Abstract
The study ascertained the potency of two statistical approaches (Binary Logistics Regression and Mantel-Haenszel Chi-Square) in detection of deferential item functioning of standardized Economics test in Nigeria test in Nigeria. Two research questions were asked and two hypotheses were formulated to guide the study. The design of this study is an inferential survey. The populaton of the study was 4,434,979 secondary school students and the sample comprised of seven hundred and ninety eight (798) students. The sample comprised of seven hundred and ninety eight (798) students involving purposive, stratified and simple random sampling techniques. The instruments for data collection is a 50 item WAEC General Economics Paper I Multiple Choice Test. Two statistics were used to test the hypotheses, binary logistic regression an Mantel-Haenszel Chi-Square statistics at 0.05 level of significance. Result of the analysis revealed among others that; Binary Logistic Regression and Mantel-Haenszel Chi-Square statistics have equal potency in the detention of gender DIF with 13 items; but 8 items were consistent with the two statistic as having having gender DIF while 5 items were not consistent. It was concluded that Binary Logistics Regression Statistics has a higher potency in detecting Binary Logistics Regression Statistics has a higher potency in detecting location DIF with 9 items than Mantel-Haenszel Chi-Square Statistics that detected 8 items; but 6 items were consistent with the two statistics as having location DIF while the other items were not consistent. Based on the findings, the researcher recommended that; when detecting location DIF, the use of Binary Logistics Regression Statistics should be encouraged since it detected more DIF in this regard.