Visual Probability Monitoring: Effects of Display Load and Signal Discriminability

Abstract

The Criterion Task Set (CTS) is a standardized battery of loading tasks developed for application to a variety of performance assessment problems, including evaluations of workload metrics and the effects of stressors on performance. One feature of the CTS that facilitates its application is a set of guidelines that specify standard loading levels for manipulation of difficulty on each task. This standardization is necessary for comparative evaluations of workload metrics and stressors, and therefore represents an essential feature of the CTS. The baseline version of the CTS includes nine primary loading tasks that represent a range of information processing functions. Perceptual input and stimulus identification functions are represented in the CTS by the Probability Monitoring (PM) task, which requires that subjects detect the occurrence of signals on a visual display. The PM display includes a variable number of dials, each consisting of a pointer that moves randomly with respect to a center reference mark. When a signal occurs, pointer movement becomes nonrandom or biased, such that a disproportionate number of moves occurs on one side of the reference mark. Difficulty of the task can be manipulated by varying the display load (number of dials) or the discriminability of the signals (Shingledecker, 1984). Signal discriminability is varied by changes in the percentage of movements that occur on the biased side of the center reference. Previous work has indicated that a 95% bias can be more readily detected than an 85% bias, and that the latter is more rapidly detected than a 75% bias. Current work with the CTS involves continued evaluation and refinement of tasks from the baseline version of the battery. The present work evaluated the effect of display load and signal discriminability in an updated version of the PM task. The new version incorporated a faster pointer movement rate (5 moves per second) than had been used in the previous task. This change was implemented to permit inclusion of more signals per experimental session than had been possible with the baseline version of the task. Since the new version differed substantially from its predecessor, the present study was conducted to evaluate the need to specify new standard loading levels for the task.