Effects of pictorial context reinstatement on correct and false recognition memory: insights from the simplified conjoint recognition paradigm

ABSTRACT

We used the Fuzzy-Trace Theory (FTT) perspective to investigate the role of context switching or reinstatement on correct and false recognition in Deese–Roediger–McDermott (DRM) and simple false memory for occurrence (FMO) tasks. In two experiments, participants studied words presented against unique picture backgrounds. The results indicated that providing specific item-context associative information in the reinstated-context test condition enhances verbatim trace retrieval of targets but does not influence their gist retrieval. In the case of false recognitions of related distracters, manipulations of retrieval context resulted in different effects across experiments, putatively because the DRM and FMO tasks involve different latent-processes implied by FTT. The results confirmed the critical role of retrieval conditions for false and correct recognition memory and showed the insights that can be gained from the process-level multinomial modelling analyses of data obtained in particular experimental paradigms.

KEYWORDS:

• DRM paradigm
• context reinstatement
• false recognition
• recollection
• Fuzzy-Trace Theory

Acknowledgements

The authors are grateful to David Kellen and anonymous reviewers for helpful comments on an earlier draft of this paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Marek Nieznański http://orcid.org/0000-0002-2457-3909

Notes

1 Welch statistic was used because variances were significantly heterogeneous across the conditions.

2 We considered an alternative multinomial model imposing no restrictions on recollection rejection parameters but restricting phantom recollection parameters to zero instead. For that purpose, we used the Fisher Information Approximation (FIA) index that represents a formal implementation of the Minimum-description-length principle for model selection (Klauer, Singmann, & Kellen, 2015; Wu, Myung, & Batchelder, 2010). The model used in Experiment 1 obtained smaller index value (2,646.9) than the alternative model (2,664.4) indicating that FIA favors the current model over the alternative one.

3 We considered an alternative multinomial model imposing no restrictions on phantom recollection parameters but restricting recollection rejection parameters to zero instead. A smaller FIA criterion value for the model used in Experiment 2 (2,750.7) than for the alternative model (2,754.6) indicates its better generalization.