Teachers' and Learners' Emotional Experiences in Class: A Field-Based Video Study

• In Event: 25.023 - Show...and Tell? Using Behavioral Observation to Study Affective Processes in Educational Settings

Abstract

How do teachers and learners feel in real-life learning environments, and how are their emotional experiences reflected in their facial expressions? These questions can hardly be answered using solely traditional methods such as self-report surveys or manual coding of video recordings. This study aimed to test a new methodological approach to examine emotions during teaching by combining retrospective self-reports with automated frame-by-frame coding of facial expressions in a real-life teaching setting. Our analyses of frequencies of teachers’ and learners’ emotional experiences in a real-life setting are largely exploratory, while our theorizing as to covariation between facial and self-report data is rooted in a component process definition of emotions (Scherer, 2005) which conceptualizes emotions as episodes of interrelated, synchronized changes in the states of multiple organismic subsystems in response to the evaluation of events that are relevant to the organism. Subjective feelings and facial expression constitute two of these emotion subsystems. As such, we propose that teachers’ and learners’ self-reported subjective feelings during class should covary with the emotions as expressed in their faces.
We videotaped N=12 lecturers (66% female) and a random sample of their students (N=78; 80% female) with separate synchronized cameras for each participant during regular university course sessions. Participation was voluntary. A range of different subjects was covered (including psychology, education, English). All videotaped courses used direct-interactive teaching methods. The videos were recorded in full HD (1920x1080p) and 30 frames per second. Using automated facial coding software (iMotions FACET, formerly known as CERT; Littlewort et al., 2011), a total of M=67,208/61,141 frames were processed and analyzed for the teacher/student data. We thus obtained occurrence scores for positive and negative emotion on a frame-by-frame basis with an average frame detection rate of M=82/73% for the teacher/student data. We further obtained immediate retrospective self-report of teacher and student emotions using single items measuring positive emotions (joy, pride; sample item: “In the past 45 minutes, I enjoyed class”) and negative emotions (anger, anxiety, shame, and boredom) on a 5-point Likert scale.
Positive emotions were facially expressed by teachers and students 11% and 4% of the time, respectively. Negative emotions were facially expressed 12% and 8% of the time, respectively. Mean scores of self-reported positive emotions were M=3.3/2.8 (SD=.77/.75) and for negative emotions M=1.4/1.7 (SD=.40/.77) for teachers and students, respectively. Positive emotion indicators for teachers correlated substantially (r=.53), and all other congruent pairs between self-reported and observed emotions covaried positively, too, but those correlations were small in size (average r=.13). Thus only teachers’ subjective positive emotional experiences seem to become clearly evident in their facial expressions in class.
This study is the first to examine the emotional expressions of teachers and learners on a frame-by-frame level and in a real-life learning environment. The results support the applicability of our methodological approach and thus provide a promising avenue for future investigations. Upcoming analyses will involve time-series across teachers’ and learners’ state emotional experiences to explore processes of emotional contagion in the classroom.

Authors

• Anne C. Frenzel, University of Munich (LMU)

• Anton Karl Georg Marx, University of Munich

• Reinhard Pekrun, University of Essex

• Corinna Reck, University of Munich

• Mitho Mueller, University of Munich