Subjective mental workload assessment technique-An introduction to NASA-TLX and SWAT and a proposal of simple scoring methods.

Abstract

The most popular and sensitive subjective mental workload assessment techniques, NASA-TLX and SWAT are introduced. However, scaling procedures to create individ ual scales such as "Paired comparisons for source of workload" in NASA-TLX and "SWAT CARD SORT" are somewhat cumbersome and complex. Furthermore, difficult and complex statistical analysis, for instance, NONMETRG and MONANOVA are necessary to calculate SWAT scores. In this paper, simple and easier new scoring methods are proposed and, traditional and new scores obtained from 108 sample data from a tracking task study are compared. 1) NASA-TLX: Rank order of raw scores for six TLX dimensions are used as weights. Weighted average of scores by means of this method (coined "AWWL": Adaptive Weighted Workload) and traditional WWL are highly correlated (r=0.964). 2) SWAT: The SWAT prototype (ex. TES, SET, etc.) is decided by asking subjects to rank three SWAT dimensions ( : time load, E: mental effort load, S: psychological stress load). Then, SWAT prototype scores (SWAT-P) are obtained from the prototype score table. Simple weighted average of three raw scores of each dimensions (equidistant SWAT score: SWAT-E) is also available. SWAT-P and SWAT-E have adequate correlations with original SWAT scores (r= 0.765 and r= 0.895). No psychometric principle were taken into account to make new scaling procedures.