Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
352830 | Contemporary Educational Psychology | 2008 | 29 Pages |
The primary purpose of this study was to examine the domain specificity (or otherwise) of motivation and engagement across academic, sport, and music domains using a survey-based methodology with parallel forms of the Motivation and Engagement Scale. The secondary purpose of the study was a methodological one seeking to test the most appropriate analytical means of assessing domain specificity. The sample comprised 329 young sportspeople (62% of total sample; 53% male; mean age 14 years) and classical musicians (38% of total sample; 40% male; mean age 14 years) in which target academic, sporting, and music domains are salient. Confirmatory factor analysis (CFA) led to the conclusion of modest domain specificity in motivation and engagement. This conclusion was based on fit indices for competing models, the size of parallel between-domain correlations, the proportion of shared variance between parallel domains, the number of within-domain correlations that were higher than parallel between-domain correlations, the decline in model fit when parallel correlations were constrained to be equal, and differences between parallel means. Notwithstanding this, constructs that were more trait-like evinced relatively less specificity. In terms of the secondary, methodological purpose, the study found that CFAs that correlated parallel uniquenesses resolved biased parameters found in raw scale score correlations that did not correct for unreliability and CFAs that did not correlate parallel uniquenesses. Together, findings hold implications for more targeted intervention, guidance for appropriate analysis in parallel domains using parallel measures, and more differentiated approaches to measurement.