Mixture Rasch models (MRMs) offer the potential for providing increased information about examinees’ abilities as well as their response strategies not captured by traditional single class IRT models. However, under some conditions overextraction of latent classes may occur, potentially leading to misinterpretation of the results. An empirical data was demonstrated first to show the extra latent classes in MRMs are arises from the underfit items exist. Three simulation studies were conducted to demonstrate how data generated to fit a one-class 2-parameter logistic (2PL) model required more than one class when fit with MRMs.
In simulation study 1, the purpose was to examine the sensitivity of the MRM to violations of Rasch model assumptions. The result shows that a one-class Rasch model estimates a 2PL ICC as well as a two-class MRM and, therefore, there is no need for two classes. In simulation study 2, it was to show that how many items were needed, how high the item discrimination needed to be, and how large a sample size was needed to cause an extra latent class for form. The result shows that the lower the discrimination index, the larger the number of items required to trigger the second class. In simulation study 3, the result shows that two classes in a MRM did not necessarily collapse into a one-class 2PL when a mixture 2PL model was used to data generated as a two-class MRM.
Question:
Q1: how to apply the view of this study when the mixture Rasch model was used?
Q2: the mixture Rasch model should be used to analysis or not?