Academic journal article Memory & Cognition

Reversing the Picture Superiority Effect: A Speed-Accuracy Trade-Off Study of Recognition Memory

Academic journal article Memory & Cognition

Reversing the Picture Superiority Effect: A Speed-Accuracy Trade-Off Study of Recognition Memory

Article excerpt

Speed-accuracy trade-off methods have been used to contrast single- and dual-process accounts of recognition memory. With these procedures, subjects are presented with individual test items and required to make recognition decisions under various time constraints. In three experiments, we presented words and pictures to be intentionally learned; test stimuli were always visually presented words. At test, we manipulated the interval between the presentation of each test stimulus and that of a response signal, thus controlling the amount of time available to retrieve target information. The standard picture superiority effect was significant in long response deadline conditions (i.e., ≥2,000 msec). Conversely, a significant reverse picture superiority effect emerged at short response-signal deadlines (<200 msec). The results are congruent with views suggesting that both fast familiarity and slower recollection processes contribute to recognition memory. Alternative accounts are also discussed.

In standard recognition memory tests, both previously studied and new items are presented and subjects are required to decide whether or not each item was presented at study. Two general classes of theories have been put forward to explain performance in recognition memory tasks. One approach to recognition memory is common to the so-called global matching models, such as TODAM (Murdock, 1982), SAM (Gillund & Shifrrin, 1984), and MINERVA 2 (Hintzman, 1986). Although they offer different representations of memory and retrieval mechanisms, global matching models maintain that subjects make familiarity judgments on the basis of a continuous index of memory strength.1 When a test item exceeds the criterion of memory strength set by the subject, the stimulus is judged as old. (For a review of global matching models of recognition memory, see Clark & Gronlund, 1996.)

A second class of theories-the so-called dual process models-posits that a single continuous quantitative index of memory strength is not sufficient to account for the results obtained in experiments in which recognition memory tasks are used (see, e.g., Jacoby, 1991; Mandler, 1980; for a review of dual process models, see Yonelinas, 2002). Two processes are proposed: one based on familiarity and the other on recollection. Despite some differences between models (see Yonelinas, 2002), familiarity is considered to be a rapid, automatic process, often observed to be sensitive to manipulations of the perceptual features of target items (see, e.g., Lamberts, Brockdorff, & Heit, 2002). Recollection, in contrast, is considered to be a slower, intentional process dependent on the retrieval of specific qualitative information about prior occurrence of the target items and predominantly sensitive to manipulations affecting semantic or conceptual encoding, such as level of processing (see, e.g., Atkinson & Juola, 1973; Boldini, Russo, & Avons, 2004; Jacoby, 1991; Mandler, 1980). Further evidence about the distinction between the time course of familiarity and that of recollection has been obtained in studies in which event-related brain potentials (ERPs) were examined through recognition memory tests. For example, Curran (2000) detected an early ERP component, at about 300-500 msec, associated with familiarity, and a later component, at about 400-800 msec, associated with recollection (for a recent review of the relevance to this issue of ERPs and functional neuroimaging studies, see Rugg & Yonelinas, 2003). It has often been suggested that familiarity and recollection influence recognition memory independently (see, e.g., Jacoby, 1991; Mandler, 1980; but see also Humphreys, Dennis, Chalmers, & Finnigan, 2000).

Various experimental procedures, such as the rememberknowprocedure(see,e.g.,Gardiner&Richardson-K]avehn, 2000) and the process dissociation procedure (see, e.g., Jacoby, 1991), have been used to assess the potential contributions of different processes to recognition memory (see Yonelinas, 2002, for a review of the techniques used to investigate the contributions of different processes to recognition memory). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.