Experimental investigation of stress-inducing properties of system response times

System response times are regarded as a major stressor in human-computer interaction. In two earlier studies short (2s) and long (8s) response times were found to have differential effects on psychological, subjective, and performance variables, but results did not favour either response time. Therefore, in another laboratory study with 48 subjects in four independent groups working at a stimulated computer workplace, system response times of 2, 4, 6 and 8s were introduced in the same error detection task as used before, during 3 training and 5 working trials of 20 min each, and the same physiological, subjective and performance measures were obtained. There were no global effects on physiological variables, possibly due to low work load as a result of missing time pressure, but subjective and performance variables clearly favoured the longer system response times. When task periods and response time-periods were analysed separately, a shift of electrodermal activity could be observed from task- to response time-periods during the course of trials in the 8s condition. This did not appear in any other condition, which points to psychophysiological excitement that develops when system response times are too long, thus providing support for the concept of optimal system response times.