Optional stopping performance under graphic and numeric CRT formatting

Two experiments sought to determine whether, and if so how, performance varies as a function of the manner in which a progressively unfolding decision problem is displayed over time. A hurricane-tracking scenario was adapted to the optional-stopping paradigm, such that subjects elected to continue sampling information or to make a terminal decision (evacuate or stay) at successive points in the storm's development based upon cumulative evidence and future expectations. The display manipulation was minimal in that it applied only to non-predictive historical data (position information). Despite this conservative test, display format had a significant effect when time pressure was involved: subjects reached earlier and better terminal decisions under the analog (graphic) than the numerical format (Experiment 2). The differences reduced to nonsignificances under self-pacing (Experiment 1), although significant improvements were obtained by use of a simple aiding device (calculation of worst case probabilities). Results are generally consistent with Hammond's Cognitive Consistency Theory.