Validity And Reliability Rubric Of Performance Assessment Geometry Study In Junior High School Using The Many Facet Rasch Model Approach

Abstract

The purpose of this study was to analyze the validity and reliability rubric of performance appraisal on geometry subject matter using the many facet rasch model approach through the Facets software. Data were collected from 100 small-scale students and 250 large-scale students in junior high schools using 3 raters. The performance assessment instrument in the form of a rubric is used to assess the student's process of working on the questions, each question has a different rubric. The Rasch Many Faceted Measurement Model (MFRM) is used to analyze data by looking at three aspects, namely the facet person, rater agreement, and difficulty domain using the Facet program. For the facet person, the rater separation ratio was 4.96, while the reliability of the separation index was 2.15 which indicates that the assessors are separated reliably. The stratum index is 3.21 which indicates that there are three strata of rater severity that differ statistically in the sample of these 4 raters. The rater agreement obtained the reliability of the rater separation of 0.87 and the correlation between each assessor and the other ranged between 0.40 and 0.63, indicating adequate agreement among the raters in assessing test participants with their level of competence. The Difficulty Domain on the variable maps shows that the hard to soft range is from about +1 to −1 logit. Ke