Establishing the Assessment Model for English Language Continuous Writing Component

Validation of assessment instruments has often been confined to ensuring that items in the test are a representative sample of the domain. However, structural validation of psychological instruments through multivariate statistical procedures can be used to enhance the construct validity of psychological scales by providing an empirical model of both the dimensionality of the construct itself and fidelity of the scale developed to measure the construct. The current research is an attempt to establish the assessment model of the English Language continuous writing component. The study uses exploratory factor analysis with orthogonal rotation techniques to map-out the dimensional structure the construct. Specifically, final examination composition and letter writing scores were submitted to a principal component analysis extraction method with Varimax rotation technique to arrive at a more parsimonious and theoretically meaningful language proficiency model. Items with significant loadings on the recovered components were subsequently used to name each component in the scale.