面向MOOC课堂反馈的学习行为分析研究
|
摘要
课堂反馈作为最直接的教师-学生互动内容,对于MOOC课程中学生的课堂学习行为研究具有重要意义.大规模的课堂反馈记录有利于观察学生的参与度和活跃度变化规律;同时,课堂反馈内容又可以反映出学生的知识掌握程度,便于教师精准地捕捉学生反馈水平和优化教学结构.基于此,本文选取中国大学MOOC平台"课堂交流区"中的5万条教师-学生反馈记录作为研究对象,结合数学统计和自然语言处理等学习分析技术,从反馈记录的统计分析和反馈内容的话语分析两个方面入手,追踪学生在课程学习过程中的参与率和活跃度变化,并进一步探讨不同学科类型课程的学习行为差异及其课堂话语分布.结果表明:学生的反馈内容主要围绕主题内容展开,正面情感比例较高,但其参与率和活跃度较低,学习积极性有待提高.就不同学科类型课程的学习行为而言,艺术类课程的参与人数比例和反馈数量呈现出最优状态,反馈内容的广度均高于其他类型课程;文科类课程呈现出参与率高而活跃度低的不协调现象,反馈内容注重主题的深层次思考;而理科类课程力求简洁准确的回复主题,学生间的反馈重复度高,适当提高其课堂参与率和活跃度也将有助于学生间的学习交流,促进知识建构行为的优化发展.
As a popular interaction between students and instructors,MOOC classroom feedback analysis is crucial to the success of learning and teaching.It may help teachers to grasp students'learning situation precisely and promote students to examine themselves actively.Traditionally,instructors have to analyze these responses in a manually and costly manner.In this work,text-based student feedback after each lecture are automatically collected.Using statistical analysis and natural language processing technol-ogy,we then investigate student learning behavior from various reply records and textbased feedback.Experimental results on a student feedback corpus from China College MOOC show how the student's engagement and activeness vary over time.We also demonstrate how students in various disciplines perform differently between their learning behavior and classroom discourse distribution.The results show that students'feedback almost focuses on these course contents themselves,and positive emotions account for a large proportion.But student's engagement and activeness are at a lower level,which need to be worth improving further.There was significant variability of learning behavior between and within different courses.For Arts courses,the rate of students'participation and the activeness of reply present the superior states,the extent of content is higher than other courses.Liberal Arts courses expose the imbalance states between higher engagement and lower activeness,and feedback content focus on the deep thought of topic.While Science courses are inclined to reply questions compactly and repeatable.The improvement of its engagement and activeness contributes to learning communication and knowledge construction process.
As a popular interaction between students and instructors,MOOC classroom feedback analysis is crucial to the success of learning and teaching.It may help teachers to grasp students'learning situation precisely and promote students to examine themselves actively.Traditionally,instructors have to analyze these responses in a manually and costly manner.In this work,text-based student feedback after each lecture are automatically collected.Using statistical analysis and natural language processing technol-ogy,we then investigate student learning behavior from various reply records and textbased feedback.Experimental results on a student feedback corpus from China College MOOC show how the student's engagement and activeness vary over time.We also demonstrate how students in various disciplines perform differently between their learning behavior and classroom discourse distribution.The results show that students'feedback almost focuses on these course contents themselves,and positive emotions account for a large proportion.But student's engagement and activeness are at a lower level,which need to be worth improving further.There was significant variability of learning behavior between and within different courses.For Arts courses,the rate of students'participation and the activeness of reply present the superior states,the extent of content is higher than other courses.Liberal Arts courses expose the imbalance states between higher engagement and lower activeness,and feedback content focus on the deep thought of topic.While Science courses are inclined to reply questions compactly and repeatable.The improvement of its engagement and activeness contributes to learning communication and knowledge construction process.
