Evaluating the Factor Structure of a Teacher Motivation Scale In the Incredible Years Intervention: Using Exploratory and Confirmatory Factor Analyses

Background and Purpose: Teacher motivation to change is a key mechanism underlying the implementation and effectiveness of school-based interventions such as the Incredible Years (IY) program (Reinke et al., 2018). Understanding how teacher motivation is conceptualized and measured is critical for identifying mechanisms of change and improving intervention outcomes (Hitchings et al., 2013). Understanding teacher motivation is critical for how school-based interventions are implemented and sustained, as it influences teachers’ adoption and use of new practices and, in turn, implementation fidelity and student outcomes. However, its underlying structure remains unclear. Without evidence on whether motivation reflects a single construct or multiple dimensions (e.g., readiness vs. action), measurement may obscure meaningful differences in teacher engagement. Therefore, this study evaluated the factor structure of a teacher-reported motivation to change scale to determine whether a unidimensional or multidimensional representation is more appropriate.

Methods: Data were drawn from teachers participating in the IY project (N = 104). Exploratory factor analyses (EFA) were conducted to examine 1-, 2-, and 3-factor solutions using maximum likelihood estimation with oblique rotation, allowing factors to correlate. Initial inspection of item distributions indicated strong ceiling effects and ordinal response patterns. To address this, response categories were collapsed to reduce sparsity, and analyses were conducted using ordinal methods. Confirmatory factor analyses (CFA) were estimated using the WLSMV estimator, which is appropriate for ordered categorical data. Competing 1-factor and 2-factor models were specified based on EFA results and theoretical considerations. Model fit was evaluated using Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), and Standardized Root Mean Square Residual (SRMR), with less emphasis on RMSEA given known limitations with ordinal data. Model refinements were guided by both theoretical justification and modification indices, including the addition of residual covariances between conceptually similar items.

Results: EFA results indicated a strong general factor, with most items loading highly on a single factor and limited support for clear multidimensional separation due to cross-loadings and imbalanced factor structures. In ordinal CFA models, both the 1-factor and 2-factor models demonstrated excellent fit (CFI ≥ .99, TLI ≥ .99, SRMR ≤ .05). The 2-factor model showed statistically better fit compared to the 1-factor model; however, the correlation between the Readiness and Action factors was extremely high, indicating substantial redundancy between the factors. Additional model modifications (e.g., inclusion of residual covariances) improved model fit further but did not meaningfully reduce the high correlation between factors. Visual inspection of factor score relationships also suggested near-linear overlap between the two factors.

Conclusions and Implications: Overall, findings suggest that the teacher motivation to change scale is best conceptualized as a primarily unidimensional construct, despite some evidence of distinguishable subcomponents. The extremely high correlation between factors raises concerns about the interpretability and practical utility of separate subscale scores. These results have important implications for measurement in school-based intervention research, supporting the use of a total motivation score while cautioning against overinterpretation of multidimensional structures. Future research should continue to refine measurement approaches to better capture theoretically meaningful distinctions in teacher motivation while maintaining psychometric validity.