Differential Item Functing (DIF) In the Anxiety and Stress Scale on Students Academic Achievement in Chemistry Using Structural Equation Modeling (SEM) (Published)
This is a study that applied two different Structural Equation Modeling (SEM) methods for the detection of gender-and age-related item bias in the anxiety and stress scale on students’ academic achievement in chemistry (AASS). Specifically, a multi-group SEM approach was used to investigate both uniform and non uniform item bias in each subscale of AASS separately, and a multidimensional SEM approach that enabled in both subscales of AASS and with regard to both gender and age simultaneously. Results from the multi group SEM approach and the multidimensional SEM approach with regard to the detection of uniform item bias were largely consistent and generally agreed with the results of the ordinal logistic regression, item response theory (IRT) and contingency tables methods reported by Cameron et al 2014; Ogidi and Iweka 2015; and Iweka 2017. Inspection of parameter estimate’s of the final model showed that there was a significant positive association between Anxiety and stress (r=0.83, P<.001) indicating that symptom severity with respect to Anxiety goes together with symptom severity with respect to stress. There was a significant negative association between Age and Anxiety (r.024) P<.001), indicating that older students score lower on Anxiety than younger students.