Evaluation of three methodologies for assessing work activity during computer use

Abstract

The overall goal of this study was to evaluate three separate methodologies for gathering work activity information among computer users. These methodologies included worker self-report, work sampling, and activity monitoring. A repeated measures design was employed whereby data were collected simultaneously on each subject (n = 51) across three consecutive workdays. Exposure information gathered included keying time, mouse usage, and time spent performing various work tasks (i.e., writing, proofreading, handling documents). Subjects were recruited to represent a wide range of keyboard activity and mouse usage. The study found that worker self-reports overestimated actual keyboard usage by a factor of approximately 1.5 for workers using the keyboard an average of 4 hours per day to a factor of 4 for workers using the keyboard an average of 30 min per day. On average, there was an approximate twofold difference between worker self-reported keying time and that obtained via activity monitoring and work sampling. This trend was similar with regard to time spent using the computer mouse. Worker self-reported mouse usage was approximately twofold higher than that obtained via activity monitoring or work sampling. Self-reported exposure information not only resulted in different estimates, but showed greater variance compared with the other methodologies. The results of this study suggest that the use of worker self-reported exposure information on keying time and mouse usage may not represent an accurate account of time spent performing these tasks. In the context of epidemiological studies work sampling and/or activity monitoring would be more suitable methodologies for obtaining such information.