Validitas dan reliabilitas instrumen asesmen kinerja literasi sains pelajaran fisika berbasis STEM

Penelitian ini merupakan bagian dari penelitian pengembangan asesmen kinerja literasi sains berbasis STEM pada pembelajaran fisika. Tujuan dari penelitian ini adalah untuk mengungkapkan validitas isi, validitas empiris, dan reliabilitas instrumen asesmen kinerja literasi sains berbasis STEM yang sebelumnya telah disusun. Instrumen yang dikembangkan berupa lembar pengamatan dan tes pilihan ganda. Analisis validitas isi dari lembar pengamatan menggunakan Koefisien V oleh Aiken sedangkan validitas isi instrumen tes dianalisis dengan menggunakan CVI (Content Validity Index) oleh Lawshe. Validitas empiris reliabilitas instrumen tes diestimasi dengan IRT (Item Response Theory). Reliabilitas lembar pengamatan ditentukan dengan ICC (Item Correlation Coefficient). Hasil dari penelitian ini menunjukan bahwa (1) Lembar pengamatan berupa rubrik penskoran dan penilaian diri  terbuktivalid dengan koefisien  V Aiken 0,75 dan reliabel dengan koefisien Reliabilitas Alfa > 0,8 dan ICC yang Excellent. (2) Instrumen tes terbukti realiabel untuk digunakan pada peserta didik dengan kategori sedang sampai dengan tinggi (-0,7 sampai dengan 6,7 ) dengan CVI=1 dan INFIT MNSQ sesuai model Rasch. Berdasarkan hasil penelitian tersebut maka asesmen kinerja Literasi Sains berbasis STEMlayak digunakan.Kata kunci:validitas isi, validitas empiris, asesmen kinerja, literasi sains, STEM VALIDITY AND RELIABILITY INSTRUMENT OF SCIENTIFIC LITERACY PERFORMANCE ASSESSMENT IN PHYSICS TEACHING BASED ON STEMAbstractThis research is part of the development of scientific literacy performance assessment based on STEM in teaching physics. The aim of this research is to reveal the validity (content and also empiric) and reliability of scientific literacy performance assessment instrument based on STEM. The kind of instruments were developed are observational sheet and multiple choice test. The content validity of observational sheet was revealed by used the Aiken’s V Coefficient. The content validity of multiple choice tests was revealed by used Content Validity Index (CVI) which proposed by Lawshe. The empirical validity and reliability of multiple choice tests was revealed by used Item Response Theory Analysis. The reliability of observational sheet was revealed by used ICC (Item Correlation Coefficient) Analysis. The results of this study are the validity from the contents and empirical trials from the developed instruments. The observation sheet from scoring rubric and self-assessment has been valid with Aiken’s V value that exceeds the standard of 0,75. The reliability of the scoring rubric has Alfa Reliability> 0.8 and Excellent of ICC. Validity values from The written test is shown with CVI of 1 and the MNSQ INFIT value which match to the Rasch model. Based on the TIC and SEM graphs, the written test is stated to be reliable for use in students with moderate to high categories (-0.7 to 6.7). STEM-based Science Literacy performance assessment with caloric material is appropriate to use.Keywords: content validity, empirical validity, performance assessment, scientific literacy, STEM