Valence and the development of immediate and long-term false memory illusions

Abstract

Across five experiments we examined the role of valence in children's and adults’ true and false memories. Using the Deese/Roediger-McDermott paradigm and either neutral or negative-emotional lists, both adults’ (Experiment 1) and children's (Experiment 2) true recall and recognition was better for neutral than negative items, and although false recall was also higher for neutral items, false recognition was higher for negative items. The last three experiments examined adults’ (Experiment 3) and children's (Experiments 4 and 5) 1-week long-term recognition of neutral and negative-emotional information. The results replicated the immediate recall and recognition findings from the first two experiments. More important, these experiments showed that although true recognition decreased over the 1-week interval, false recognition of neutral items remained unchanged whereas false recognition of negative-emotional items increased. These findings are discussed in terms of theories of emotion and memory as well as their forensic implications.

Keywords:

• Valence and memory
• False memory development
• DRM paradigm

Acknowledgements

This research was supported by a grant to MLH from the Economic and Social Research Council of Great Britain (RES-062-23-0452) and by a grant to IC from the Netherlands Organisation for Scientific Research (NWO 451-03-013).

Notes

¹No differences were found as a consequence of whether emotional lists appeared first or second. Because different orders were constructed for methodological not theoretical reasons, the data were collapsed across order for all of the analyses reported in this article.

²No differences were found across the various versions of the recognition tests. Because these different versions were constructed for methodological not theoretical reasons, the data were collapsed across test version for all of the analyses reported in this article.

Across all five experiments in this article there were no effects of blocking or gender. As there were no theoretical hypotheses concerning these variables in this article, and because neither variable affected performance, they were not included in subsequent analyses for any of the experiments.

Throughout this article it is important to note that the recognition findings were unaffected by the prior recall test. That is, when recognition probabilities were conditionalised on whether an item had been recalled earlier, the patterns of findings were no different than when the unconditional recognition probabilities were analysed (for similar findings, see Marche, Brainerd, Lane, & Loehr, 2005). In addition, using a separate but smaller sample of 5-, 7-, and 11-year-olds and adults, we examined recognition performance without an intervening recall test. Analyses of these data showed the same pattern of results as the ones reported here that were preceded by a recall test.

For Experiments 2, 4, and 5 involving children, the recognition results were the same regardless of whether raw scores or A′ was used in the analyses (for a discussion of the use of signal detection analyses, see Pastore, Crawley, Berens, & Skelly, 2003). Because there were no differences in the outcome of the analyses, and because we report the raw score recognition findings for adults in Experiments 1 and 3, we report the untransformed, raw score analyses for the recognition data from all participants across the five experiments in this article.