Abstract

The study examined the quality of ADUS selection tests. Specifically, the study sought to examine the level of error of measurement in selection of faith mission selected students (FMSS) by computing and analyzing the item-difficulty parameter, item discrimination parameter and reliability level of the test items. Due to scarcity of research studies on these parameters on faith selection (ADUS selection) tests, it was worthwhile to conduct the study. A total of 1003 Standard 8 examinees out of 8569 were randomly sampled and 2015 ADUS selection test was administered to generate data (scores). Descriptive research design with cohort longitudinal survey and quantitative approach were employed. Subject Matter professionals scored scripts of the test to obtain test scores of examinees. The responses of the examinees were computed using BILOG 3.0 to obtain item-difficulty, item-discrimination and test reliability. The study revealed that ADUS selection test items were from the syllabus of Malawi curriculum. However, examinees performed very low on the test. The findings on parameters indicated that most of the items (83%) were very difficult. As for the discrimination power most items (89%) were out of desirable range of +1.0 to +2.0. The large percentage of items being poor on discrimination (power) parameter decreased drastically the quality of test. On reliability the test produced statistical value of 0.65. The test indicated low reliability since failed to achieve minimum recommended statistic value of 0.70. The study implies that the majority of items in ADUS test had poor test quality parameters and could increase errors in selection process of form one examinees. The study recommended doing ‘item analysis’ as one way to improve the quality of ADUS selection tests to minimize the level of error of measurement in selection process of faith mission selected students (FMSS).

More details

School	: School of Education
Issued Date	: 2020

Download full document