THE EFFECTIVENESS OF CLASSICAL TEST AND ITEM RESPONSE THEORIES IN SECONDARY SCHOOL GEOGRAPHY ACHIEVEMENT TEST CONSTRUCTION IN OYO STATE
Sunday O. Adedigba, Ayoka M. Olusakin, & Ngozi A. Osarenren
Department of Educational Foundations
University of Lagos Akoka, Lagos
Measurement theories are important to practice in educational measurement because they supply a background for addressing measurement problems. One of the most essential challenges is handling the Measurement Errors. A reliable theory can help in understanding the role of errors they play in measurement. Therefore, the purpose of
this study was to investigate the effectiveness of Classical Test and Item Response Theories in secondary school Geography Achievement Test construction, in Oyo State.
The descriptive survey research design was used for this study. The population for the study comprised of all SS2 students offering geography in the three educational zones of the State. The sample for the study was 1200 students. Multi stage stratified sampling technique was used to get this sample. A 100 item GAT was the instrument for the
study. The hypotheses were tested using paired sampled t- test. Results revealed that item statistics obtained from the two theoretical frameworks were comparable.
However, item statistic obtained from IRT model showed more stability than those from CTT. Moreover, for item selection process, IRT model led to deletion of fewer items than CTT model. This result indicates that test developers and public examining bodies should incorporate IRT model into their test development processes because through IRT model test constructors would be able to generate more reliable items than in the CTT model which is being employed currently in schools. It was further recommended that CTT framework could be used as a complement to IRT.
Keywords: Item analysis, Classical test theory, Item response theory, Item statistics, Item parameters.