An investigation of item biasness of early reading literacy test developed for pre-school children in Oyo State

Abstract

Fairness and equity of test items across sub-groups of pre-school children is very important as vital decisions on policy, placement, certification, intervention, scholarship award and the likes are arrived at depending on the pre-school children's scores for the test. It is very imperative for test developers to avoid test items that might function differently across sub-groups. This study empirically investigated Differential Item functioning of Developed Early Reading Literacy Test among pre-school children in Oyo State. The study advanced four research questions, where the underlying dimensions and test biasness in terms of gender, school type and location were investigated. The research type is instrumentation while counterbalance design was adopted. Five hundred and sixty pupils were drawn using multi stage sampling procedures from forty (40) pre-schools in twelve municipalities in Oyo State. The instrument used for data collection was self-developed Early Reading Literacy Test (emp rel =0.84). Stout Test of Essential Dimensionality (STED) was used to assess the dimensionality of the test while Item response theory and DIF method were used to detect gender, school type and location biasness. Findings from this study showed that the traits that underlie questions are unidimensional in nature and 3 1.0% of the questions were gender biased, 42.0% were biased towards school type while 47. I % were biased with relevance to school location. Thus, the author recommended that test developers should endeavor to determine biasness of their test items so as to avoid it favoring a specific sub group.