Subjective time estimates of work tasks by office workers

Job information supplied by workers has many valuable uses in industry. While the reliability of self‐reports is well established in the literature, few validity studies have been reported. This study used an observational coding system to provide criterion data for assessing the accuracy of three kinds of worker self‐report: activity identification, activity rank ordering, and activity time estimation. Work activities for 36 office workers were coded by observers and compared with worker estimates obtained from a card‐sorting task and a written questionnaire. Accuracy of the self‐report activity data decreased as more quantitative information was requested: workers were best at identifying the tasks which they had performed, they were less able to judge the relative amounts of time spent in those activities, and they were least able to say how much time they had spent in any one activity. We conclude that in the analysis of work, self‐report techniques provide greater quantification only at the expense of accuracy.