Assessing the Effects of Misleading Post-Event Information Using Multinomial Processing Tree Models

Abstract

Multiple studies have shown that misleading post-event information can alter an individual's memory. Three hypotheses (no-conflict, coexistence, destructive updating) have been proposed to explain the fate of the original memory trace and have subsequently been mathematically formalized to gain a more comprehensive understanding of the predictions regarding false memory formation (Wagenaar & Boer, 1987). We utilized multinomial processing tree (MPT) models to test these hypotheses concurrently. In two experiments, we implement the Loftus (1978) misinformation paradigm to subsequently apply MPT models to the data. In Experiment 1 and Experiment 2, we found support for both the no-conflict and coexistence models, but due to the no-conflict model being the most parsimonious model, we defaulted to the no-conflict model. However, when only the top-performing participants were examined, we found strong evidence for the coexistence model. In Experiment 2, we also categorized participants based on their perceptions of what happened to their original memory and used these distinctions to determine if there was a correspondence between participants’ intuitions and model fits. We found some correspondence between participants categorized as endorsing No-Conflict and minimal support for Coexistence and the respective models. Surprisingly, we did not replicate the overall misinformation effect that we found in Experiment 1. However, we did find a misinformation effect for the participants classified as Coexistence and Destructive Updating, suggesting that participants who acknowledged a conflict were affected by the conflicting information. Future research should continue to apply these models to different sub-sets of participants to examine the extent to which participants are aware of a conflict.