Document Type : Research Paper
Authors
1 National Organization for Assessment and Evaluation of Education System
2 Department of Psychometrics, Faculty of Psychology and Education, Allameh Tabataba'i University, Tehran, Iran
3 Ph.D of Assessment and Measurement, National Organization for Assessment and Evaluation of Education System, Tehran, Iran
Abstract
Conducting two nationwide tests in one year necessitates creating alternate forms of the tests and equating their scores. Equating the alternate forms needs specific requirements. In high-stakes tests, pilot testing of items is not possible due to security concerns. This study aimed to investigate the requirements for equating alternate forms of the first and second 2023 English foreign language nationwide tests. The current research is applied and psychometric, based on secondary data analysis. A total of 13,091 candidates from the first test and 19,915 candidates from the second test of the 2023 English language were randomly selected, and the data from the tests conducted in January 2023 and July 2023 were analyzed. The mean difficulty and discrimination coefficients of the two tests were reported to be close to each other. Equal construct requirement of the two tests with single-factor confirmatory factor analysis, comparing the performance of male and female candidates, and the relationship between test scores and high school GPA were confirmed. Furthermore, the equal consistency requirement was examined, and its value was 0.95 in both test forms. Equating alternate forms necessitates adherence to requirements, where judgmental estimation of item difficulty and using the test specification table can aid test developers in creating alternate forms. Guidelines have been provided to meet the requirements of equating and writing items with similar difficulty.
Keywords
)