Acts of generating and their sources: Predicting the effects of imagery encoding on false recognition errors

Abstract

In the two experiments reported here the basis of the beneficial effects of generating images on false recognition errors is investigated. Acts of generating (descriptions, images, or both) were manipulated while examining the effects of the source of descriptions guiding imagery generations (participant vs peer). False recognition errors were relatively high across encoding conditions except when imagery generations were based on participants' own descriptions (Experiments 1 and 2). These differences in the acts of generating were not attributable to differences in the cohesiveness of descriptions themselves. Acts of generating led to greater “remember” responses than “know” responses only when participants were not the source of the descriptions used to generate images (Experiment 2). Results highlight the importance of examining the effects of the source of descriptions for guiding imagery (participant or peer) when testing predictions about the effects of imagery encoding on false recognition errors.

Keywords:

• False memory
• Imagery encoding
• Source monitoring

Acknowledgements

This research was supported by NSF Grant #BCS 1023890. We express appreciation to Carly Hughes, Sarah Landsberg, Zack Rowen, and Wendy Roedriquez for their help with data collection and scoring. We also extend appreciation to Rebecca Brooke Bays, Adina Fried, Steve Lindsay, and an anonymous reviewer for their constructive feedback on an earlier version of this manuscript.

Notes

¹Participants' descriptions were also examined for the presence of explicit self-referential and autobiographical comments. A response was denoted as self-referential if a participant made an explicit reference to his or her self (e.g., reporting “I” or “my”) or to some kind of self-relevant detail (e.g., referring to a relative, a family trip, etc.). Because these frequencies were quite low, they were not discussed further. They are mentioned here for the sake of completeness. Furthermore, conditions involving participant-generated or yoked (peer-generated) materials had comparable likelihood ratings.