A significant discrepancy was found between subject self-reporting and actual measurement of time duration spent using a computer and a mouse in a recent European study. Agreement was seen 16 percent of the time for mouse use and 18 percent of the time for computer use. The variation was due mostly to overestimation of time duration from the self-reporting.
For duration of computer work:
- 51 percent of the subjects overestimated their time duration by one hour
- 30 percent of the subjects overestimated their time duration by at least two hours
For duration of mouse work:
- 35 percent of the subjects overestimated their time duration by one hour
- 48 percent of the subjects overestimated their time duration by at least two hours
The authors suggest that these self-report exposure inaccuracies may actually lead to an understatement of computer related risks in epidemiologic studies. Musculoskeletal symptoms may occur at greater frequency/intensity at a lower exposure than reports indicate due to study subject overestimation of exposure.
In the same report, a second project found subjects were significantly inconsistent when asked on two occasions over a two-week interval to estimate their daily duration of computer and mouse usage. Participants provided markedly different responses 25 percent and 33 percent of the time when stating their daily duration of computer usage and mouse usage, respectively.
Study Design – Test-Retest Reliability Investigation
Subjects
Eighty-one university research department employees were recruited.
Process
An internet based questionnaire was completed twice with a two week interval. Subjects self-reported the total number of hours per day they used their computer at work. Subjects also self-reported the total number of hours per day they used their mouse use at work. The response categories were never, 0-1, 1-2, 2-4, 4-6, 6-8, and greater than 8 hours per day.
Key Outcomes of Interest
- Percentage of agreement (i.e., 0-2 hours/day initial test versus 0-2 hours/day on retest) for total computer usage
- Percent misclassified by one category (i.e., 0-2 hours/day initial test versus 2-4 hours/day on retest) for total computer usage
- Percent misclassified by two categories (i.e., 0-2 hours/day initial test versus 4-6 hours/day on retest) for total computer usage
- Percentage of agreement (i.e., 0-2 hours/day initial test versus 0-2 hours/day on retest) for total mouse usage
- Percent misclassified by one category (i.e., 0-2 hours/day initial test versus 2-4 hours/day on retest) for total mouse usage
- Percent misclassified by two categories (i.e., 0-2 hours/day initial test versus 4-6 hours/day on retest) for total mouse usage
Study Design – Validity Investigation
Subjects
Participants of this study were a completely different group – 572 office workers who were volunteers in a separate research project.
Process
Through software, the duration of computer activity was measured over a three-month period. Total computer usage was calculated from the time duration between consecutive active events (keyboarding, clicking the mouse, moving the mouse). If two consecutive active events occurred within 30 seconds, the time between the active events was considered as continuous and calculated as part of the total computer use duration. If the time between active events exceeded 30 seconds, the intervening duration was considered a break.
Total mouse usage was based on the duration between consecutive active mouse events (clicking or moving the mouse). Time was considered continuous if mouse events occurred within 5 seconds.
The mean daily duration for total computer use and mouse use was calculated.
After this information was collected by the software, subjects filled out the same questionnaire as those in the test-retest investigation (self-report of the total number of hours per day they used their computer at work; self-report of the total number of hours per day they used their mouse use at work; the response categories were never, 0-1, 1-2, 2-4, 4-6, 6-8, and greater than 8 hours per day).
Key Outcomes of Interest
- Percentage of agreement (i.e., 0-2 hours/day measured versus 0-2 hours/day reported) for total computer usage
- Percent misclassified by one category (i.e., 0-2 hours/day measured versus 2-4 hours/day reported) for total computer usage
- Percent misclassified by two categories (i.e., 0-2 hours/day measured versus 4-6 hours/day reported) for total computer usage
- Percentage of agreement (i.e., 0-2 hours/day measured versus 0-2 hours/day reported) for total mouse usage
- Percent misclassified by one category (i.e., 0-2 hours/day measured versus 2-4 hours/day reported) for total mouse usage
- Percent misclassified by two categories (i.e., 0-2 hours/day measured versus 4-6 hours/day reported) for total mouse usage
Study Strengths and Weaknesses
1) Strengths – large sample size; evaluation of reliability and validity in one paper; use of accepted measuring method.
2) Weakness – definition of “computer use at work” may have been confusing to some subjects when estimating time duration (is sitting in front of the computer considered “computer use at work”).
The Bottom Line – How This Applies To Ergonomists
Misclassification of risk exposure (computer use/mouse use) jeopardizes the authenticity of study associations and conclusions. In this investigation, self-reporting study subjects misestimated their computer usage 82 percent of the time and mouse usage 84 percent of the time – mostly through overestimating their exposure duration.
The authors suggest that exposure overestimating may actually lead to an understatement of computer-related risks. The use of objective measurements are encouraged for defining computer activity exposure. When using subject self-reporting, the authors advocate epidemiologic studies use a cut off of 4 hours for total computer usage and 2 hours for mouse usage.
Article Title: Test-retest reliability and validity of self-reported duration of computer use at work
Publication: Scandinavian Journal of Work, Environment & Health, 34(2), 113-119, 2008
Authors: S I Jmker, J N M Leijssen, B M Blatter, A J van der Beek, W van Mechelen, P M Bongers
This article originally appeared in The Ergonomics Report™ on 2008-06-18.