Abstract
Recently, the Subjective Workload Assessment Technique (SWAT) has become a popular tool for assessing mental workload. SWAT attempts to estimate an individual's internal representation of mental workload from a rank ordering of combinations of mental workload factors. This paper presents an experiment in which these internal values were simulated on a computer. Conjoint analysis was performed on rank orders produced from the internal values with and without noise added. Results indicated that conjoint analysis was able to accurately estimate the “true” rankings and values, even when noise was present. In addition, linear regression was found to be a poor estimator of the “true” values, while a simpler technique, rank summing, was found to be as accurate an estimator as conjoint analysis.
Get full access to this article
View all access options for this article.