INFLUENCE OF ARTIFICIAL INTELLIGENCE COMPUTERIZED MULTIPLE CHOICE ADAPTIVE ACHIEVEMENT TEST (AI-CAT) ON ACADEMIC PERFORMANCE OF UNDERGRADUATES IN LAGOS STATE UNIVERSITY OF EDUCATION, LAGOS STATE, NIGERIA

Abstract

This study examined how the use of Artificial Intelligence Computerized Multiple Choice Adaptive Achievement Test (AI-CMCAAT) influenced undergraduates’ performance at Lagos State University of Education, Lagos State, Nigeria. Two hypotheses were formulated for the study. A quasi-experimental research design, specifically a pretest-posttest control group design, was adopted. The population comprised 2,345 third-year undergraduate students at Lagos State University of Education. One-tenth of the population was used, resulting in a sample of 285 undergraduates selected through a multi-stage sampling procedure. Data were collected using the AI-CMCAAT, which covered seven selected topics in Educational Measurement and Evaluation. The instrument contained 35 items using a four-option response format. The validity of the test was established by three experts in Measurement and Evaluation. Cronbach’s alpha was used to determine the internal consistency of the items, yielding a reliability coefficient of r = .779. An independent-samples t-test was used to analyze the data, with the hypotheses tested at the .05 level of significance. The findings revealed that students’ performance was generally not as significant as expected when compared with findings from previous studies. The study also found that although AI-generated adaptive assessment provides personalized assessment, immediate feedback, and improved test precision, the conventional test (Teacher-Made Test) was perceived as more learner-friendly, more closely related to learners’ learning conditions and time, and more considerate of students’ ability levels during test development. The authors concluded that while AI adaptive tests remain an effective assessment innovation capable of improving students’ academic performance, teachers need to be trained in globally accepted structured formats for formulating test items, particularly in the era of digital learning. The study recommended that educational stakeholders should complement the use of AI adaptive tests in classrooms to enhance learning outcomes.