Ibrahim Shuaibu, Aminu Idris Harbau, Yalwaji Shehu and Sani Ahmd Katsayal
Differential Item Functioning in Senior Secondary School Certificate Examination Multiple-Choice Items: Implication for Technology Adoption in large scale Assessment
This research investigated differential item functioning of Senior Secondary School Certificate Examination Multiple-choice Mathematics items in Kano state Nigeria, with a view to finding out items that exhibits differential item functioning (DIF) on gender basis. The ex-post-facto research design was adopted. The population comprised of 65,899 candidates who sat for the examination in 2017. A sample size of 1,000 students was selected using three stage a cluster sampling technique. The population was naturally divided into schools and zones; the instrument for data collection was a sixty-item multiple-choice examination constructed and administered by National Examination Council (NECO) in 2017. The quality estimates of the instrument were ensured by NECO. The data were analyzed using two DIF detection methods (i.e. logistic Regression model, and Mantel-Haenszel statistics). The major findings reveal that items exhibit gender-based uniform and non-uniform DIF with small and moderate effect sizes in NECO 2017 multiple-choice Mathematics questions. It was recommended that, NECO which is involved in development, construction and administration of large-scale examinations should put their hands on deck to prioritize subjecting tests items in to modern statistical evaluations of DIF for a better effect.
Ibrahim Shuaibu, Aminu Idris Harbau, Yalwaji Shehu and Sani Ahmd Katsayal
Differential Item Functioning in Senior Secondary School Certificate Examination Multiple-Choice Items: Implication for Technology Adoption in large scale Assessment
This research investigated differential item functioning of Senior Secondary School Certificate Examination Multiple-choice Mathematics items in Kano state Nigeria, with a view to finding out items that exhibits differential item functioning (DIF) on gender basis. The ex-post-facto research design was adopted. The population comprised of 65,899 candidates who sat for the examination in 2017. A sample size of 1,000 students was selected using three stage a cluster sampling technique. The population was naturally divided into schools and zones; the instrument for data collection was a sixty-item multiple-choice examination constructed and administered by National Examination Council (NECO) in 2017. The quality estimates of the instrument were ensured by NECO. The data were analyzed using two DIF detection methods (i.e. logistic Regression model, and Mantel-Haenszel statistics). The major findings reveal that items exhibit gender-based uniform and non-uniform DIF with small and moderate effect sizes in NECO 2017 multiple-choice Mathematics questions. It was recommended that, NECO which is involved in development, construction and administration of large-scale examinations should put their hands on deck to prioritize subjecting tests items in to modern statistical evaluations of DIF for a better effect.