There is a crisis of measurement in memory research, with major implications for theory and practice. This crisis arises because of a critical complication present when measuring memory using the recognition memory task that dominates the study of working memory and long-term memory (“did you see this item? yes/no” or “did this item change? yes/no”). Such tasks give two measures of performance, the “hit rate” (how often you say you previously saw an item you actually did previously see) and the “false alarm rate” (how often you say you saw something you never saw). Yet what researchers want is one single, integrated measure of memory performance. Integrating the hit and false alarm rate into a single measure, however, requires a complex problem of counterfactual reasoning that depends on the (unknowable) distribution of underlying memory signals: when faced with two people differing in both hit rate and false alarm rate, the question of who had the better memory is really “who would have had more hits if they each had the same number of false alarms”. As a result of this difficulty, different literatures in memory research (e.g., visual working memory, eyewitness identification, picture memory, etc) have settled on a variety of distinct metrics to combine hit rates and false alarm rates (e.g., A’, corrected hit rate, percent correct, d’, diagnosticity ratios, K values, etc.). These metrics make different, contradictory assumptions about the distribution of latent memory signals, and all of their assumptions are frequently incorrect. Despite a large literature on how to properly measure memory performance, spanning decades, real-life decisions are often made using these metrics, even when they subsequently turn out to be wrong when memory is studied with better measures. We suggest that in order for the psychology and neuroscience of memory to become a cumulative, theory-driven science, more attention must be given to measurement issues. We make a concrete suggestion: the default memory task should change from old/new (“did you see this item’?”) to forced-choice (“which of these two items did you see?”). In situations where old/new variants are preferred (e.g., eyewitness identification; theoretical investigations of the nature of memory decisions), receiver operating characteristic (ROC) analysis should always be performed.

When we ask people to hold a color in working memory, what do they store? Do they remember colors as point estimates (e.g. a particular shade of red) or are memory representations richer, such as uncertainty distributions over feature space? We developed a novel paradigm (a betting game) to measure the nature of working memory representations. Participants were shown a set of colored circles and, after a brief memory delay, asked about one of the objects. Rather than reporting a single color, participants placed multiple bets to create distributions in color space. The dispersion of bets was correlated with performance, indicating that participants' internal uncertainty guided bet placement. Furthermore, relative to the first response, memory performance improved when averaging across multiple bets, showing that memories contain more information than can be conveyed in a single response. Finally, information about the item in memory was present in subsequent responses even when the first response would generally be classified as a guess or report of an incorrect item, suggesting that such failures are not all-or-none. Thus, memory representations are more than noisy point estimates; they are surprisingly rich and probabilistic.

Previous research shows that a single visual working memory item can guide visual attention towards objects that match the feature held in mind, but the results are mixed as to whether this attentional guidance occurs for multiple concurrently active working memory items or not. Evidence in favor of a single-item guidance account has been taken as evidence for the structure of working memory comprising multiple distinct states where one item is prioritized over all others by being placed within a special focus of attention. The present study was designed to test attentional guidance effects for single and multiple working memory items, and to test the hypothesis that there are special distinct states in working memory. To do so, we asked participants to remember one or two colors, then perform a visual search task, and then report the items held in mind. We demonstrate that a single working memory item robustly biases attention towards items that match the color maintained in working memory during visual search (Exp. 1). Although we found reliable guidance when participants remembered two items, we show that these effects can largely be explained by a single item guiding attention on a proportion of trials (Exp. 2). Next, by precisely measuring memory for individual items we show that items naturally vary in their representational fidelity, and that only the item with the strongest representation guides attention (Exp. 3). Importantly, we demonstrate that no special focus of attention is necessary to explain these single-item guidance effects but that natural variation in the fidelity between items — which arises through independent noise — can account for the effects (Exp. 4). These findings challenge current models of working memory guidance and propose a simpler account for how working memory and attention interact: through natural variation in the representational fidelity of memories, one item tends to be the dominant item guiding attention on any individual trial.

Social interactions are dynamic, and unfold over time. To make sense of social interactions, people must aggregate sequential information into summary, global evaluations. But how do people do this? To address this question, we conducted 9 studies (N= 1,583), using a diverse set of stimuli. Our focus was a central aspect of social interaction, namely the evaluation of others’ emotional responses. Results suggest that when aggregating sequences of images and videos expressing varying degrees of emotions, perceivers overestimate the sequence’s average emotional intensity. This tendency for overestimation, which we term the sequential amplification effect, is driven by stronger memory of more emotional expressions. A computational model further supports the memory account, and shows that amplification cannot be driven merely by perception. The current paper is the first to demonstrate amplification in social cognition of sequential information, which is especially important given the prevalence of such information in many social interactions.

“Similarity” is often thought to dictate memory errors. For example, in visual memory, memory judgements of lures are related to their psychophysical similarity to targets: an approximately exponential function in stimulus space (Schurgin et al. 2020). However, similarity is ill-defined for more complex stimuli, and memory errors seem to depend on all the remembered items, not just pairwise similarity. Such effects can be captured by a model that views similarity as a byproduct of Bayesian generalization (Tenenbaum & Griffiths, 2001). Here we ask whether the propensity of people to generalize from a set to an item predicts memory errors to that item. We use the “number game” generalization task to collect human judgements about set membership for symbolic numbers and show that memory errors for numbers are consistent with these generalization judgements rather than pairwise similarity. These results suggest that generalization propensity, rather than “similarity”, drives memory errors.