Analysis of the difficulty level of physics cognitive test items using the Rasch approach: Implementation of analysis with R Software

This study aims to analyze the difficulty level of Physics cognitive test items using the Rasch approach and the implementation of analysis through the application of R. The instrument consisted of 10 dichotomous items with three cognitive levels: C2 (comprehension), C3 (application), and C4 (analysis). A total of 100 students from SMAN 9 Gowa participated as respondents. The analysis was carried out using the Rasch 1PL model to obtain difficulty parameters (b), fit items and test information. The results showed that the difficulty of the item varied widely in the range of 0.12 to 2.22 logits, so that it was able to distinguish students' abilities, especially at the intermediate level, as evidenced by the peak of the Test Information Function at θ ≈ 1. However, some C2 items appear more difficult than the theory, which indicates the need for revisions to the redaction and effectiveness of the distractors. Items C3 and C4 were relatively consistent, although some items were classified as very difficult. The ICC analysis also showed differences in characteristics between items according to the principles of the Rasch model. Overall, the use of Rasch through R has proven to be effective in providing objective information about the quality of the items so that it can be used as a basis for improving the Physics cognitive assessment instrument to be more valid and in accordance with students' abilities.