Research On CAT Placement Test For International Students In China Based On IRT

With the in-depth advancement of the "Belt and Road" initiative,the world's Chinese teaching has flourished.The number of foreign students studying in China has grown steadily.As a prospective test before teaching,the placement test plays an important role in the smooth development of subsequent teaching.With the rapid advancement of Internet technology,computer adaptive testing has become possible.The Research on CAT Placement Test for International Students in China Based on IRT has theoretical and practical significance for improving the reliability and validity of class placement,reducing the workload of class placement.By enhancing the teaching relevance for foreign students with different Chinese abilities and levels,this research can improve the teacher 's teaching effects and students' learning efficiency.The research work of this paper is mainly divided into three stages.The first stage summarized and analyzed the domestic and foreign research status of language testing,second language placement test and computer adaptive placement test.Through the statistics and analysis of the test paper structure,question structure and question frequency of standardized language tests at home and abroad,determined the test paper structure that combines components and skills which includes four parts: vocabulary,grammar,listening comprehension and reading comprehension.The second stage is the development of test questions.First,through comparative analysis,a comparison table of the series of textbooks of GO for CHINESE Elementary Level and the new HSK language proficiency levels was established.Then,based on the common words and grammar points in the textbooks and the new HSK,determined the range of test points for the test questions.On these bases,according to the characteristics of the CAT test,specific question types were determined and specific test questions were developed.The third stage is the construction of a question bank.First,the Chinese international education cloud platform was used to enter the question bank,form papers,publish the test questions,and to collect and score the test papers.Then used EXCEL,SPSS,BILOGMG3.0 for data statistics,hypothesis testing of project response theory,parameter estimation of test questions and question quality analysis,etc.,inferior quality test questions were eliminated,and a question bank was then established.The research work in this paper shows that online testing based on Item Response Theory and relying on computer cloud computing is currently the most suitable test form for the placement test for international students coming to China.First of all,as far as the test format is concerned,students only need to have a smartphone or computer with Internet access to conduct a graded test anytime,anywhere;secondly,as far as the structure of the test paper is concerned,the discrete written test format that combines ingredients and skills can be accurate Evaluate language proficiency level;third,in terms of question type setting,the four-choice one type has the characteristics of convenience and quickness when answering,and can realize automatic grading;finally,in terms of work placement,The IRT-based CAT placement test form can realize the saving of human and material resources,and the teaching institution has more time to reasonably arrange classes and teachers and formulate teaching plans.The innovation of this article lies in the following aspects: Firstly,a comparison table of test paper structure,question structure and question frequency of language standardization tests at home and abroad was constructed;secondly,a series of textbooks and new HSK language proficiency level comparison table;thirdly,480 test questions suitable for computer adaptive testing were developed;fourthly,online pre-testing work was carried out based on the cloud platform;and lastly,the test questions were screened using the method of project response theory,Constructed a question bank that can be applied to the computer adaptive test.The shortcomings of this article include two aspects.Firstly,there is no network-based automatic question classification system;secondly,the setting of some question types needs further consideration.