The present research aims to develop an instrument called the two-tier multiple choice test for assessing students’ misconceptions in science, exploring students’ misconceptions in science, evaluating item difficulty patterns across science disciplines such as Physics, Chemistry, and Biology, and lastly performing comprehensive evaluation for students’ inductive reasoning skills. Before conducting a research activity and data collection, A systematic literature review had been conducted to reveal topics in science that distribute misconceptions to students. The most widely used diagnostic test was found as multiple-tier tests (33.06%). In this study, the two-tier multiple choice test was developed and tested in the pilot and main study to assess students’ misconceptions in Science. To assess inductive reasoning, the adapted inductive reasoning test in the Indonesian version. The data collection in this study was conducted in two-stage using a cross-sectional studies design. The first stage is data collection for the pilot study with 153 students at senior high schools. The second stage is data collection for main studies with 856 senior high school students and a pre-service science. Rasch measurement with the unidimensional model is used as the main statistical analysis assisting parametric and nonparametric statistical analysis. The finding confirmed that the developed two-tier multiple-choice test in the pilot study is valid and reliable based on Rasch parameters. The difficulty level of items on various science concepts is not universally based on science topics, but they are connected or similar across science disciplines, especially in physics, biology, and chemistry. The exploration of students’ misconceptions revealed students have misconceptions in science subjects physics (33.4%), biology (35.22%), and chemistry (47.97). In comparing school grades, based on school grade regression analysis , the 25.2% variance of student misconceptions in science was explained. Stepwise regression showed that only the school grade predictor could significantly explain 25.2% of the variance in mean scores, F (152) = 10.208, p