Is it all in the details? Description content and false recognition errors

In three experiments, we examined the effect of embellished content on memory errors for thematically related items as well as whether an encoding manipulation, specifically instructions to visualize content, further affects those errors. Using a modified Deese-Roediger-McDermott (DRM) paradigm, participants listened to subsets of DRM items embedded within scene descriptions. Some descriptions embellished item connections, weaving them into cohesive scenes. Other descriptions only made general reference to scenes, mentioning the items in list-like format. Listening to more detailed descriptions, compared to general scene references, elicited higher false recognition errors (Experiments 1 and 3). However, when description details varied within the same encoding series, as expected, false recognition errors did not differ significantly (Experiments 2 and 3). Results further suggest that more detailed content was more vividly imagined, providing one possible mechanism for increases in false recognition (Experiment 3). Implications for theoretical discussions of encoding task effects on false recognition errors are discussed.