The Effect of Proportion and Position of Anchor Items Toward Test Equating

This study examines how the proportion and position of anchor items influence test equating results, a key component in ensuring fairness and accuracy in standardized assessments. In this experimental design, the independent variables are the proportion and position of anchor items, and the dependent variable is the absolute difference in ability parameters before and after the equating process. To assess this, we analyzed test response data from 1,000 respondents, each completing 40 items across 24 data sets, generated through Monte Carlo simulations for reliability. The analysis was conducted using bi-factor cell mean analysis, a method that explores the interaction between various factors influencing equating outcomes. Results show that (1) a higher proportion of anchor items enhances the accuracy of test equating, (2) the position of anchor items significantly affects the equating outcomes, with positions at the beginning showing the highest accuracy, (3) the interaction of both proportion and position plays a crucial role in improving equating results, and (4) the position of anchor items has a more significant impact on equating accuracy than the proportion of anchor items.