探索高校网络题库与考试系统的数据分析在教学实践中的应用  

Exploring the Application of Data Analysis in University Online Question Banks and Examination Systems in Teaching Practice

作　　者：王彦[1] 李骢[2] 刘冰心[3] 刘明[4] 赵赫男[2] WANG Yan;LI Chong;LIU Bing-xin;LIU Ming;ZHAO He-nan(Scientific Research Center,The Second Hospital of Dalian Medical University,Dalian 116021,China;Department of Pathophysiology,College of Basic Medical Sciences,Dalian Medical University,Dalian 116044,China;Department of Joint and Sports Medicine,The Second Hospital of Dalian Medical University,Dalian 116021,China;Department of Anesthesiology,The Second Hospital of Dalian Medical University,Dalian 116021,China)

机构地区：[1]大连医科大学附属第二医院科研中心,大连116021 [2]大连医科大学病理生理学教研室,大连116044 [3]大连医科大学附属第二医院关节与运动医学科,大连116021 [4]大连医科大学附属第二医院麻醉科,大连116021

出　　处：《高校医学教学研究(电子版)》2024年第6期57-63,共7页Medicine Teaching in University (Electronic Edition)

基　　金：2023年度大连医科大学校级教学改革研究立项(DYLX23004)。

摘　　要：目的通过分析网络题库与考试系统的相关数据,探索高校利用信息化数据资源辅助、指导教学实践的新思路和方法。方法对数据库试题类型的分布特征进行描述,采用矩阵分析探索课时设置与不同类型试题的关系,对数据库中相关参数进行逻辑回归分析,对影响学生A1型试题实测难度的影响因素进行了单因素及多因素回归分析。结果A1型试题数量占比达42.14%,显著高于其他题型(F=30.30,P均<0.001),且与授课学时数高度相关。试题“质量”判定与题源、认知分类、知识点分布、实际难度及实际学习效果间的相关关系均有统计学意义。各章节中,学生的实际学习效果均未出现显著<60%的情况(单因素及多因素Cox回归分析中P均>0.05)。题源、认知分类及主题词目录级别中,各亚组出现学生实际学习效果<60%的风险均较低。在知识点分类分析中,学生对“超纲题”的实际学习效果反而较好,多因素回归分析显示其P值<0.05。试题预测难度及试题质量判定分析均显示,无论单因素还是多因素分析,HR均>1,且P均<0.001。“需要修改”类试题其单因素回归分析显示HR>1,且P=0.001。结论本研究通过对题库相关数据进行分析,探索教学效果的影响因素,为高校调整教学实践策略提供参考。Objective Through the data analysis of the online question bank and examination system,we explore the new ideas and methods for colleges and universities to use the massive data resources of the database to assist and guide the teaching practice.Methods The distribution characteristics of test questions in the database were analyzed.Through matrix analysis,the relationship and rationality between the lesson time setting and different types of test questions were explored.Logistic regression analysis was performed on the relevant parameters in the database.Taking A1-type test questions as an example,univariate and multivariate regression analysis was carried out on the influencing factors affecting the difficulty of students’actual tests.Results The number of A1 questions accounted for 42.14%,which was significantly higher than that of other question types(F=30.30,P<0.001),and correlating strongly with the number of teaching hours.There was a statistically significant correlation between the“quality”of test questions and their source,cognitive classification,knowledge point distribution,actual difficulty,and student learning outcomes.The actual learning effect for students in each chapter was at least 60%(P>0.05 in both univariate and multivariate Cox regression analyses).The risk of fewer than 60%of students achieving the desired learning effect was lower across all subgroups based on question source,cognitive classification,and subject heading.In knowledge point classification,“over-the-syllabus questions”yielded better learning outcomes,with“over-curriculum questions”showing a significant result(P<0.05)in multivariate regression analysis.Furthermore,both univariate and multivariate analyses indicated that the prediction of test question difficulty and quality judgments had an HR>1 and P<0.001.Univariate regression analysis of“need to modify”test questions also showed HR>1 and P=0.001.Conclusion Using the massive data in the university database can explore deeply the influencing factors of teac

关 键 词：网络数据系统 数据分析 教学实践 

分 类 号：G63[文化科学—教育学]