引文
[1]冷静,吴小芳,顾小清.面向深度学习的在线课程活动设计研究——基于英国开放大学的案例剖析[J].远程教育杂志,2017,35(2):56-65.LENG J,WU X F,GU X Q.Research on the deep learning online course activity design:Based on case analysis to the open university of UK[J].Journal of Distance Education,2017,35(2):56-65.(Ch).
[2]耿宏丽.课堂反馈:现实偏差及矫正[J].教育测量与评价:理论版,2009(10):38-40.GENG H L.Classroom feedback:Realistic deviations and corrections[J].Educational Measurement and Evaluation,2009(10):38-40.(Ch).
[3]吴永和,陈丹,马晓玲,等.学习分析:教育信息化的新浪潮[J].远程教育杂志,2013(4):11-19.WU Y H,CHEN D,MA X L,et al.Learning analytics:The new coming wave of education information[J].Journal of Distance Education,2013(4):11-19.(Ch).
[4]JOHNSON L,SMITH R,WILLIS H,et al.The 2011horizon report[J].New Media Consortium,2011(3):40-44.
[5]JIN H,LIU D,GAO L,et al.An analysis of the NMC horizon report:2016higher education edition[J].Journal of Distance Education,2016(5):27-33.
[6]刘智,张文静,孙建文,等.云课堂论坛中的学习者互动话语行为分析研究[J].电化教育研究,2016(9):95-102.LIU Z,ZHANG W J,SUN J W,et al.Analysis of learners'interactive discourse behaviors in the cloud classroom forum[J].E-education Research,2016(9):95-102.(Ch).
[7]曹传东,赵华新.MOOC课程讨论区的社会性交互个案研究[J].中国远程教育:综合版,2016(3):39-44.CAO C D,ZHAO H X.Social interaction case study in MOOC course discussion area[J].Distance Education in China,2016(3):39-44.(Ch).
[8]孙洪涛,郑勤华,陈丽.中国MOOCs教学交互状况调查研究[J].开放教育研究,2016(1):72-79.SUN H T,ZHENG Q H,CHEN L.A survey of MOOC instructional interaction in China[J].Open Education Research,2016(1):72-79.(Ch).
[9]ANDERSON A.Engaging with massive online courses[C]//Proceeding:WWW'14Proceedings of the 23rd International Conference on World Wide Web,Seoul,Korea,2014:687-698.
[10]徐彬,张昱,李封,等.MOOC课程论坛活跃度衡量指标及影响因素浅析[J].计算机教育,2016,261(9):130-134.XU B,ZHANG Y,LI F,et al.Analysis on the measurement indexes and influencing factors of MOOC curriculum forum activity[J].Computer Education,2016,261(9):130-134.(Ch).
[11]牟智佳.MOOCs学习参与度影响因素的结构关系与效应研究——自我决定理论的视角[J].电化教育研究,2017(10):126-133.MOU Z J.Research on structural relationship and effects of influential factors of MOOCs learning participation:From perspective of self-determination theory[J].E-education Research,2017(10):126-133.(Ch).
[12]吴江,贺超城,马磐昊.基于迭代超中心度的MOOC论坛用户知识互动超网络研究[J].现代图书情报技术,2017,1(8):1-8.WU J,HE C C,MA P H.Analyzing interaction of MOOC users with iteration super centrality[J].New Technology of Library and Information Service,2017,1(8):1-8.(Ch).
[13]蒋卓轩,张岩,李晓明.基于MOOC数据的学习行为分析与预测[J].计算机研究与发展,2015,52(3):614-628.JIANG Z X,ZHANG Z M,LI X.Learning behavior analysis and prediction based on MOOC data[J].Journal of Computer Research and Development,2015,52(3):614-628.(Ch).
[14]赵慧琼,姜强,赵蔚,等.基于大数据学习分析的在线学习绩效预警因素及干预对策的实证研究[J].电化教育研究,2017(1):62-69.ZHAO H Q,JIANG Q,ZHAO W,et al.Empirical research of predictive factors and intervention countermeasures of online learning performance on big data-based learning analytics[J].E-education Research,2017(1):62-69.(Ch).
