Academic journal article Memory & Cognition

Super Memory Bros.: Going from Mirror Patterns to Concordant Patterns Via Similarity Enhancements

Academic journal article Memory & Cognition

Super Memory Bros.: Going from Mirror Patterns to Concordant Patterns Via Similarity Enhancements

Article excerpt

When memory is contrasted for stimuli belonging to distinct stimulus classes, one of two patterns is observed: a mirror pattern, in which one stimulus gives rise to higher hits but lower false alarms (e.g., the frequency-based mirror effect) or a concordant pattern, in which one stimulus class gives rise both to higher hits and to higher false alarms (e.g., the pseudoword effect). On the basis of the dual-process account proposed by Joordens and Hockley (2000), we predict that mirror patterns occur when one stimulus class is more familiar and less distinctive than another, whereas concordant patterns occur when one stimulus class is more familiar than another. We tested these assumptions within a video game paradigm using novel stimuli that allow manipulations in terms of distinctiveness and familiarity (via similarity). When more distinctive, less familiar items are contrasted with less distinctive, more familiar items, a mirror pattern is observed. Systematically enhancing the familiarity of stimuli transforms the mirror pattern to a concordant pattern as predicted. Although our stimuli differ considerably from those used in examinations of the frequency-based mirror effect and the pseudoword effect, the implications of our findings with respect to those phenomena are also discussed.

In typical recognition memory experiments, participants are asked to discriminate items they have seen previously during a study phase (i.e., "old" items) from those they have not (i.e., "new" items). Classifying an old item correctly is called a hit, and mistakenly classifying a new item as old is called a false alarm. The difference between hits and false alarms provides an indicator of memory strength, and to assess effects on memory, differences can be contrasted across stimulus categories or empirical manipulations.

The present article focuses primarily on differences between stimulus classes; within the memory theory literature, two patterns of effects have informed memory theories. In cases where Stimulus Class A gives rise to both greater hits and lower false alarms that does Stimulus Class B the resulting pattern generally is called a mirror pattern, because the advantage of Class A in terms of more hits is mirrored in terms of fewer false alarms to that class. However, when Stimulus Class A gives rise to both higher hits and higher false alarms relative to Class B, the result is called a concordant pattern, because the hits and false alarms move in the same direction across the stimulus classes. The primary goal of the present work is to gain a better understanding of the factors that underlie the emergence of mirror patterns in some contexts and concordant patterns in others.

To aid in this goal and to provide a more explicit theoretical starting point for our considerations, we focus on a specific mirror effect and on a specific concordant effect, and we consider the potential factors relevant to each. With respect to the mirror pattern, the frequency-based mirror effect is the finding that low-frequency words give rise to more hits and fewer false alarms than do highfrequency words (e.g., Glanzer & Adams, 1985). With respect to the concordant pattern, the pseudoword effect is the finding that pseudowords (i.e., nonwords or very rare words) elicit more hits and more false alarms than do words (Greene, 2004; Hintzman & Curran, 1997; Whittlesea & Williams, 2000; Wixted, 1992). Although typically they are considered separately, we consider these two phenomena conjointly in order to assess the theoretical parallels and differences.

Both single- and dual-process accounts have been proposed to account for the frequency-based mirror effect. Traditional single-process accounts (e.g., Gillund & Shiffrin, 1984; Hintzman, 1988; Murdock, 1982) suggest that for each type of stimulus, an old and a new distribution can be placed on a scale of memory strength, on which participants adopt a criterion above which they call an item to be old. …

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.