Civil Engineering Reference
In-Depth Information
the correlation between the subjective workload and performance ratings. Both techniques showed
similar sensitivity to different tasks manipulations and task difficulty. However, both techniques were
not able to detect resource competition effects in dual tasks performance, demands of response execution
processing, and the dynamics of the difficulty changes.
Hill et al. (1992) compared four subjective workload ratings scales: NASATLX, OW, SWAT, and MCH
scale. These techniques were compared in reference to the following criteria: (1) sensitivity (measured by
factor validity), (2) operator acceptance, (3) resource requirements, and (4) special procedures. The
results showed that NASATLX had the highest factors validity (the greatest correlation with the operator
workload factor), while the OWhad the second highest average factor validity. The TLX was liked best by
the operators, and OWwas the easiest to complete. Verwey and Veltman (1996) compared sensitivity and
diagnosticity of several workload assessment methods such as SWAT, RMSE, workload secondary task,
ratings, heart rate, and eyebinks. The workload assessment techniques were compared for the short
periods of elevated visual and mental workload during driving. The results revealed that the secondary
performance technique, SWAT, and RSME were sensitive to the visual workload peaks. Secondary per-
formance technique and RSME were also sensitive to the mental workload peaks, while SWAT was less
sensitive to the mental workload peaks.
Rubio et al. (2004) evaluated psychometric properties (sensitivity, diagnosticity, and validity) of three
instruments: the NASA TLX, SWAT, and WP. Two laboratory tasks were evaluated: Sternberg's memory
searching tasks and tracking tasks. The sensitivity and diagnosticity of WP ratings were higher than
NASA TLX and SWAT ratings. The assessment of concurrent validity by correlation of ratings with
performance showed that the NASA TLX ratings had higher correlation than SWAT and WP.
Nygren (1991) made a theoretical analysis of the psychometric properties of NASATLX and SWAT. He
concluded that psychometric properties of the unidimesional scale (such as Cooper-Harper or OW)
make them less sensitive to differences in workload than either TLX or SWAT, and relatively more vari-
able than TLX and SWAT scores. According to Nygren (1991), the advantage of SWATover TLX is that it
is a psychological model of subjective judgment, which may be oversimplified, but may also reflect
cognitive mechanisms and biases that actually affect the process of mental workload judgments. The
advantage of TLX over SWAT is based on the general linear model, where six dimensions derived
from extensive multivariate analyses based on numerous studies and different workload domains.
SWAT is more relevant for empirical testing of the appropriateness of particular additive model of work-
load assessment. TLX has greater potential for solving workload problems in many applied settings, by
accurately predicting operator workload levels across a variety of tasks (Nygren, 1991). The results of
numerous validation studies indicate that MCH, Bedford scale, SWAT, and TLX procedures represent
globally sensitive measures of operator workload (Wierwille and Eggemeier, 1993). Since, both SWAT
and TLX are multidimensional scale and they can therefore provide some diagnostic information
concerning causes of workload represented by the subscales (Moroney et al., 1995).
37.5 Summary
Several authors concluded that subjective ratings scales are among the simplest and most efficient of
workload estimation instruments that can be used for ergonomics applications (Skipper et al. 1986;
Nygren, 1991; Wierwille and Eggemeier, 1993; Tsang, 2001). These scales are the most sensitive, most
transferable, and least intrusive techniques for workload estimation, are easy to administer, and
require little effort or no equipment. Several scales demonstrated global sensitivity and thus can
provide appropriate workload indicators in test and evaluation situations. Some authors stated that sub-
jective rating scales constitute the most relevant method of mental workload assessment, since” subjective
scaling is the most direct measure of such subjective experience” (Sheridan, 1980). However, the subjec-
tive rating scales have also some serious disadvantages. The source of the resource demands is hard to
introspectively diagnose within a dimensional framework. Subjects may not be able to distinguish
mental demands from other type of demands such as physical or manipulative (O'Donnell and
Search WWH ::
Custom Search