[15]李良,乔海英,王淑平.基于Moodle平台的学习者社会性交互特征研究[J].电化教育研究,2012(7):48-53.LI L,QIAO H Y,WANG S P.Study on learners'social interaction features based on moodle platform[J].E-education Research,2012(7):48-53.(Ch).
[16]史慧姗,郑燕林,马芸.MOOC论坛教师参与度对学生参与的影响研究[J].现代远距离教育,2017(6):47-56.SHI H S,ZHENG Y L,MA Y.Research on the influence of teachers'participation on students'participation in MOOC forum[J].Modern Distance Education,2017(6):47-56.(Ch).
[17]LUO W,LIU F,LIU Z,et al.Automatic summarization of student course feedback[C]//Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies.2016:80-85.
[18]WISE A F,CUI Y,VYTASEK J.Bringing order to chaos in MOOC discussion forums with content-related thread identification[C]//International Conference on Learning Analytics&Knowledge,ACM,2016:188-197.
[19]胡艺龄,祝嘉钰,顾小清.以话语分析挖掘社会性学习价值——访国际知名教育心理学专家凡妮莎·登嫩教授[J].开放教育研究,2017,24(2):4-12.HU Y L,ZHU J Y,GU X Q.Exploring the value of social learning by discourse analysis:An interview with Prof.Vanessa P.Dennen[J].Open Education Research,2017,24(2):4-12.(Ch).
[20]吴林静,刘清堂,毛刚,等.大数据视角下的慕课评论语义分析模型及应用研究[J].电化教育研究,2017(11):43-48.WU L J,LIU Q T,MAO G,et al.A semantic analysis method of MOOC online posts from the perspective of big data[J].E-education Research,2017(11):43-48.(Ch).
[21]CROSSLEY S,PAQUETTE L,DASCALU M,et al.Combining click-stream data with NLP tools to better understand MOOC completion[C]//International Conference,ACM,2016:6-14.
[2]耿宏丽.课堂反馈:现实偏差及矫正[J].教育测量与评价:理论版,2009(10):38-40.GENG H L.Classroom feedback:Realistic deviations and corrections[J].Educational Measurement and Evaluation,2009(10):38-40.(Ch).
[3]吴永和,陈丹,马晓玲,等.学习分析:教育信息化的新浪潮[J].远程教育杂志,2013(4):11-19.WU Y H,CHEN D,MA X L,et al.Learning analytics:The new coming wave of education information[J].Journal of Distance Education,2013(4):11-19.(Ch).
[4]JOHNSON L,SMITH R,WILLIS H,et al.The 2011horizon report[J].New Media Consortium,2011(3):40-44.
[5]JIN H,LIU D,GAO L,et al.An analysis of the NMC horizon report:2016higher education edition[J].Journal of Distance Education,2016(5):27-33.
[6]刘智,张文静,孙建文,等.云课堂论坛中的学习者互动话语行为分析研究[J].电化教育研究,2016(9):95-102.LIU Z,ZHANG W J,SUN J W,et al.Analysis of learners'interactive discourse behaviors in the cloud classroom forum[J].E-education Research,2016(9):95-102.(Ch).
[7]曹传东,赵华新.MOOC课程讨论区的社会性交互个案研究[J].中国远程教育:综合版,2016(3):39-44.CAO C D,ZHAO H X.Social interaction case study in MOOC course discussion area[J].Distance Education in China,2016(3):39-44.(Ch).
[8]孙洪涛,郑勤华,陈丽.中国MOOCs教学交互状况调查研究[J].开放教育研究,2016(1):72-79.SUN H T,ZHENG Q H,CHEN L.A survey of MOOC instructional interaction in China[J].Open Education Research,2016(1):72-79.(Ch).
