Nonignorable Consequences of (Partially) Ignoring Missing Item Responses: Students Omit (Constructed Response) Items Due to a Lack of Knowledge

In recent literature, alternative models for handling missing item responses in large-scale assessments have been proposed. Based on simulations and arguments based psychometric test theory, it is argued this literature that should never be scored as incorrect scaling but rather treated ignorable or handled a model. The present article shows these limited validity illustrates the consequences country comparison using PIRLS 2011 study. It students omit (constructed response) items because they do not know correct answer. A different treatment of than scoring them leads to significant changes rankings, which induces nonignorable regarding results. Additionally, two response are proposed assumptions responses. first pseudo-likelihood approach, particular student replaced by score ranges between zero model-implied probability computed non-missing items. second predicted latent propensity variable itself. were applied study, demonstrating comparisons change under modeling