Attention Does Not Improve Impaired Understanding Of Variability In Spatial Prediction

Abstract

Understanding variability of uncertain systems is often vital for decision makers, yet is habitually disregarded in favor of developing superior capability to predict most likely outcomes. One potential path to improving appreciation of variability is simply to attend more carefully to it. The present study explores a trade-off in the ability to predict average trajectories and estimate the variability in a spatial prediction task. Through instructional and task manipulations, some participants were encouraged to emphasize variability in a task that had previously shown good performance of mean prediction but poor estimation of variability. Overall estimation of variability was poor in both the prediction emphasis condition, and crucially in the variability emphasis condition. The results suggest overlooking variability is not just the result of neglecting that dimension, but rather represents a more systematic limitation in human performance. The implications of these findings are discussed.