Recollection-based recognition eliminates the revelation effect in memory

0 This research was conducted while D.L.W. was supported by Na tional Institute of Child Health and Human Development Mental Re tardation Research Training Grant 5-t32-HD07176. I thank Robert L. Greene , Tim Curran, Douglas K. Detterman, and Barry S. Layton for cerning this article should be addressed to D. L. Westerman, Depart ment of Psychology, State University of New York , P.O. Box 6000, Binghamton, NY 13902 ( The revelation effect refers to the tendency to call an item on a recognition test old if it is preceded by a cognitive task that involves the processing of a similar stimulus (Watkins & Peynircioglu, 1990). It has been proposed that the revelation effect occurs because of an increase in the familiarity of the test items in the revelation condition (Luo, 1993; Westerman & Greene, 1998). In the present experiments, the revelation effect was investigated in recognition tasks that were not based solely on the familiarity of the test items but,also, on a recall-like retrieval process. A revelation effect was not found on an associative-recognition task or on a plurality recognition task. The results of this study show that the revelation effect does not occur when the contribution of familiarity to recognition decisions is reduced by factors that encourage the recollection of the study episode. - Lately, a great deal of attention has been devoted to trying to understand the factors that lead a person to claim to remember something that did not actually occur (e.g., Hyman & Pentland, 1996; Pezdek, Finger, & Hodge, 1997; Read, 1996; Roediger, 1996; Zaragoza & Mitchell, 1996). Numerous studies have demonstrated that so-called memory illusions can be brought about by specific encod ing conditions. For instance, it is well known that study ing a list of conceptually related words, such as, bed, dream, and pillow, will make a person likely later to (in correctly) recall or recognize the word sleep (e.g., Deese, 1959; Payne, Elie, Blackwell, & Neuschatz, 1996; Roedi ger & McDermott, 1995; Tussing & Greene, 1997). Memory illusions can also occur because of factors that influence the memory test (see, e.g., Jacoby & White house, 1989; Lindsay & Kelly, 1996; Reinitz, Lammers, & Cochran, 1992; Whittlesea, 1993; Whittlesea, Jacoby, & Girard, 1990). One robust finding that has been dem onstrated primarily on tests of recognition memory is the revelation effect (Watkins & Peynircioglu, 1990). As it was originally reported, the revelation effect referred to the finding that a word on a recognition test will tend to be called old (i.e., previously studied) if it first appears in a distorted fashion-for instance, as a word fragment or as an anagram (e.g., rcukt)-and the research partici pant has to solve the word by putting it into its correct form (truck) prior to the recognition judgment. The ten dency to claim that items that are revealed had been pre sented previously occurs both for items that actually had been presented (targets) and for items that had not been presented (lures); however, the magnitude of the revela tion effect is typically larger for lures than for targets (see Hicks & Marsh, 1998, for a meta-analysis). As a result, revealing an item on a recognition test leads to a greater increase in false alarms than in hits; therefore, recogni tion performance tends to be less accurate in the revela tion condition. Recent research has shown that the term revelation effect appears to be a misnomer, since the ef fect does not depend on the revelation of the test word and has been found with a variety of tasks interleaved among test trials. A revelation effect has been found when the word that was revealed was not the word that was subsequently recognized but, rather, a different word that had not been presented on the study list and that bore no apparent relationship to the test word. For instance, un scrambling an anagram of the word vineyard made par ticipants more likely to claim that a different test word that followed it, such as raindrop, had been previously stud ied, as compared with test words that did not follow an anagram task (Westerman & Greene, 1996). Other cog nitive tasks, such as counting letters, generating synonyms, and recalling a short list of letters, have also been found to produce a revelation effect. That is, test items that im mediately followed these tasks were more likely to be called old than were test items that did not follow these tasks (Westerman & Greene, 1998). A theoretical explanation of the revelation effect has been elusive. One possible explanation is that the cognitive task that is performed during the revelation phase some how increases the familiarity of the test item that follows it (LeCompte, 1995; Luo, 1993; Peynircioglu & Tekcan, 1993; Westerman & Greene, 1998). Evidence in support ofthis notion was reported by LeCompte, who conducted experiments that used the remember/know procedure (Tulving, 1985) and the process dissociation technique (Jacoby, 1991) to study the contribution of familiarity to recognition decisions in revelation effect experiments. The results of LeCompte's experiments were consistent with the idea that the revelation task increases the famil iarity of the test item that follows it; both procedures yielded higher familiarity estimates for test items in the revelation condition. Several investigators (Luo, 1993; Peynircioglu & Tekcan, 1993) have suggested that an en hanced sense offamiliarity for revealed test items might be due to a misattribution of familiarity (see Jacoby & Whitehouse, 1989). According to this account, the pro cessing ofa stimulus during the revelation phase increases its familiarity, and the enhanced familiarity is misattrib uted to the test item's presence on the study list. Although a misattribution account can explain the revelation effect when the items that are processed during the revelation phase match those items that are recognized, it cannot account for a revelation effect when the stimulus that is processed during the revelation phase is completely dif ferent than the word that is recognized. For instance a fa miliarity misattribution account cannot explain why, after solving an anagram of the word vineyard, partici pants are more likely to claim that a different word (e.g., raindrop) had appeared on the study list (Westerman & Greene, 1996, 1998), nor can it explain why participants are more likely to call a test word old after having com pleted a letter-span task (Westerman & Greene, 1998). Westerman and Greene (1998) have offered an alter native account of the revelation effect that is similar to the familiarity misattribution account, insofar as it at tributes the revelation effect to the enhanced familiarity ofthe test items in the revelation condition. This account has been described in terms of the global matching ap proach to recognition memory (e.g., Gillund & Shiffrin, 1984; Hintzman, 1988; Murdock, 1993; for a review, see Clark & Gronlund, 1996). According t (...truncated)