PSYCHOMETRIC FAIRNESS OF NECO MATHEMATICS ESSAY QUESTIONS: EVIDENCE FROM DIF DETECTION, TYPE AND MAGNITUDE IN IMO STATE

DIF analysis is an important aspect of psychometric evaluation because it helps to determine whether test items are fair and unbiased across different groups of respondents. The study examined the psychometric fairness of NECO Mathematics essay questions: Evidence from DIF detection, types and magnitudes in Imo State. Four research questions and hypotheses guided the study. The study employed an ex post facto research design. The population of the study comprised 10,194 Senior Secondary School III (SSIII) Mathematics students from 310 public and private secondary schools in Imo State, Nigeria. The students were distributed across the three education zones: Okigwe (3,322), Orlu (2,694), and Owerri (4,178), and studied Mathematics as a certificate subject. A sample of 728 SSIII Mathematics students was selected using a multistage sampling procedure. The data pertaining to the research questions for this study were analysed using the Generalized Partial Credit Model (GPCM) within the framework of Item Response Theory (IRT), and null hypotheses were tested at 0.05 level of significance. Findings, among others, revealed that some 2024 NECO Mathematics essay items functioned differentially based on gender under the Generalized Partial Credit Model (GPCM). Based on the findings, it was recommended, among other things, that examination bodies, such as the National Examinations Council, should conduct routine Differential Item Functioning (DIF) analyses of Mathematics essay items before operational use to identify and eliminate biased items.