[9]ANDERSON A.Engaging with massive online courses[C]//Proceeding:WWW'14Proceedings of the 23rd International Conference on World Wide Web,Seoul,Korea,2014:687-698.
[10]徐彬,张昱,李封,等.MOOC课程论坛活跃度衡量指标及影响因素浅析[J].计算机教育,2016,261(9):130-134.XU B,ZHANG Y,LI F,et al.Analysis on the measurement indexes and influencing factors of MOOC curriculum forum activity[J].Computer Education,2016,261(9):130-134.(Ch).
[11]牟智佳.MOOCs学习参与度影响因素的结构关系与效应研究——自我决定理论的视角[J].电化教育研究,2017(10):126-133.MOU Z J.Research on structural relationship and effects of influential factors of MOOCs learning participation:From perspective of self-determination theory[J].E-education Research,2017(10):126-133.(Ch).
[12]吴江,贺超城,马磐昊.基于迭代超中心度的MOOC论坛用户知识互动超网络研究[J].现代图书情报技术,2017,1(8):1-8.WU J,HE C C,MA P H.Analyzing interaction of MOOC users with iteration super centrality[J].New Technology of Library and Information Service,2017,1(8):1-8.(Ch).
[13]蒋卓轩,张岩,李晓明.基于MOOC数据的学习行为分析与预测[J].计算机研究与发展,2015,52(3):614-628.JIANG Z X,ZHANG Z M,LI X.Learning behavior analysis and prediction based on MOOC data[J].Journal of Computer Research and Development,2015,52(3):614-628.(Ch).
[14]赵慧琼,姜强,赵蔚,等.基于大数据学习分析的在线学习绩效预警因素及干预对策的实证研究[J].电化教育研究,2017(1):62-69.ZHAO H Q,JIANG Q,ZHAO W,et al.Empirical research of predictive factors and intervention countermeasures of online learning performance on big data-based learning analytics[J].E-education Research,2017(1):62-69.(Ch).
[15]李良,乔海英,王淑平.基于Moodle平台的学习者社会性交互特征研究[J].电化教育研究,2012(7):48-53.LI L,QIAO H Y,WANG S P.Study on learners'social interaction features based on moodle platform[J].E-education Research,2012(7):48-53.(Ch).
[16]史慧姗,郑燕林,马芸.MOOC论坛教师参与度对学生参与的影响研究[J].现代远距离教育,2017(6):47-56.SHI H S,ZHENG Y L,MA Y.Research on the influence of teachers'participation on students'participation in MOOC forum[J].Modern Distance Education,2017(6):47-56.(Ch).
[17]LUO W,LIU F,LIU Z,et al.Automatic summarization of student course feedback[C]//Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies.2016:80-85.
[18]WISE A F,CUI Y,VYTASEK J.Bringing order to chaos in MOOC discussion forums with content-related thread identification[C]//International Conference on Learning Analytics&Knowledge,ACM,2016:188-197.
[19]胡艺龄,祝嘉钰,顾小清.以话语分析挖掘社会性学习价值——访国际知名教育心理学专家凡妮莎·登嫩教授[J].开放教育研究,2017,24(2):4-12.HU Y L,ZHU J Y,GU X Q.Exploring the value of social learning by discourse analysis:An interview with Prof.Vanessa P.Dennen[J].Open Education Research,2017,24(2):4-12.(Ch).
[20]吴林静,刘清堂,毛刚,等.大数据视角下的慕课评论语义分析模型及应用研究[J].电化教育研究,2017(11):43-48.WU L J,LIU Q T,MAO G,et al.A semantic analysis method of MOOC online posts from the perspective of big data[J].E-education Research,2017(11):43-48.(Ch).
[21]CROSSLEY S,PAQUETTE L,DASCALU M,et al.Combining click-stream data with NLP tools to better understand MOOC completion[C]//International Conference,ACM,2016:6-14.