Abstract
Mean Computer Subsystem (CSS) response times and task type were studied to determine their effect on operator performance in an interactive computer system. The mean CSS response times were categorized into four groups; “zero” seconds, “low” (4.0 to 7.9 sec.), “medium” (16.8 to 17.3 sec.), and “high” (24.2 to 30.3 sec.). A simple data entry task and a data correction task were used to study the task type effect.
Five directory assistance operators were used as subjects. Each operator performed the tasks using a Teleray 3700 CRT terminal over a 5-day period. The operators experienced every CSS response time condition for about one-and-a-half hours.
An analysis of typographical errors revealed a significantly lower error rate on the data correction task than on the data entry task. Overall speed of performance results showed that all operators performed significantly faster during the data correction task. Improvement of overall typing rate was observed in the data correction tasks with increased response time. The data also were analyzed to study the “search and preparation” time that the operator used prior to entering data on the terminal. This analysis revealed that operators efficiently utilized the CSS response time to search for errors and to prepare to enter data. The analysis also showed that increased delays in CSS response time caused no identifiable reduction in typing speed once the operators began entering data on the terminal.
Utility Indices were calculated to estimate the relative importance of each independent variable in influencing the operators' performance. The calculated indices indicated that mean CSS response time had only a small effect on performance of the operators under the conditions studied. Within the limits of this study, the data suggest that operator performance is better controlled by paying close attention to the design of the operators' tasks than by providing short CSS response times.
Get full access to this article
View all access options for this article.