| In 2004, along with the new basic education curriculum reform, original graduation examination of high school is replaced by academic proficiency test. Although graduation examination is the predecessor of academic proficiency test, there is an essential difference between them. The main difference is that academic proficiency test is one of the references to which colleges and universities recruit students. Academic proficiency test begins as a basis for college recruitment. However, due to the baton’s role of college entrance examination this property of academic proficiency test has not been fully played in many areas. Until 2014, the State Council and Ministry of Education issued related policies and the innovation programme of college entrance examination published in Shanghai and Zhejiang, academic proficiency test truly came into the college entrance examination. Follow is the community’s focus on academic proficiency test. So the quality and professional of test is widely regarded.There are many indicators to measure the quality of items, such as reliability, difficulty, discrimination, etc. Among them the difficulty is most important. The difficulty parameter is contained in not only one-parameter model but also two-parameter model and three-parameter model of item response theory. It demonstrates the importance of difficulty which measures the quality of items. Therefore, how to control the difficulty of item to achieve test goal becomes a hotspot. But the academic proficiency test is not tested before it is conducted, the difficulty of item is mainly based on judgment from proposition personnel experience, there is no practical quantitative method.This study presented 14 factors that affect the difficulty of chemistry academic proficiency test based on research, surveys and interviews. They are amount of knowledge, length of title, module, cognitive level, inference, distracter, acquaintance degree, problem type, operation, skill, and diagram, information of material, macro micro-breaks, and openness. This study establishes an assignment criterion for these 14 factors and analyzes reliability and validity based on past research. Test results prove the reliability and validity of assignment criteria well, and can be used for statistical analysis.In order to verify whether the 14 factors have significant effects, this study takes a statistical analysis of every factor and ultimately removed no significant factors, retaining factors that have significant influence. The analysis method depends on variable types. Pearson correlation test and non-parametric rank test are used for continuous variable. Independent t-test and box-type figure are used for dichotomous variables. One-way ANOVA analysis and box-type figure are used for rank variables. Through the test, after discounting invalid factor, the retained factors are amount of knowledge, cognitive level, acquaintance degree, operation, diagram, macro micro-breaks, and openness.Finally, based on assignment criterion and these 7 factors, the prediction formula is P=P0-0.05(X1+X2-2)-0.05(X3-1)-0.05X4-0.05X5-0.05X6-0.05X7. It turns out that the estimated difficulty agrees well with the measured difficulty, which has some reference to the research of chemistry item difficulty. |
