Thomas E. Nygren

The Subjective Workload Assessment Technique: A Scaling Procedure for Measuring Mental Workload

Mental workload is proposed to be a multidimensional construct that can be largely explained by three component factors: Time Load, Mental Effort Load, and Psychological Stress Load. In this paper, we describe a subjective scaling approach, the Subjective Workload Assessment Technique (SWAT), that captures this multidimensional nature of mental workload. We describe the SWAT procedure as a two-phased method that includes (a) a scale development phase based on conjoint measurement and nonmetric scaling, and (b) an event scoring phase. The development of SWAT and its measurement foundations are discussed. Recent research illustrating SWATs widespread utility and its sensitivity as a measure of perceived mental workload is summarized.

Psychometric properties of subjective workload measurement techniques: implications for their use in the assessment of perceived mental workload

This paper examines fundamental measurement issues that have largely been ignored in scaling techniques for assessing perceived mental workload. It is argued that both theoretical and practical advances in workload assessment have been unnecessarily slowed by a lack of concern for measurement issues. Typical of this impedance are some recently reported “empirical results“ comparing various workload assessment techniques which actually follow directly from some misunderstood statistical properties of the scales. The commonly used subjective workload assessment procedures have distinct psychometric properties that make each procedure sensitive to different kinds of workload measurement. Examination of the properties for the National Aeronautics and Space Administration Task Load Index (NASA TLX) support its potential as a general prediction model for experienced workload and for the Subjective Workload Assessment Technique (SWAT) as a cognitive model sensitive to individual differences.

Reacting to Perceived High- and Low-Risk Win–Lose Opportunities in a Risky Decision-Making Task: Is It Framing or Affect or Both?

Affect research has suggested that in high-risk situations, a positive mood often results in an enhanced sensitivity to losses, leading to strong risk-averse behavior relative to neutral or negative mood, but when a situation is seen as being low risk, a reversal occurs and positive affect will often result in more risk-seeking behavior. It was hypothesized that the simple “framing” of a gambling task to emphasize either potential gains or potential losses could act as an affect inducer and would produce similar results. In Experiment 1 the effects of an induced positive or negative affective state on risk-taking behavior in a gambling task were examined. Results replicated the risk-averse/risk-seeking reversal phenomenon described above. In Experiment 2 the affect conditions were replaced with a simple Winning vs. Losing framing manipulation where an instructional emphasis was placed either on accumulating points or avoiding the loss of points. Results demonstrated that a reversal pattern in risk taking like that found in Experiment 1 for affect could also be obtained via this simple framing manipulation. An affective-cognitive model of pre-choice framing and a theoretical link between the effects of framing and the effects of mood manipulation based on mood management theory are presented and discussed.

The Impact of Educational Message Framing on Adherence to Continuous Positive Airway Pressure Therapy

Although of proven health benefit to persons with obstructive sleep apnea (OSA), adherence to continuous positive airway pressure (CPAP) therapy is suboptimal, with patterns of use that are established early and that are not easily altered after the initial experience. In a randomized controlled trial, 70 participants with OSA and cardiovascular disease were assigned to receive either positively or negatively framed education about CPAP. Objective adherence was measured following 30 days of home CPAP therapy. Daytime sleepiness, dispositional optimism, self-efficacy, and depression were also evaluated at baseline and after 30 days. CPAP use was greater in the group receiving negative message framing (p = .015).

Assessing Individual Differences in Decision Making Styles: Analytical vs. Intuitive

Decision strategies are often characterized as being intuition-based or analytically-based. The use of these strategies is proposed to be associated with individual differences in propensity toward using different decision making styles. A reliable self-report measure, the Decision Making Styles Inventory (DMI), consisting of 15 items on each of three scales was constructed. The items were found to differentiate among an “analytical”, an “intuitive”, and a “regret-based” emotional decision making style. The analytical and intuitive scales were found to predict differences in performance in a complex dynamic decision making task. on a decision making subtask, a greater general reliance on an analytical decision making style was found to lead to poorer performance. Greater reliance on a more intuitive approach had no effect on this subtask, but was found to predict better performance as workload levels increased. These findings suggest that human performance may be significantly influenced when either a more intuitive or analytical decision style is used. Implications for training the adaptive decision maker are discussed.

