Digital Assessment Tools: As a Media for Assesing High School Science Learning

Abstract

This research is to produce a suitable digital assessment tool for high school students' in chemical literacy with validity, reliability, levels of difficulty and discrimination power. The research metodology is Research and Development (R&D) using the Oriondo and Dallo-Antonio model. The research stages include planning the test instrument, testing, determining empirical validity, determining reliability, and interpreting test scores. In the planning (instrument creation), where the validation score estimates by experts in content, language, construction, and media are above 0.8, indicating that the test instrument is considered suitable for further testing to estimate validity, reliability, difficulty levels, and discrimination power. Out of the 25 developed questions, 24 fit the Rasch model. The reliability of Instrument Part I is 0.67, and Part II is 0.65, falling into the "sufficient" category. The questions exhibit varied difficulty levels, ranging from very easy to very difficult, with a standard deviation (SD) of 0.21 for Part I and 0.14 for Part II. The separation achieved by Instrument Part I is into 4 ability groups, while Instrument Part II separates into 2 ability groups. The distractors for multiple-choice questions fall into the "good" and "very good" categories, fulfilling their intended function effectively.