Task complexity and serial performance under steadily increasing input rates.

Abstract

Twenty-seven subjects were trained on a multichannel information processing task, under three levels of stimulus complexity, and tested in these conditions in each of three 5-min. sessions of rapidly increasing input rate. Correct responses, errors, and omissions were examined for each condition. During testing, performance decrements which followed initial performance peaks occurred at lower input rates were least severe, and recovery was quickest when stimulus complexity was greatest. Subjects adapted to high input rates by filtering and making omissions rather than incorrect responses. In addition two types of attention strategies were identified, one of which produced superior performance at low input rates, and the other superior performance at high input rates. Error and response rates were positively correlated for all code complexity levels when pooled across subjects. However, at the most complex level this relation only held for one of the two attention strategies when the two sets of data were analyzed separately.