Application Of Item Response Theory On Development Of Achiement Test In Secondary School Ecnonmics In Enugu Educational Zone, Enugu, Nigeria
Bernadine Amarachi Ugwuanyi
Department of Science & Vocational Education (Measurement & Evaluation), Godfrey Okoye University (GOU), Enugu.
Genevieve Chimaoge Ebulum
Department of Psychology, University of Nigeria Nsukka
Keywords: Item Response, Achievement Test, Economics
Abstract
The study aimed at applying the Item Response Theory (IRT) on the development of achievement test in Secondary Schools in Enugu Education Zone. Four research questions guided the study. The study adopted instrumentation research design. The population for the study was 3105 SS2 students. A sample size of 1015 was obtained using multiple stage random sampling technique. The instrument for data collection of this study was Economics Achievement Test (EAT) which consists of 60 test items, face validated by three specialists, one from Economics Department and two from Measurement and Evaluation unit; all from Godfrey Okoye University (GOU), Enugu. The reliability coefficient of 0.77 was obtained using Kuder – Richardson 20 (K-K20). A total of 1005 out of 1015 copies were correctly filled and retrieved immediately and were used for the study. The data collected on the study were analyzed using BILPG – MG for IRT analyses including probability level, item goodness of fit and others. The results from findings revealed that 95% of the multiple choice test in Economics had high reliability in measuring the students’ ability. Fifty eight percent (58%) of the multiple choice test in Economics fitted the two parameter (2PL) model because the items were equal and above 0.05 level of significance. Educational Implications of the findings of the study show that: IRT is a reliable measure and essential to the field of Economics as well as, to the study of latent abilities of students, IRT is a valid methodological framework for modeling response data from assessments in Economics, A didactic application of IRT highlights some of these advantages for psychological test development. Often, classical methods exclusively are used to offer evidence of validity and reliability to new tests and this evidence is undoubtedly important. But the findings of this study provide evidence that IRT results can be extremely helpful to complement this evidence with information regarding the quality of the measurement at specific points of the measuring scale