Mood state-dependent memory: A meta-analysis

Abstract

Published research (from 1975 to 1985) on mood state-dependent memory is examined with meta-analytical techniques. Research on mood-congruent memory is also examined when relevant to a mood state-dependent hypothesis. Effect sizes are shown to vary systematically as a function of several variables. Mood state-dependent memory is more often observed with positive rather then negative moods. Also, the strongest effects were found in studies that: (a) used real-life events as the material to be retrieved; (b) selected contrasting moods, i.e. positive vs. negative rather than simply positive or negative vs. neutral moods; (c) used mood states that were specific to the items; (d) lasted longer; (e) had smaller sample size; (f) selected subjects according to direct mood criteria; (g) paid subjects for their participation in the experiment; and (h) used adults rather than students or children. Present findings are consistent with other reviews and with common theories of mood state-dependent memory. However, the findings also highlight the importance of methodological variables, more particularly variables related to the degree of commitment of subjects to the experiment, and variables related to the nature and complexity of the experimental environment. Two dimensions, realism of the experiment and demand sensitivity, compete to account for the methodological results and for some of the theoretical results.