Application of Generalizability Theory in Measurement Error in 2019 WAEC Mathematics Objective Examination in Benin Metropolis

Abstract

: This study investigated measurement error in 2019 WAEC senior secondary school examination using generalizability theory. The study was specifically concerned with identifying and analyzing measurement error in the senior secondary school 2019 WAEC mathematics objective examination using generalizability theory, and also to determine the highest contribution of facets: students, items and teachers to measurement error. Four research questions were raised to guide the study. The study was survey which adopted a random effect two-facet fully crossed s×t×i design for a generalizability (G) and decision (D) studies. The population consisted of fifty-six thousand, seven hundred and ninety-seven (5697) senior secondary three (SS3) students in the seventy-five (75) public secondary schools in Benin metropolis for the 2019/2020 academic session. The instrument for data collection was a fifty (50) multiple choice WAEC, mathematics 2019 examination. The instrument has been validated by the West African Examination Council (WAEC). The reliability of the items was ascertained using the Kuder – Richardson 20 (KR 20) to obtain internal consistency. It gave a value of 0.92. Data collected were analyzed using a software EduG version 6.0-e based on analysis of variance (ANOVA) and generalizability. The findings which emerged from the study were: the highest contribution to measurement error in examination scores was the students - teacher interaction which accounted for 68.9%, this was followed by the student factor (27.5%) and the residual, that is, interaction of student, teachers and items (3.6%). A generalizability coefficient of 0.97 high enough to rank order students according to their relative abilities in examinations was obtained when the number of teachers was increased to 78. Based on the findings, it was therefore recommended that generalizability analysis should be carried out by researchers, test developers and examination bodies so as to reduce or eliminate measurement error and hence maximize reliability.