The Allocation of Attentional Resources in a Dynamic Environment

Abstract

Three experiments are reported in which subjects performed a primary task, whose difficulty varied quasi-randomly over time, concurrently with a secondary task, whose difficulty remained constant. Three models of resource allocation were described which varied in the degree of optimality with which operators can mobilize resources in response to the demand changes. These are quantified by linear time-series analysis. In experiment 1 employing two tracking tasks, allocation behavior was non-optimal, but indicated a trend toward optimality with practice. In experiment 2 when tracking difficulty was varied concurrently with a dynamic system monitoring task, behavior was somewhat optimal. In experiment 3 which again paired tracking and monitoring, but varied monitoring task demand, resource allocation was extremely non-optimal. Some reasons for this departure from optimality are considered.