Macias-Cervantes, Carla-Keyanna Rose. The strategic taxation of memory: preschooler and adult attention to expectation-violating events and the impact on learning of future content. Retrieved from https://doi.org/doi:10.7282/t3-vtc5-wm83
DescriptionYoung children learn a lot about the world within their first few years of life. Learning is critically intertwined with working memory—the ability to maintain and manipulate information in mind— yet children have limited working memory resources early on in development. How might we reconcile a developing working memory system with this rapid learning? One possibility is that children have some control over what kinds of information are prioritized in memory—consistent with work showing that other cognitive capacities, like attention and action, can be rationally directed (e.g. Bonawitz et al., 2012; Kidd et al., 2012). Through empirical studies with adults and children, this dissertation investigates whether learners are selective in the kinds of information they prioritize for processing in working memory, whether this comes at the cost of learning other information, and if the processing of information can be manipulated by the promise of an explanation.First, how do learners decide what information to attend to and remember for later use? Previous literature suggests that Expectation-Violating events are more likely to provide opportunities for learning, and be better remembered than Expectation-Consistent events (Stahl & Feigenson, 2015). I suggest that learners selectively attend to Expectation-Violating information as a rational learning strategy, at the cost of learning subsequent (unrelated) information. We explore this question in preschoolers using a modified surprise paradigm (Stahl & Feigenson, 2017) and in older children and adults using age-appropriate trivia tasks (Wade & Kidd, 2019). Adult behavioral data support this prediction; however, behavioral data from preschoolers’ and older children do not.
Second, how might memory for information be impacted when there is a promise of a future explanation? Explanations play an important role in learning. I propose that when learners believe that an explanation for an Expectation-Violating event will be available later, they will no longer display a deficit in subsequent learning. Results from adult behavioral data do not support this predication.
Together this work informs our understanding of how memory is influenced by different kinds of expectation-related information. These results suggest that the impact expectation-related information has on memory may be age and/or task dependent.