ABSTRACT
Creative problem solving has been important for the advent of new technologies. In this study, we hypothesized that subjective ratings of answers should be useful for evaluating the answer quality in creative problem solving. To test this hypothesis and extract objective indicators of the subjective ratings of answers, we evaluated the relationship between subjective ratings of task performance and behavioral and autonomic nervous activities during a creative problem-solving task performed via online conversation. The task involved an answerer and a supporter, and in the experiment, each pair performed 10 trials. The trials were categorized as highly or lowly rated according to the answerer's confidence in the answer. The task performance and behavioral and autonomic nervous activities were then compared between these categories of trials. Behavioral activity was evaluated via movements and speech activities, while for autonomic nervous activity, sympathetic nervous activity (SNA) was evaluated via skin conductance. The task performance was significantly better in the highly rated trials, whereas there were no significant differences in the behavioral activities between the highly and lowly rated trials. Moreover, in the highly rated trials, the skin conductance of the answerer was significantly high, whereas that of the supporter was significantly low. The results support the hypothesis and suggest that contrasting differences in SNA between an answerer and a supporter are indicators of the subjective ratings of answers in creative problem solving.