Explore feasibility of using multiple imputation for dimensionality analyses in multistage testing

• In Event: Multistage Testing

Abstract

This study examines the feasibility of using multiple imputation by chained equations (MICE) for dimensionality analyses in multistage test. We compare four MICE based methods and evaluate using DIMTEST and DETECT indexes. Results show that MICE with random forest has highest accuracy to recover the true factor structure.

Authors

• Danqi Zhu, FORDHAM UNIVERSITY

• Louis Roussos, Cognia

• Han Yi Kim, Cognia

• Sandra Sweeney, Cognia

• Qi Qin, Cognia