Application of Fuzzy Rating Methods for an Effective Subjective Assessment of Emotional States at Work

Abstract

For the quantitative assessment of mental stress for workers such as operators, it is needed to collect physiological, subjective, behavioral and environmental measures effectively, and also analyze correlations among these parameters multidimensionally. Concerning subjective evaluation methods, it is possible that constraints in measurement procedures may have caused inconsistencies in subjective rating scores. Thus, psychometric methods including fuzzy rating methods are compared by conducting three experiments. The methods investigated in this study are a category rating scale, a graphic rating scale, a fuzzy category rating scale, and a fuzzy graphic rating scale. In the first experiment, twenty subjects participate in the psychological assessment task with questionnaires. The results of the first experiment indicate that subjects prefer verbal expressions for an emotional state assessment, though the most suitable rating method differs from person to person. In the second experiment, twelve out of twenty subjects in the first experiment participate in the psychological assessment task with computerized rating scales using a system, which is developed to concurrently collect behavioral and environmental data including subjective measures triggered by physiological response patterns. The results of the second experiment indicate, again, that the most suitable rating method and preferable input device differs from person to person. Finally, a fuzzy categorical rating method using voice input is chosen as an effective means of eliminating constraints on subjects' work. In the third experiment with this method, subjective assessment data are collected from twelve subjects using voice input, by prompting a subject when some typical patterns of the estimated physiological parameters are detected. It is possible to conclude that these computerized fuzzy rating methods can be effective for subjective assessments of emotional states at work, though it is desiable to chose a suitable method by concidering the subject' a preference and characteristics of stimuli to be evaluated.