Deepening Understanding of Common Persons Design in Score Equating: A Simulation Study

• In Event: Virtual Poster Session
  In Virtual iPresentation Only: Division D - Section 1: Measurement, Psychometrics, and Assessment Virtual Poster Session

Abstract

The Common Persons (CP) equating design benefits high-security testing by avoiding anchor exposure and handling non-equivalent groups, but lacks clear implementation guidance. Addressing this gap, this comprehensive Monte Carlo simulation examined 8 influencing factors and compared four equating methods (identity, IRT true-score, linear, equipercentile) using NRMSE and %Bias. Key findings reveal: (1) CP sample characteristics have minimal impact on accuracy; (2) Test factors, especially difficulty shifts and test length, are decisive; (3) Equipercentile and linear methods are most robust. A minimum of 30 CPs across the score range is sufficient for reliable equating. These results provide the first evidence-based framework for CP implementation, resolving security-vs-accuracy tradeoffs in high-stakes equating (e.g., credentialing exams) and enabling novel solutions like synthetic respondents.

Authors

• Jiayi Liu, Peking University

• Zhehan Jiang, Peking University

• Tianpeng Zheng, Peking University

• Shicong Feng, Peking University