基於語義分析網路即時回饋系統輔助線上討論成效之有效行為模式研究

Research on Mining the Effective Behavior Patterns Based on Semantic Network Instant Feedback System to Facilitate Online Discussion

蔡曉婷(Tsai Hsiao-Ting)

指導教授 ： 陳志銘

國立政治大學/文學院/圖書資訊與檔案學研究所/碩士(2019年)

https://doi.org/10.6814/NCCU201900501

摘要

隨著數位學習的蓬勃發展，線上討論被廣泛運用於輔助數位學習促進互動討論。然而，許多研究指出線上討論面臨無法掌握議題討論方向、討論缺乏深度與廣度等的問題。因此，如何促進學習者在非同步線上討論中的學習成效成為重要的研究議題。本研究採用「語義分析網路即時回饋系統(Semantic Network Instant Feedback System, SNIFS)」輔助學習者進行線上討論，並蒐集學習者的學習歷程行為進行行為分析，藉此了解學習者在使用SNIFS輔以學習過程中的有效討論行為，以引導學習者進行更有效的討論行為，進而促進線上非同步討論的學習成效。
本研究採用單組前實驗研究法，以台北市某高中二年級的學生為研究實施對象，有效樣本為34人，進行「大安溪濕地公園建置與石虎保育」主題之線上討論，過程中蒐集學習者的學習歷程數據，實驗後結合統計分析、分群分析(cluster)與滯後序列分析(LSA) 探討學習者系統操作行為模式與學習成效的關聯，以及不同電腦中介溝通(Computer-Mediated Communication, CMC)能力與人格特質學習者的學習行為模式，探討不同學習成效的學習者，是否具有不同的操作行為模式，以及不同的學習成效、不同電腦中介溝通能力以及不同的人格特質的學習者是否有具有不同的行為轉移模式。
研究結果發現，學習者使用SNIFS進行線上討論時，若能多加深入了解各貼文的完整內容與前後文，可以有效幫助學習者獲取更多資訊，對於討論議題的認知也更佳。在SNIFS回饋圖的部分，列出該組內著重的討論主題以提供統整概念的資訊呈現方式，能有效的幫助學習者增進對討論議題的認知。而學習者在討論過程中若不多方了解大家的想法，僅關注所有學習者的共同想法，在學習成效上較無法有效提升。而不同電腦中介溝通能力與人格特質的學習者其行為模式也有所不同，高CMC能力學習者將SNIFS做為查看額外資訊的工具，低CMC能力學習者則將SNIFS視為了解整體討論概念的工具；高外向性學習者較關注於本組異於他組的想法，低外向性學習者會多了解他組的想法與本組有何不同；高開放性學習者會進一步了解他組異於本組的討論內容，低開放性學習則較習慣查看大家共有的想法；高嚴謹性學習者會反覆查看許多不同資訊的完整貼文內容，低嚴謹性學習者僅會查看部分貼文的完整貼文內容，次數較少，也不常重複查看其他貼文。
最後基於研究結果，本研究提出SNIFS應用於教學之建議，以及未來可以繼續發展的研究方向。整題而言，本研究透過行為分析可以發現不同學習行為會造成不同的學習成效，並且不同電腦中介溝通能力以及不同人格特質的學習者也會有不同的學習行為，對於改善線上討論的教學方式具有貢獻。

關鍵字

線上討論 ； 社會性科學議題 ； 學習歷程分析 ； 行為分析 ； 分群分析 ； 滯後序列分析 ； 電腦中介溝通 ； 人格特質

並列摘要

Along with the boom of e-learning, online discussion is broadly applied to assist in e-learning and facilitate interactive discussion. Nevertheless, a lot of studies pointed out the problems of online discussion in the grasp of issue discussion directions and the lack of discussion depth and width. How to facilitate learners’ learning effectiveness in asynchronous online discussion therefore becomes a primary research issue. “Semantic Network Instant Feedback System (SNIFS)” is applied in this study to assist learners in online discussion and to collect learners’ learning process behaviors for behavior analyses. It is expected to understand learners’ effective discussion behaviors in the SNIFS assisted learning process and to guide learners to more effective discussion behaviors so as to facilitate the learning effectiveness of online asynchronous discussion.
With single-group prior experimental research, G11 students of a senior high school in Taipei City are the research objects. 34 effective samples precede the online discussion about the topic of “the establishment of Da-an River wetland park and the conservation of leopard cat”. Learners’ learning process data are collected in the process. Statistical analysis, cluster analysis, and lag sequential analysis (LSA) are combined, after the experiment, to discuss the correlation between learners’ system operation behavior models and learning effectiveness as well as the learning behavior models of learners with different computer-mediated communication (CMC) ability and personality. It aims to discuss various operation behavior models of learners with different learning effectiveness and different behavior shift models of learners with distinct learning effectiveness, computer-mediated communication ability, and personality.
Research results show that learners more deeply understanding the complete content and context of posts, during online discussion with SNIFS, could effectively acquire more information and present better cognition of the discussion issue. In regard to the SNIFS feedback diagram, the discussion topics emphasized in the group are listed for the information with integrated ideas to effectively assist learners in enhancing the cognition of discussion issues. Learners not understanding others’ opinions but merely concerning about the common idea of all learners in the discussion process would not effectively promote the learning effectiveness. Learners with distinct computer-mediated communication ability and personality would appear different behavior models. Learners with high CMC ability would use SNIFS for looking over extra information, while learners with low CMC regard SNIFS as the tool to understand the overall discussion concept. Learners with high extroversion concern more about the ideas different from other groups, while learners with low extroversion would understand more of the difference in the ideas from other groups. Learners with high openness would further understand different discussion contents of other groups, while learners with low openness are used to look over common ideas of all. Learners with high rigor would repeatedly look over the complete post content of different information, while learners with low rigor merely look over the complete post content of some posts, with fewer times and less review of other posts.
Based on the research results, suggestions for the application of SNIFS to teaching and the future development direction are eventually proposed in this study. Overall speaking, behavior analyses reveal that different learning behaviors would result in distinct learning effectiveness and learners with different computer-mediated communication ability and personality would show various learning behaviors. It would contribute to the improvement of teaching with online discussion.

並列關鍵字

online discussion ； socio-scientific issue ； learning process analysis ； behavior analysis ； cluster analysis ； lag sequential analysis ； computer-mediated communication ； personality

參考文獻

黃雅翎 （2018）。發展語意分析網路即時回饋系統促進線上討論成效。圖書資訊與檔案學研究所碩士論文。

朱思齊（2016）。Goldberg’s 大五因素特質量表國際人格題庫50題範本之繁體中文版發展（碩士論文）。取自臺灣博碩士論文系統。

Althaus, S. L. (1997). Computer-mediated communication in the university classroom: An experiment with on-line discussions. Communication Education, 46(3), 158–174. doi:10.1080/03634529709379088

Amichai-Hamburger, Y., Gazit, T., Bar-Ilan, J., Perez, O., Aharony, N., Bronstein, J., & Sarah Dyne, T. (2016). Psychological factors behind the lack of participation in online discussions. Computers in Human Behavior, 55, 268–277. doi:10.1016/j.chb.2015.09.009

Andresen, M. A. (2009). Asynchronous discussion forums: success factors, outcomes, assessments, and limitations. Journal of Educational Technology & Society, 12(1), 249–257.