This paper investigated that the effect of the several fit statistic on the two kinds of misspecification of CDM. The saturated model, DINA, and A-CDM were compared. An empirical data was applied to illustrate how the statistic in the simulation study can be analyzed in the real situation. However, my question is that it seems no evidence to conclude which statistic is better to be adapted when the real data are analyzed. The other question is that the author just explore the phenomenon of using the statistics to check model fit but didn’t explain why and what consequents the phenomenon will lead to.