performance was better in the second half. Performance differed depending on whether the descending order (16-12-8-4) of the planes occurred during the first or second half of the task. If it was encountered during the first half, the number of planes acquired and the number acquired correctly were the lowest when 16 planes were displayed. There was a slight increase with 12 planes displayed and then a gradual decline was seen with 8 and 4 planes. In contrast, when the descending order occurred during the second half, the 16 planes displayed condition resulted in more planes accessed and accessed correctly than in any other condition. Performance than showed a gradual decline with 12, 8, and 4 planes displayed. Regarding reaction time, when the ascending order was presented before the descending order, reaction time decreased with the ascending order and increased with the descending order. If the descending order was presented first, reaction time showed a decrease for 16, 12, and 8 planes, followed by a flattening out of performance.
The poor performance when the task began with the highest number of planes on the screen may have contributed to this 3-way interaction. This does not totally account for the results. It appears that when a task progresses from easy to difficult performance improves. By contrast when a task starts as hard and becomes easier participants initially put more effort into the task but they begin to reduce their effort as the task becomes easier. Fowler ( 1980) found a similar pattern of performance with air traffic controllers. Errors tended to be more prevalent during periods of low workload that immediately followed a high workload period.
This research suggests that changes in presentation of the same task from a high to a low level of difficulty or from a low to high level of difficulty can change performance results. Additional study of this issue is necessary before drawing any conclusions.