Detection of potential errors in measurement results of madrasa admission instruments in Indonesia

Madrasa (Islamic boarding school) in Indonesia have a strategic role in character building. At present madrasa education is still considered second class education. Besides, to improve the quality of madrasas can be started by improving the quality of the student national admission to all madrasas in Indonesia. This study aimed to trace the potential errors in the measurement results of Students National Admission of Madrasah Aliyah Negeri (SNPDB MAN-IC) 2020. Tracing was carried out on two aspects: i) Equality between test sets used based on evidence of test responses; and ii) Further tests on equality between question sets based on evidence of relationship between variables, taking into account the origin of the participating schools (MTs/JHS) and the origin of the participating regions (West, Central and East of Indonesia). This study involved 13,115 participants in 23 MAN-ICs throughout Indonesia in 2020. The materials tested comprised learning potential and academic ability (Mathematics, Natural Sciences, Social Studies, English, Arabic, and Islamic Religious Education). The study used achievement test with mathematics as a sample of test subjects. Based on the test response evidence, it was found that seven of the 15 questions were thought to have an indication of inequality between item sets. The results of tracing the evidence between variables indicated that it was the participants' origin of institutions that influenced the inequality between item sets. On the other hand, regional origin did not affect the inequality between item sets because the majority of participants came from the western region of Indonesia.