Effects of Item Relatedness on Output Interference in Recognition Memory

In recognition memory tasks, output interference (OI) effects manifest as a decrease in performance over the course of a test list. While this interference effect has been shown in many different experimental contexts, it is unclear how it is influenced by different properties of study and test items. This work investigates the relationship between semantic similarity and OI in memory, comparing semantically related (animal names or emotion words) and unrelated items to better understand memory and decision processes. Because the related items share similar features and are more confusable, it was predicted that there would be a greater amount of OI for related items. The first experiment used a single item recognition task and showed OI for each stimulus type, but no difference between related and unrelated words. However, there were differences in memory bias amongst the emotion condition that make it difficult to interpret discriminability measures. To remove the potential confound of memory bias, Experiments 2 and 3 use a two-alternative forced choice version of the task. There were mixed findings for the effects of OI in these experiments, such that there were no significant effects in the emotion condition and weak to moderate effects in the animal condition. Additionally, related words only produced a greater amount of OI in the final experiment and only in the animal condition. These results show that related features for emotion words lead to an increased memory bias but no difference in OI, whereas related features for animal names lead to an increase in OI but no difference in memory bias. These results are discussed with potential implications to current theories and models of recognition memory processing.