Predicting the future position of a moving target.

Three experiments were conducted to examine the processes underlying the prediction of the future position of a moving target. A target moved horizontally across a computer screen at constant velocity, disappearing partway across the screen, and subjects responded when they estimated the target would have passed a point on the far side of the screen, had it continued on its path. The first experiment demonstrated that visual tracking of the target is not necessary for successful position estimation. In the second experiment, the time over which the prediction was made rather than the interval for which the target was visible, the distance over which the prediction was made, or the velocity of the target, was found to affect performance. Finally, performance was not affected when markers were placed parallel to the trajectory of the target; the presence of gratings which masked portions of the target's path did not affect subjects' performance. The previous literature suggests that the spatial interval over which predictions are made is the important variable; we find that temporal factors are the major determinants of prediction.