Pillars of judgment : how memory abilities, task feedback, and cognitive load guide judgment strategies

Abstract

Making judgments is an essential part of everyday life and how people form a judgment has instigated a plethora of research. Research in judgment and categorization has particularly contrasted two types of judgment strategies: rule-based and similarity-based strategies. Recent research suggests that people can make use of both rule- and similaritybased strategies and frequently shift between these strategies. To select between strategies, contingency approaches propose that people trade off the strategies’ accuracy against the effort needed to execute strategy so that the selected strategy matches the demands of the task environment and the capabilities of the decision maker. This dissertation presents three papers investigating how accuracy-effort trade-offs between rule-based and similarity-based judgment strategies change strategy selection in judgment and categorization tasks. The first paper studies how reducing working memory by imposing a cognitive load may foster shifts to a less demanding similarity-based strategy and, in turn, enhances judgment performance in tasks well solved by a similarity-based strategy, but not in tasks for which rules are better suited. The second paper compares judgment strategies to strategies people apply in categorization. It shows that the same task characteristics, namely the number of cues and the functional relationship between cues and criterion, foster shifts between rulebased and similarity-based strategies in judgment and categorization. The third manuscript explores which memory abilities underlie rule-based and similarity-based judgments. Specifically, it shows that working memory predicts to a stronger degree how well people solve rule-based judgment tasks, whereas episodic memory is more closely linked to judgment performance in similarity-based tasks. Furthermore, episodic memory also predicts selecting a similarity-based strategy, but not working memory.