A well-known phenomenon of memory is the bizarreness effect which refers to enhanced memory for objects that are highly incongruent with people’s prior expectations. This phenomenon was recently extended into the visual domain of color, showing enhanced memory for objects paired with expectation-incongruent (or bizarre) colors. Here, we explore whether the enhanced memory for bizarre/expectation-incongruent objects extends to memory for the object-color binding and whether this binding is well-preserved long-term. Using a 4-Alternative forced choice task, we assessed memory for object colors as a function of expectation-congruency on one day and 3 days later. Our results revealed no significant difference in recognition memory for bizarre colors compared to expectation-congruent colors, and no enhanced memory for bizarre colors in long-term memory. These findings highlight conditions where the enhanced memory for expectation-incongruent information is limited, providing an interesting challenge to current mechanistic accounts of memory for expectation-related information.

How might surprise influence memory and learning? Isolating an item from an established category induces surprise andresults in better memory. However, it is less clear whether the degree of induced surprise correlates with better memory,or whether – regardless of degree –surprise simply triggers a uniform improvement in memory. To investigate whether thedegree of surprise has an incremental effect on memory outcomes, we gave 158 participants lists of words, varying thedegree to which a single word in the list surprisingly conflicted with the lists overarching category. Although there wasan overall boost in learning for surprising words, we found no evidence of an effect of amount of surprise on memory.Lack of evidence is not evidence of a lack, however these results provide some suggestive evidence for a threshold modelof memory enhancement from surprise. Distinguishing these accounts has important implications for affective models ofmemory and learning.

Expectations play a critical role in children's learning. Prior studies suggest that children selectively focus on and better remember details of expectation-violating events (Stahl & Feigenson, 2017; 2019). Yet, it remains unclear whether this enhanced memory persists across varying degrees (e.g., somewhat vs. very surprising) and types of expectation violations (core-knowledge vs. schema-based violations). Adapting a surprise storybook paradigm from Foster and Keane (2019), we measure children's (5-8 years; N=20) surprise and recognition memory for six stories that span different expectation-related domains and contain outcomes that are expectation-congruent, somewhat expectation-violating, or completely violating. While preliminary data revealed no significant difference in recognition accuracy by level of surprise, a trend towards better memory for violations of well-entrenched versus schema-based expectations was observed. This preliminary work points to potential differences in how varying types of expectations influence memory in young children and has important implications for learning.

Previous research evaluating the influence of category knowl-edge on memory found that children, like adults, rely on cat-egory information to facilitate recall (Duffy, Huttenlocher, &Crawford, 2006). A model that combines category and targetinformation (Integrative) provides a superior fit to preschoolersrecall data compared to a category only (Prototype) and targetonly (Target) model (Macias, Persaud, Hemmer, & Bonawitz,in revision). Utilizing data and computational approaches fromMacias et al., (in revision), we explore whether individual andage-related differences persist in the model fits. Results re-vealed that a greater proportion of preschoolers recall was bestfit by the Prototype model and trials where children displayedindividuating behaviors, such as spontaneously labeling, werealso best fit by the Prototype model. Furthermore, the best fit-ting model varied by age. This work demonstrates a rich com-plexity and variation in recall between developmental groupsthat can be illuminated by computationally evaluating individ-ual differences.

A growing body of research suggests that infants and children are sensitive to signals of information gain. However, thevalue of a piece of information may also change as the learner knows more. How do changes that occur naturally inchildrens intuitive theories contribute to their subsequent learning? Here we tested whether children who are at differentstages of understanding an intuitive theory also differ in their interest in acquiring more information in the same domain.We tested childrens performance in three distinct domains, including intuitive biology, psychology, and beliefs about psy-chosomatic events. We found that children at earlier stages of their intuitive theories were more likely to seek informationin the related domain than children with mature knowledge. These results are the first to show the relationship betweennatural changes in childrens existing knowledge and childrens future learning preferences.

In the present study, we test the theory that humans selectively encode incoming sensory information on an as-necessary basis, when the information would not be accessible otherwise, in order to compensate for cognitive limitations on the quantity of new information that they can encode. We investigate whether external informational sources—much like high-level knowledge obtained from previous experiences—can spare learners from having to encode all new information in fine-grained detail. If this is true, we would expect to observe differences between the way human learners encode new information that they know to be easily available through external informational resources (e.g., names of actors in a movie, the date of a historical event) and those that they know are not (e.g., names of new acquaintances, the date of a wedding anniversary). Specifically, we would expect learners to encode far less detail for information that is available through known external informational resources than for information that is not. We present evidence from a study run on Amazon Mechanical Turk that human memory preferentially encodes information that is not expected to be available from external informational resources.