SENS: Network analytics to combine social and cognitive perspectives of collaborative learning
SENS: Network analytics to combine social and cognitive perspectives of collaborative learning
复制标题
DOI：
10.1016/j.chb.2018.07.003
复制
发表时间：
2019-03-01
影响因子：
9.9
通讯作者：
Shaffer, David Williamson
中科院分区：
文献类型：
--
作者：
Gasevic, Dragan;Joksimovic, Srecko;Shaffer, David Williamson
在本文中，我们提出了一种分析协作学习的新方法。该方法假设，从社会关系和话语内容分析中出现的协作学习的不同维度可以建模为网络。因此，社交网络分析 (SNA) 和认知网络分析 (ENA) 分析的结合可以检测有关学习者对协作学习文献所描述的角色的制定的信息：认知和社会维度的集合，其特征是与适当的人就适当的内容进行交互。所提出的方法被称为社会认知网络签名（SENS），并被定义为这两种互补网络分析技术的组合。所提出的 SENS 方法对通过主要 MOOC 平台提供的大规模开放在线课程 (MOOC) 中执行的协作活动产生的数据进行了检查。对 MOOC 中收集的数据集进行的一项研究结果表明，SNA 和 ENA 产生了互补的结果，这些结果可以 i) 解释塑造社会关系创建以及与不同网络角色相关的协作过程； ii) 描述低绩效学习者群体和高绩效学习者群体之间的差异； iii) 显示从 SNA 和 ENA 得出的综合属性如何预测学业成绩。
In this paper, we propose a novel approach to the analysis of collaborative learning. The approach posits that different dimensions of collaborative learning emerging from social ties and content analysis of discourse can be modeled as networks. As such, the combination of social network analysis (SNA) and epistemic network analysis (ENA) analysis can detect information about a learner's enactment of what the literature on collaborative learning has described as a role: an ensemble of cognitive and social dimensions that is marked by interacting with the appropriate people about appropriate content. The proposed approach is named social epistemic network signature (SENS) and is defined as a combination of these two complementary network analytic techniques. The proposed SENS approach is examined on data produced in collaborative activities performed in a massive open online course (MOOC) delivered via a major MOOC platform. The results of a study conducted on a data set collected in a MOOC suggest SNA and ENA produce complementary results which can i) explain collaboration processes that shaped the creation of social ties and that were associated with different network roles; ii) describe differences between low and high performing groups of learners; and iii) show how combined properties derived from SNA and ENA predict academic performance.