Relatedness in memory and metamemory: Benefits, costs, and beliefs

Abstract

This dissertation examines the influence of semantic relatedness on both memory and metamemory. Related items tend to be better remembered than unrelated items in most memory tasks, and people are usually able to anticipate this in their memory predictions. In this dissertation, I report a novel case where inter-item relatedness produces a memory cost, specifically, in a location memory task. Despite this cost, participants predict that relatedness should be beneficial in this task, showing a misalignment between their beliefs and their performance. I then examine the mechanisms underlying the relatedness cost in location memory performance and what I refer to as the relatedness halo in metamemory. The latter phenomenon in particular is used to provide novel insights into the nature of metamemory beliefs and how they are updated in response to new information. I advance a theoretical framework for understanding beliefs as cue-dependent judgments that are constructed from multiple sources of retrieved information.