基於社交網路分析之線上討論活動同儕互動分析

Abstract

許多網路合作學習相關研究肯定同儕互動對學習的正向影響，因此深入探討同儕互動的行為與態度如何影響學習是相當重要的。然而多數研究對於互動行為的分析通常以整個群體或是個別學習者的表徵數據為衡量指標，容易忽略隱藏在社群成員中的隱性互動關係。透過社交網路分析的技術可以瞭解出社群成員個體與群體的相互關係，而能更清楚客觀地描述網路學習中互動行為的意義。 因此本研究利用社交網路分析技術，探究互動行為、互動態度、與學習表現間的關係。我們建構了一個社交學習平台，模擬坊間流行社交平台之互動模式，讓學生可利用藉此平台進行討論互動，透過平台紀錄，討論學習者在網路學習平台上的分享、閱讀、提問、回應以及贊同等行為，並分析線上討論平台上的社交互動如何影響學習，以期能歸納出學習者之互動行為、態度、與學習表現的關係。針對台北市高中二年級兩班學生74人進行實驗分析，結果發現： 1. 學習者在進行線上討論活動時，其討論品質的高低與學習表現有關，而低學習成就組的學生透過閱讀活動，可能有較好的學習成就 2. 學習者的回應行為與學習表現有關係，群體中位於討論互動中心的成員也有較高的學習表現。 3. 觀察及反思態度較積極的學習者，其發表的材料受關注的程度也相對較高，也比較容易得到同儕支持。 4. 競爭態度較高的學習者，會有較少的討論回應，在缺少反思、回饋的情況下，學習者可能有較低的學習成就。 藉由討論平台上的互動分析的結果，我們得以更深入了解同儕互動行為、學習態度、與學習成就的關聯，教師可藉此設計更合宜的討論活動或社交學習平台，以促進合適的互動行為與態度，進而有更好的學習表現。

Much existing research of online collaborative learning argue that peer interaction plays an important role in effective collaborative learning. To investigate how peer interaction affects learning, several studies have devoted to discuss peer interaction by analyzing interaction behaviors based on subjective observation or questionnaire study, which lacks a precise and objective analysis of interactive relationships among group members. Instead of using traditional methodologies, this study employs the techniques of social network analysis to mine students’ interaction behaviors in a more structural way to investigate more deeply about the relationships among the interaction behaviors, the attitude toward peer interaction, and learning performance. This research conducted an empirical study to investigate the relationships among interaction behaviors, attitude toward interaction, and learning performance based on social network analysis. We developed a social learning platform embedded with interactive functionalities to simulate a social media for collaborative learning, on which students could discuss with peers about the learning topics by actions of posting, reading, inquiring, replying, and liking. Seventy-four eleventh-grade students in northern Taiwan participated in the experiment and all of their online discussion behaviors were recorded and then analyzed by social network analysis. The research findings are: 1. The discussion quality was highly correlated with learning performance. That is, students performed better in the achievement test had higher quality discussions. In the other hand, reading could improve low achievers’ scores in the test. 2. The reply frequency was correlated with learning performance related with learning effect, and the members situated in the discussion center of social community are better in learning effect. 3. The learning materials posted by the students who had more positive attitudes toward observation and reflection were accessed by peers more frequently. 4. The competitive oriented students involved less in replying peers’ discussions. Since they had fewer opportunities to reflect or give feedback to the discussion, they performed worse than others in the achievement test. From the experiment results, we understand more about the relationships among the interaction behaviors, attitude toward interaction, and learning performance, based on which instructors/researchers can develop effective online discussion activities or learning platforms to inspire suitable interaction behaviors and attitude, and accordingly improve students’ learning performance platform.