit Associations, Suggestions, and Episodic Traces
Total Page:16
File Type:pdf, Size:1020Kb
Memory & Cognition 2005, 33 (5), 759-769 Source monitoring in eyewitness memory: Implicit associations, suggestions, and episodic traces STEVE T. HEKKANEN University of Tampa, Tampa, Florida and CATHY MCEVOY University of South Florida, Tampa, Florida Both the distinctiveness heuristic and discrepancy detection hypotheses were investigated by inde- pendently manipulating both schema consistency and incidental suggestion in an eyewitness memory paradigm. A sequence of slides was shown, followed by a postevent questionnaire that contained both schema-typical and schema-atypical information. Fifteen minutes later, a source-monitoring task was administered. In Experiment 1, the proportion of source misattribution errors was greater for schema- typical items than for schema-atypical items, and the proportion of errors on suggested items was greater than that on control items. Suggestion affected schema-typical and schema-atypical items equally, providing no support for the predictions of either hypothesis. In Experiment 2, the interval be- tween the questionnaire and the source test was manipulated. The results of Experiment 1 were repli- cated under the short delay, whereas the proportion of errors increased under the long delay. An asso- ciative network model involving two types of episodic traces was used to account for the results. In the past decade, theoretical attention has focused on using the Deese/Roediger–McDermott (DRM) paradigm. explaining various types of errors that occur during the Roediger and McDermott’s (1995) study exemplifies how recollection of an event. Source misattribution errors occur this paradigm has been employed. They presented to par- when the times, locations, or sources of items are improp- ticipants several lists of words in succession. On a typical erly identified. The rate at which misattribution errors list (e.g., kid, adult, adolescent, etc.), each word was arise appears to be influenced by the use of the distinc- associatively related to a critical nonpresented word (e.g., tiveness heuristic (Dodson & Schacter, 2002; Schacter, child). Even though the critical words were never pre- Israel, & Racine, 1999): a metacognitive strategy that sented during the study phase, the participants frequently helps determine which recollected items were part of the identified them as having been presented. experience. Vivid details and distinctive information are The identification of those critical unstudied words as assumed to be part of the memorial representation of an “old” items on a recognition test was assumed to be due experience. When an item is recollected but lacks this in- to the activation of preexisting associates by the list words, formation, it is identified as new or novel. Sometimes making available implicit associative responses during problems occur, and an item that is novel is identified as the study phase (Roediger & McDermott, 1995; Under- part of the experience, resulting in a recall, recognition, wood, 1965). Because each of the list words was related or source misattribution error. Presumably, the novel to the critical word, that word was repeatedly activated item was somehow encoded along with vivid details that during study. McEvoy, Nelson, and Komatsu (1999) pro- made it indistinguishable from the items that were actu- vided evidence in support of this interpretation when ally encoded from the experience. The purpose of the they showed that the probability of falsely recognizing a present study was to investigate the use of the distinc- critical word significantly increased as the strength of the tiveness heuristic in the interpretation of source misattri- connections between the list words and the critical word bution errors in the context of an eyewitness paradigm. increased. Both Israel and Schacter (1997) and Schacter et al. Two factors may be responsible for the false recogni- (1999) evaluated the distinctiveness heuristic hypothesis tions that are observed in a DRM paradigm: Implicit as- sociative responses are activated by the list words during the study phase, and the use of the distinctiveness heuris- This research was supported by a Delo grant from the University of tic by the participants during the test phase would make Tampa to S.T.H. and Grant AG13973 from the National Institute on it difficult for them to distinguish between studied words Aging to C.M. We thank James Woodson and the reviewers for their and critical words. To evaluate this possibility, Schacter helpful and insightful comments. Correspondence concerning this arti- cle should be sent to S. Hekkanen, Department of Psychology, 401 West et al. (1999) presented each word in the list with a de- Kennedy Blvd., University of Tampa, Tampa, FL 33606 (e-mail: scriptive picture and found that this technique resulted in [email protected]). a reduced number of critical words being falsely remem- 759 Copyright 2005 Psychonomic Society, Inc. 760 HEKKANEN AND MCEVOY bered as studied words. Presumably, the words from the study and in other eyewitness studies may have been due lists were encoded with vivid pictorial details, whereas both to the activation of implicit associative items during the critical words were not. During the memory test, the encoding of the experience and to the combined in- these vivid details were used to distinguish between the fluence of suggestion and activation that occurred during studied words and the critical words, leading to improved the reading of the questionnaire. accuracy. Smith and Studebaker’s (1996) results provide some The occurrence of implicit associative responses and support for this interpretation. Participants in the first of the use of the distinctiveness heuristic may help to ex- their three experiments listened to an audiotaped de- plain why participants in an eyewitness paradigm tend to scription of an experience, followed by a questionnaire include suggestions in their recollection of an experience. that contained information about details of the experi- In a typical eyewitness testimony procedure (see, e.g., ence. The consistency of the subsequent information Zaragoza & Lane, 1994), participants study a pictorial ex- with the audio experience was manipulated in the fol- perience and then read a set of statements/questions with lowing manner: The information biased the participants embedded misinformation that is always consistent with toward a typical interpretation of an experiential detail, the pictorial experience. After a short delay, a source- biased them toward an atypical interpretation, or did not monitoring task shows that the suggested misinforma- bias them toward any interpretation. After completing tion is frequently identified as having occurred in the the questionnaire, the participants took a forced-choice witnessed event. recognition test in which each question had three alter- The pictorial experience in an eyewitness paradigm is native responses: a typical response, an atypical response, usually organized around a schema. For example, in and a response that indicated that no such information Zaragoza and Lane’s (1994) study, a workman comes to was given in the audiotaped experience. The results an office and is instructed to fix a chair by an assistant. showed an interaction between consistency and type of When the assistant leaves, the workman rummages response: More typical responses were chosen after bi- through the drawers and bags in the office and takes a cal- asing by typical information, but when participants were culator and money. At least two sources of memory errors biased by atypical information the proportion of typical could have occurred in this study. The first could have responses decreased and the proportion of atypical re- occurred when the pictorial experience was presented to sponses increased. Smith and Studebaker concluded that the participants. The office schema and the items in the prior knowledge predisposed the participants toward office activated semantically associated information in typical information. an associative network (Anderson, 1976, 1985), much as Unfortunately, as the authors noted, the testing format the words in the study phase of a DRM paradigm activate may have caused the participants to pick typical rather implicit associative responses (Roediger & McDermott, than atypical responses, a problem that has been exten- 1995). Because the suggested items were consistent with sively examined by others (e.g., McCloskey & Zaragoza, the office schema, some of them may have been among 1985). Even though the problem was investigated in their the set of activated implicit items. The activation of im- second and third experiments, the contributions of con- plicit items while the pictorial sequence was being pro- sistency and suggestion remain unclear. In the present set cessed would have increased their levels of vividness. At of experiments, we manipulated both degree of schema the time of the test, the participants would have found consistency and type of incidental suggestion in an eye- the suggested typical items to be indistinguishable from witness study, for two purposes: to evaluate the relative ef- the other items of the experience, because both the pic- fects of the two factors and to determine whether they in- torial and the suggested typical items were encoded with teract. The misinformation items in the experiment were distinctive details. Thus, some source misattributions in either consistent with the pictorial experience (typical) or the Zaragoza and Lane study may have been the result of inconsistent