The effect of computer-displayed knowledge of results on a two-variable optimization task

Twenty-five subjects were tested on a two-variable optimization task. In the task, subjects had to find the maximum of an unseen bivariate-normal ‘hill’ by choosing a pair of co-ordinates and receiving feedback on the height achieved. Five levels of displayed knowledge or results were tested, ranging from the current value only through the previous three, five or seven values to all values in the current trial. Restricting information below seven previous steps reduced performance. Using a computer for execution of the task in place of the pencil-and-paper mode (tested in an earlier paper) had no effect on performance.