Open AccessANALISIS ESTIMASI KEMAMPUAN SISWA DENGAN PENDEKATAN ITEM RESPONSE THEORY PENSKORAN DIKOTOMUS DAN POLITOMUS
Author(s) -
Duden Saepuzaman,
Edi Istiyono,
Haryanto Haryanto,
Heri Retnawati,
Yustiandi
Publication year - 2021
Publication title -
karst
Language(s) - English
Resource type - Journals
eISSN - 2655-1276
pISSN - 2622-9641
DOI - 10.46918/karst.v4i1.948
Subject(s) - polytomous rasch model , item response theory , psychology , statistics , test (biology) , reliability (semiconductor) , mathematics , psychometrics , power (physics) , paleontology , physics , quantum mechanics , biology
This study compares students' abilities using the Item Response Theory (IRT) approach to dichotomous and polytomous scoring. This research is quantitative descriptive. The research subjects were 1175 high school students in class XI spread across West Java and Banten provinces, consisting of 450 male students and 725 female students. Response data with dichotomous scoring were analyzed using the item response theory approach with the BILOG-MG program, while the polytomous scoring was analyzed using the GPCM approach using the R program. The results of the model fit test showed that the items most fit with the 2PL model. The instrument used is a work and energy material test instrument tested for validity d, reliability, distinguishing power, and difficulty level. The results showed that the average ability with a polytomous score was greater than that of a dichotomous, even though compared to having almost the same ability but with a relatively different distribution. The distribution of students' abilities with polytomous scoring is closer to the normal curve than the dichotomous scoring. The relationship between students' ability scores with these two scoring approaches is shown by a correlation coefficient score of 0.990 and a determination index of 0.9808 with a prediction line y= 0.9735 x +0.0036.