“Math scares me, but English is fun!” : The domain specificity of academic emotional experiences

Abstract

Few studies have analyzed the domain specificity of students emotional experiences; yet, a clear argument can be made in favor of a domain-specific organization of emotional experiences. First, there is cumulative evidence on the domain specific organization of a number of psychosocial constructs other than emotions (e.g. self-concept, self-efficacy, causal attributions, task values, achievement goals). It can be assumed that the mechanisms underlying the domain specificity of these constructs are operating in emotional experiences as well. Second, according to Pekrun s (2000) social-cognitive model of academic emotions, control- and value-related appraisals can be assumed to be key antecedents of students emotional experiences. As both control-related and value-related constructs pertaining to academic learning and achievement are known to be structured in largely domain-specific ways, a domain-specific organization of students academic emotions can be assumed as well.Our sample consisted of N = 721 students (54% female) from grades 7 to 10. Rating scales were used to measure enjoyment, anxiety, and boredom in the domains of mathematics, German, English, and Latin. Each of the three emotions was assessed in each of the four domains in parallel ways (3 emotions x 4 domains; e.g. How much do you enjoy [domain]? ). Achievement data were obtained by students final scholastic grades. Using confirmatory factor analysis (LISREL 8.53), we performed competitive tests of a number of models implying factor structures differing in emotion- and domain-related generality. Results corroborated our assumption that emotional experiences are largely organized along domain-specific lines. Furthermore, through a multi-level analysis procedure (HLM 6.01) we found that the emotions under investigation differed in their degree of domain specificity, and that they were more domain-specific than were students grades. Enjoyment turned out to be most domain-specific, followed by boredom and anxiety. Implications for future research and practice are discussed.