This dissertation examines how different emotion measurement approaches—self-report, other-report, and artificial intelligence (AI)-based emotion detection—relate to one another and predict outcomes in employment interview contexts. Participants completed structured mock interviews in which emotions were assessed using self-report ratings, observer ratings, and AI-based analyses of facial expressions, vocal tone, and language. Interview outcomes included performance, likability, and social skills ratings. Results indicated limited convergence across emotion measurement approaches, with correlations among self-report, other-report, and AI-based emotion measures largely nonsignificant. Factor analyses further suggested that emotion indicators clustered partly by measurement source rather than reflecting shared underlying emotional constructs. Observer-rated emotions were the most consistent predictors of interview performance, likability, and social skills. In contrast, neither self-reported emotions nor AI-based emotion measures were significantly related to the interview outcomes. These findings suggest that different emotion measurement approaches capture different aspects of emotional experience and that emotions perceived by others may be most relevant for predicting outcomes in evaluative social contexts such as interviews.
