This study explores the factor structure of a writing assessment framework through structural equation modelling (SEM). The structure is validated for different components (Content, Organisation, Vocabulary, Grammer and Mechanics) with the help of objectives for the present study. Using data from 1,204 writing samples, a one-factor model is tested to determine a single latent construct underlies five domains of writing quality: Content, Organization, Vocabulary, Grammar, and Mechanics. A descriptive survey design was employed.” is considered as method. While model fit indices showed mixed results, the high factor loadings and explained variances support the theoretical framework of writing assessment as a unidimensional construct. The TLI (0.962) and CFI (0.981) both exceeded the conventional threshold of 0.95, indicating excellent fit. The SRMR (0.011) was well below the 0.08 threshold, also suggesting excellent fit. This study validates the idea of writing proficiency as a cohesive construct, highlighting its explanatory power. It suggests that educators can create streamlined assessment tools that focus on developing this proficiency, while still providing feedback on specific strengths and weaknesses.
Ghosh et al. (Wed,) studied this question.