An Exploratory Study in Measuring Interviews Through Self-Perceptions, Mutual Gaze, and Qualitative Language Analysis

Abstract

Interviews are frequently used in clinical, media, and research settings to elicit rich information, share experiences, and develop common understandings. Despite their prevalence, it is difficult to determine the objective quality of an interview because assessments usually rely on participants’ subjective post-interview impressions. To establish the basis for more objective assessments, we used two different approaches to test whether gaze behavior during interviews may serve as a useful predictor of interview quality. First, we tested whether global gaze synchrony could be reliably assessed and could positively predict interview outcomes. Second, we rated the interviews and analyzed what qualitatively was common between the higher and lower rated interviews. We corroborated these measures against two basic measures of interview quality: subjective ratings of interviews and cognitive load. We recorded gaze data using Tobii Pro Glasses 3 worn by dyad pairs while they discussed biographical experiences. Participants completed a post interview questionnaire assessing their cognitive load and subjective experiences, and interviews were judged for quality and interestingness of elicited experiences. We tested whether our derived measures were valid by assessing their technical feasibility, their basic psychometric properties, and the degree to which they predicted interview quality. Though gaze did not predict interview quality, participants demonstrated a propensity to overestimate their performance in the interview when compared with outside raters. Additionally, it seemed that gaze was significantly lower in interviews than otherwise reported in literature.