Reasoning and memory : multiple simple response strategies are used in visual working memory for color-orientation binding
Recent research has examined the how effectively people use the information in working memory (WM) when reasoning about that information is required in a recognition memory task (Chen and Cowan, 2013; Hardman and Cowan, 2016). In the featurematching task, a WM task, deductive reasoning (Deduction) can be used to determine that knowledge of one binding between colors and orientations (blue/left) precludes a different binding (blue/up), given that each color can only be paired with one other orientation. Hardman and Cowan (2016) found that the use of Deduction with the contents of WM is attempted about half of the time. In the reasoning-only task, which is very similar to the feature-matching task but lacks the requirement that items be stored in WM, however, Deduction was used accuracy and often, even by young children (Hardman and Cowan, in prep). Thus, there is a discrepancy in how people use Deduction depending on the task constraints. In this study, we sought the cause of this Deduction Discrepancy. We hypothesized that the Deduction Discrepancy may be due to (a) the requirement of WM maintenance in the feature-matching task, but not the reasoning-only task; (b) incorrect psychometric modeling results; or (c) limits the ability to use Deduction due to WM load. We examined these possibilities with novel tasks and psychometric models. Our results provided partial support for the first and third hypotheses, but not to an extent that we can fully explain the Deduction Discrepancy. New hypotheses and future directions are discussed.