Development of a test instrument using multiple representations to assess students understanding: A Rasch Model Analysis

Abstract

Today's test instruments used in schools are still ineffective in assessing students' multi-representational understanding of the "rate of reaction" topic. Hence, teachers need to accurately and comprehensively know the depth of students' understanding. In this study, a test instrument was produced to assess the level of macroscopic, submicroscopic, and symbolic understanding of high school students in the rates of reaction topic. The method used is Research and Development (R&D) based on the Rasch modeling approach, which is modified in ten stages, i.e., (1) defining construct, (2) identifying question indicators, (3) compiling items, scoring rubrics, scoring guidelines, and the guidelines to analyze student's level of understanding, (4) conducting the pilot test, (5) analyzing data with the Rasch model, (6) reviewing the suitability of items, (7) reviewing the Wright map, (8) repeating steps 4-7 until all items fit, (9) claiming the instrument's quality, and (10) documenting the instrument. All items proved valid because they met the fit criteria of MNSQ, ZSTD, and PT Measure scores. Instrument reliability reaches the score of 0.89 (good). The instrument has a four-difficulty index, dominated by moderate difficulty (70.6%). This instrument is relatively easy. The instrument also has four discrimination indexes: very difficult, difficult, medium, and easy. It proves that this assessment instrument is of good quality and suitable for school use.