NASELA PILA
Department of Educatoona Foundations, Faculty of Education,
Benue State University, Makurdi
and
J. G. ADEWALE
Institute of Education, University of Ibadan, Ibadan
CONSTRUCT VALIDITY AND RELIABILITY OF BASIC EDUCATION CERTIFICATE EXAMINATIONS IN NATIONAL VALUES EDUCATION IN BENUE STATE – NIGERIA
This study assessed construct validity and reliability of Basic Education Certificate Examinations (BECE) in national values education (NVE) for 2017. Test bias with respect to testees’ gender and school location was used to detect construct validity. Ex-post facto research design was adopted while Item Response Theory was employed for reliability estimation. Two research questions were answered. The study population was Junior Secondary School (JSS3) students in Benue State. Eight local government areas from two senatorial districts (4 from each) were randomly sampled; 1500 JSS3 students from48 secondary schools across rural and urban settlements were selected using stratify sampling. National-BECE NVE-2017 was the instrument used to collect data which were analysed using exploratory structural equation modelling (ESEM) and IRT-Logistics Parameter. The findings revealed that: National-BECE NVE-2017, with 2-underlying factors, had no construct validity. It was biased to testees’ subgroups (gender and school location). The two-factor model describing examinees’ performance (construct validity) in National-BECE based on gender and school location differences was viable. Two-underlying factors have marginal reliability coefficient: (F1) r=0.8550695,(F2) r=0.6022473 implying that only F1 is highly reliable and indicating high internal consistency of items. It was recommended that public examining bodies should establish test construct validity and reliability before administration.
Keywords: construct validity, reliability, national values education,
NASELA PILA Department of Educatoona Foundations, Faculty of Education, Benue State University, Makurdi and J. G. ADEWALE Institute of Education, University of Ibadan, Ibadan
CONSTRUCT VALIDITY AND RELIABILITY OF BASIC EDUCATION CERTIFICATE EXAMINATIONS IN NATIONAL VALUES EDUCATION IN BENUE STATE – NIGERIA
This study assessed construct validity and reliability of Basic Education Certificate Examinations (BECE) in national values education (NVE) for 2017. Test bias with respect to testees’ gender and school location was used to detect construct validity. Ex-post facto research design was adopted while Item Response Theory was employed for reliability estimation. Two research questions were answered. The study population was Junior Secondary School (JSS3) students in Benue State. Eight local government areas from two senatorial districts (4 from each) were randomly sampled; 1500 JSS3 students from48 secondary schools across rural and urban settlements were selected using stratify sampling. National-BECE NVE-2017 was the instrument used to collect data which were analysed using exploratory structural equation modelling (ESEM) and IRT-Logistics Parameter. The findings revealed that: National-BECE NVE-2017, with 2-underlying factors, had no construct validity. It was biased to testees’ subgroups (gender and school location). The two-factor model describing examinees’ performance (construct validity) in National-BECE based on gender and school location differences was viable. Two-underlying factors have marginal reliability coefficient: (F1) r=0.8550695,(F2) r=0.6022473 implying that only F1 is highly reliable and indicating high internal consistency of items. It was recommended that public examining bodies should establish test construct validity and reliability before administration. Keywords: construct validity, reliability, national values education,