Validation of the Unified National Testing in Kazakhstan: Evidence-Based Framework for Fair Testing

Abstract

Worldwide, high-stakes entrance examinations are subject to rigorous requirements: they must be as objective, reliable, and fair as possible. This entails that the processes of their development, administration, and scoring are carefully controlled and standardized to eliminate any possibility of bias. In Kazakhstan, university admission is based on the results of the state entrance examination—the Unified National Testing (UNT). The aim of this study is to investigate the validity of the UNT results in Kazakhstan, focusing on the mathematics component. Particular attention is given to two sources of validity evidence:  based on internal test structure and based on relationship with other variables, specifically criteria of students’ subsequent academic success in higher education. The theoretical framework for this research is grounded in S. Messick’s validity theory. A series of validation studies were conducted in accordance with the methodological standards set forth by American Educational Research Association, American Psychological Association and National Council on Measurement in Education, employing classical test theory, modern test theory approaches, and statistical analyses including hierarchical linear modeling. The findings indicate that the UNT mathematics exam demonstrates strong psychometric properties. Furthermore, both the overall UNT score across five subjects and the mathematics effectively predict students’ future academic performance at university. The quality of the statistical models used and the proportion of variance explained in students’ first-semester grades are largely consistent with results reported by researchers from other countries, including Russia and the United States.