Challenges and Prospects of Emotional State Diagnosis in Command and Control Environments

Abstract

The present investigation examines the correlations of emotional user states and performance in command and control environments. The aim is to gain insights into whether and how the integration of the emotional state into a diagnostic component of an adaptive human-machine system can take place. While positive valence is expected to be associated with high performance (Hypothesis 1), negative valence is assumed to be associated with low performance (Hypothesis 2). In a laboratory experiment, a command and control task in the domain of anti-air warfare was performed. Emotional valence was assessed with a software for recognition of emotional face expressions (Emotient FACET). To measure performance, we assessed performance decrements that occurred whenever a subtask was not accomplished in time. Data from 22 participants were used for the analysis (45% female, 17–52 years, M = 30.96, SD = 9.76). Regression analyses were conducted to test the hypotheses. Contrary to expectations, there were no correlations between emotional valence and performance at the group level. Even though empirical results failed to support the hypotheses, significant correlations between positive valence and performance were found for 36% of the subjects, 45% for negative valence and performance. These results indicate that individual models are necessary for the analysis of emotional user state. We discuss practical implications and suggest improvements to the paradigm.