ASSESSMENT OF STANDARD ERROR ESTIMATES OF ITEM RESPONSE THEORY PARAMETERSOF BASIC SCIENCE IN BASIC EDUCATION CERTIFICATE EXAMINATION IN TARABA STATE, NIGERIA

Abstract

This study examined the Standard Error Estimates of Item Response Theory parameters of Basic Science in Basic Education Certificate examination in Taraba State, Nigeria. The study adopted a descriptive research design of survey type. The population of the study was 19,455 Upper Basic III Students, out of which a sample of 622 students was selected using proportional random sampling, from a total of 304 public Upper Basic Schools in Taraba State. The study adopted three sampling techniques at different stages in selecting the schools and students. Data were obtained using 2021 and 2022 Basic Education Certificate Examination Basic Science multiple choice test items developed by the Taraba State Education Resource Centre. The instrument was validated by three experts from Faculty of Education Taraba State University, Jalingo. The reliability of the examination was determined using Split-half formula which yield coefficients of 0.758 and 0.884 respectively. Data collected were analyzed using IRTPRO 4.2 to determine the Standard Error of Estimates parameter indices using the Three Parameter Logistic Model. All hypotheses were tested at 0.05 significance level using an independent sample t-test. Findings showed that not less than 70% of the items were considered good as having their parameters SEEs which lie within the acceptable limit. It was concluded that Basic Science BECE questions for the years 2021 and 2022 were good based on the standard error of estimates. It was recommended among others that items that failed to meet the minimum threshold value of SEE should be eliminated or revised by the examination body.