Robust Maximum Marginal Likelihood (RMML) Estimation for Item Response Theory Models
Self-report data are common in psychological and survey research. Unfortunately, manyof these samples are plagued with careless responses due to unmotivated participants. Thepurpose of this study is to propose and evaluate a robust estimation method in order to detectcareless, or unmotivated, responders while leveraging Item Response Theory (IRT) person fitstatistics. First, we outline a general framework for robust estimation specific for IRT models.Subsequently, we conduct a simulation study covering multiple conditions to evaluate theperformance of the proposed method. Ultimately, we show how robust maximum marginallikelihood (RMML) estimation significantly improves detection rates for careless responders andreduce bias in item parameters across conditions. Furthermore, we apply our method to a realdataset to illustrate the utility of the proposed method. Our findings suggest that robustestimation coupled with person fit statistics offers a powerful procedure to identify carelessrespondents for further review, and to provide more accurate item parameter estimates inpresence of careless responses.