Measuring Enjoyment, Anger, and Anxiety During Teaching : Development and Validation of the Teaching Emotions Scales (TES)

Abstract

Objectives. Teachers’ emotions during teaching are critically important for classroom instruction as well as teachers’ well-being and professional careers (Author, 2014; Schutz & Zembylas, 2009). However, research on teacher emotions is slow to emerge, and self-report scales to assess discrete emotions with respect to teaching are largely lacking. In the present research, we constructed English and German language versions of scales measuring three emotions considered most relevant in the context of teaching: enjoyment, anger, and anxiety (Author, 2014; Sutton & Wheatley, 2003). The study aimed to examine (a) the internal and external validity of the scales; (b) the level of situational specificity at which teacher emotions should be measured (teaching in general vs. teaching a specific group of students); and (c) their measurement invariance across the two cultural contexts (North American vs. German teachers).Method. Samples included N = 944 teachers from elementary and secondary schools in Canada and Germany. Teaching enjoyment, anger, and anxiety were assessed with four items each. There were two versions for each scale. One pertained to emotions during teaching students generally (e.g., “I generally enjoy teaching” for enjoyment, “I often have reasons to get angry when teaching” for anger, “I am tense and nervous when teaching” for anxiety), and the second pertained to emotions during teaching a specific class (e.g., “I generally enjoy teaching these students”). Social desirability, general affect, emotional exhaustion, job satisfaction, teacher self-efficacy, and student ratings of teachers’ classroom instruction were assessed to validate the scales.Results and Discussion. For all three emotions, results showed that the scales were sufficiently internally consistent, and that class-specific scales should be preferred over general scales. This was demonstrated applying CFA, comparing the model fit of six different models: A single-factor (general emotionality), a two-factor solution (positive vs. negative affect), and a three-factor solution (enjoyment, anger, and anxiety) for both the general and the class-specific scales. Clearly, across cultural contexts, the three-factor solution using class-specific scales had the best fit (CFI = .96/.98, RMSEA = .081/.052 in the German/Canadian samples, respectively). Furthermore, across cultural contexts, results suggested that answers to the emotion scales were not biased by social desirability and showed sufficient convergent and divergent validity in terms of relations with general positive and negative affect, and emotional exhaustion.Interestingly, only anxiety, but not enjoyment or anger, were correlated significantly with job satisfaction and teacher self-efficacy, across cultural contexts. In addition, consistent relationships existed with student ratings of teaching behaviors. Finally, measurement invariance analyses revealed that loadings and residual variances were largely invariant across the two language versions, but some intercepts and covariances varied across the two samples. This implies that the TES will not be suitable for mean-level comparisons across English versus German samples but overall, both language versions of the presented teacher emotion scales are highly valid and reliable. With the TES, we hope to present a sufficiently economic, objective, valid and reliable instrument that proves helpful in diagnosing maladaptive emotional patterns and stimulates the design of interventions for optimizing teachers’ emotional lives and reducing their emotional exhaustion.