Abstract
Guided by emotional response theory, we examined students' emotional reactions to perceptions of classroom injustice. Undergraduates from three universities participated by completing questionnaires. Students most frequently reported procedural injustice, but experienced the most severe and most negative emotional responses to violations involving interactional justice and a combination of procedural and distributive injustice. Students felt hurt by classroom injustice, perceived it as severe, and reacted with low pleasure, and moderate to high arousal and dominance. Students associated unfairness with low emotional support from instructors and generally low emotion work with instructors. Student perceptions of injustice severity were also associated with stronger student emotional responses to classroom injustice.
Notes
[1] Tests of homogeneity (internal consistency) indicated that no error calculated between observed and expected correlations was greater than sampling error. Furthermore, all factor loadings were greater than .60.
[2] Tests of homogeneity (internal consistency) indicated that for arousal, no error calculated between observed and expected correlations was greater than sampling error. For pleasure only one error (.01; 6.67%) was greater than sampling error, but for dominance 40% of the errors were greater than sampling error (mean error = .09). When three items (dominant, powerful, and domineering) were removed, internal consistency tests indicated that no error calculated between observed and expected correlations was greater than sampling error. In addition, the items demonstrated their highest factor loadings on their theoretically specified factors, and all but one item, excited (factor loading = .34), had factor loadings above .40 on the specified factor.
[3] Tests of homogeneity (internal consistency) indicated that for emotion work and emotional valence, no error calculated between observed and expected correlations was greater than sampling error, but for emotion support, 25% of the errors were greater than sampling error (mean error = .06). After eliminating four items (“This instructor is willing to help me make decisions about academic issues,” “I can count on this instructor when things go wrong in my personal life,” “This instructor is NOT responsive to my concerns and feelings,” and “I can count on this instructor when things go wrong with school issues”), internal consistency tests indicated that no error calculated between observed and expected correlations was greater than sampling error. In addition, all items demonstrated their highest factor loadings on their theoretically specified factors, and all such loadings were greater than .40.