Abstract The partial credit model was used to analyze the first simulation test of Shanghai chemistry college entrance examination from a double-layer, one-dimensional perspective. The study draws the following conclusions by analyzing 3 864 samples. The reliability and discrimination of the test is good (the test reliability is 1.00, the discrimination is 26.20), and the average difficulty of the test is lower than the average individual ability of 1.31 logits, which is relatively low. The overall one-dimensionality of the test is good, but some test questions are poorly fitted. The DIF analysis shows that there is no item function differentially for examinees from different gender. For different physics learning groups and different political learning groups, there are 7 items and 1 item respectively to discover the presence of DIF, but this has little effect on the overall evaluation. At the end of the paper, combined with expert consultation, the low-quality test questions are analyzed. At the same time, from the three aspects of test quality, scoring rules and teaching, it gives relevant suggestions on how to form a good evaluation analysis.
XU Jia-Min, SHEN Dian. Analysis of the First Simulation Test of Chemistry College Entrance Examination of Certain District of Shanghai Based on Rasch Model[J]. Chinese Journal of Chemical Education, 2021, 42(11): 10-17.
