Estimating capacity limitations in ensemble averaging using set size manipulations and individual differences in search performance

When confronted with a group of stimuli which are similar along one or multiple feature dimensions, the ability to extract a statistical summary representation (SSR) of the group is thought to be one way of decreasing cognitive load. The broad set of processes by which this summarization occurs is known as ensemble coding. Predominant theories of ensemble coding state that the SSR is calculated before selective attention can take hold, thus the process utilizes each stimulus in the group. Other theories state that attention may select a subsample of stimuli from which to form the SSR. This subsample would be subject to the capacity limits of selective attention. To investigate these possibilities, we tested capacity limits by manipulating set size on two visual search tasks and an ensemble averaging task, using similar stimuli in each. This manipulation allowed us to view potential set size effects on ensemble averaging as well as compare individual differences in set size effects for the search tasks and the ensemble averaging task. We established evidence of a relationship between individual differences in search performance and ensemble averaging performance. These findings suggest that SSRs may not be calculated preattentively but are influenced by attentional selection of stimuli.