Using the environment to predict memory performance.

Past research has shown great similarity in the patterns with which information appears in different natural environments. Most memory experiments present information in very different patterns than these natural patterns. Experiment 1 used a continuous recognition design in which subjects saw words in an order that mirrored the order in which items appeared in one data source, tweets from highly followed tweeters. The fluency with which they could recognize these words (measured as inverse efficiency) had high rank-order correlation with the environmental probability that they would occur again. The inverse efficiencies can be predicted by the reciprocal square-root law, derived as the optimal mapping between probability in the environment and speed of memory access. A generalized prediction algorithm is developed for predicting the results of any continuous recognition experiment from environmental probabilities. Experiment 2 tests predictions of this algorithm for three orders of presentation: natural as used in Experiment 1, random, and spaced wherein items occur with equal spacing and frequency. As a further test, the generalized prediction algorithm is applied to the results from another continuous recognition paradigm (Bright et al., 2022). The article discusses the challenges that must be addressed to enable environmental analyses to predict memory performance more generally. (PsycInfo Database Record (c) 2026 APA, all rights reserved).