Person Fit Statistics to Identify Irrational Response Patterns for Multiple-Choice Tests in Learning Evaluation

This study aims to analyze and describe the response patterns of school exam participants based on the person fit method. This research is a quantitative study with a focus on research on social science elementary school examinations as many as 15 multiple choice items and 137 participant answer sheets. Data collection techniques were carried out with documentation by collecting the participants' exam results and then being scored based on the answer key. The data analysis technique used is statistical person fit analysis with the Rasch Model. Data were analyzed using the R Program with the package latent trait model (ltm). The results showed that as many as 93 or about 67.9% of examinees detected fit or were categorized as having rational response patterns and as many as 44 examinees or around 32.1% of examinees were detected as not fit or categorized as having irrational response patterns. Based on the results of the study, it was concluded that most elementary school exam participants had a response pattern that was fit (had a rational response pattern).