An Individual Differences Approach to SWAT Scale Development

A refinement to the scale development phase of the Subjective Workload Assessment Technique (SWAT) provides for forming scales for homogeneous subject groups. Groups are formed by determining which of the three dimensions, time load, mental effort load or stress load subjects judge to be the most important contributor to workload. The group scales are then transformed into a SWAT scale that ranges from 0 for the lowest defined workload condition to 100 for the highest workload condition. This procedure should increase the precision of workload measurement while minimizing the effects of individual subject ranking errors.

AN OPEN-TRANSFORMED SCALE FOR CORRECTING CEILING EFFECTS AND ENHANCING RETEST RELIABILITY : THE EXAMPLE OF PAIN

The use of closed scales (with anchors at each end) to measure pain was found to produce ceiling effects characterized by a deceleration of ratings toward the upper end of the scale. This was consistent with previous research. Apart from producing nonlinear functions, the closed scale also limited test-retest reliability because of subjects’ tendencies to correct their distorted ratings in subsequent trials. However, an open-ended scale coupled with transformation of reported ratings into a decile scale virtually eliminated the ceiling effect, thus producing consistently linear functions and maximizing test-retest reliability. This finding may have implications for the measurement of other sensory and psychological phenomena, especially those in which the property evaluated varies in a continuous fashion.

An Examination of Conditional Violations of Axioms for Additive Conjoint Measurement

Axiomatic conjoint measurement methodology of fers a useful approach for evaluating different compo sition rules as potential models for fitting the compo nents of multidimensional stimuli. The usefulness of this methodology has been somewhat hindered in ap plied settings because of a lack of an adequate error theory for testing the fit of data to the axioms. This paper presents the results of an attempt to provide a basis for an examination of errors of the conjoint mea surement axioms. Specifically, this paper describes a means of evaluating the fit of an additive conjoint measurement model to a three-factor design. For each of the critical axioms of axiomatic conjoint measure ment, the proportions of errors that would be expected by chance for different levels of satisfaction of the simple independence property are examined. The re sults indicate that violations of these axioms occur much less often than intuitively might be expected. Error proportion tables based on monte carlo analyses are presented to aid in comparisons with empirically obtained results. It is shown that two types of viola tions of the axioms can be defined and used to differ entiate between systematic and unsystematic errors in fallible data.

A two-stage algorithm for assessing violations of additivity via axiomatic and numerical conjoint analysis

An algorithm for assessing additivity conjunctively via both axiomatic conjoint analysis and numerical conjoint scaling is described. The algorithm first assesses the degree of individual differences among sets of rankings of stimuli, and subsequently examines either individual or averaged data for violations of axioms necessary for an additive model. The axioms are examined at a more detailed level than has been previously done. Violations of the axioms are broken down into different types. Finally, a nonmetric scaling of the data can be done based on either or both of two different badness-of-fit scaling measures. The advantages of combining all of these features into one algorithm for improving the diagnostic value of axiomatic conjoint measurement in evaluating additivity are discussed.

Limitations of Additive Conjoint Scaling Procedures: Detecting Nonadditivity When Additivity Is Known to Be Violated

Two sets of three-outcome gambles were construc ted to vary factorially along the factors Amount to Lose, Amount to Win, Probability of Losing, and Probability of Winning. Single stimulus ratings of attractiveness and risk were obtained for each of the constructed gambles from 19 subjects. In addi tion, paired comparison strength of preference and difference in risk judgments were obtained for a subset of these gambles. Two additive conjoint scal ing procedures, Carrolls (1972) MDPREF and Johnsons (1975) NMRG, were used to generate pre dicted paired comparison preference and risk judg ments from the single stimulus ratings for each subject. These predictions were then compared with the observed paired comparison judgments. Results indicated that although the goodness-of-fit mea sures associated with each of the scaling models in dicated that the subjects data were being fit very well by the additive models, additivity among the payoff and probability factors was clearly violated. A procedure for detecting nonadditivity is outlined and illustrated with the data. The limitations of using these additive conjoint scaling procedures as predictive techniques when additivity is violated are shown and their implications are discussed.