An Experimental Study of Dynamic Human Interactions —Measurement of Individual Traits through Dynamic Human Relations—

Abstract

In order to assess dynamic personality traits that are difficult to tap by questionnaires, we attempted to analyze interaction scripts in computer simulation situations. To achieve home energy saving in a virtual family setup, subjects interacted with either demanding or non-demanding computerized counterparts (family members) regarding the use of electric appliances. We measured interpersonal cooperative actions and interpersonal amicability ratings as a time-series. Based on the correlations between the measures, we classified the subjects into two groups and extracted typical time-series trends separately in each group using the principal component analysis. The time-series trends of interpersonal amicability ratings were found to be affected both by the counterpart's manners of speaking and the cognitive coherence between them and actual behaviors. The finding that the measured dynamic individual traits did not correlate with the static traits measured by a questionnaire implied that the analysis of dynamic interactions is necessary for the study of personality.View full abstract