Understanding topic duration in Twitter learning communities using data mining. (10th December 2021)

• Record Type: Journal Article
• Authors: Arslan, Okan; Xing, Wanli; Inan, Fethi A.; Du, Hanxiang
• Is Part Of: Journal of computer assisted learning. Volume 38:Number 2(2022)
• Page Start: 513
• Page End: 525
• Publication Date: 2021-12-10
• Subjects: educational data mining -- learning analytics -- learning networks -- social media -- teacher professional development; Computer-assisted instruction -- Periodicals; 371.334
• DOI: 10.1111/jcal.12633
• Languages: English
• ISSNs: 0266-4909

Abstract

Abstract: Background: There has been increasing interest in online professional learning networks in a variety of social media platforms, especially in Twitter. Twitter offers immediacy, personalization, and support of networks to increase professional knowledge and the sense of membership. Knowing the topics discussed in Twitter and the factors that affect the duration of a topic would help to sustain and reconstruct Twitter‐based professional learning activities. Objectives: The purpose of this study is to analyse the topics discussed and what factors affect the duration of a specific topic in 6 years within a virtual professional learning network (VPLN) using #Edchat in Twitter, based on media richness features. Methods: Internet‐mediated research and digital methods are used for data collection and analysis. Various text, natural language processing, and machine learning algorithms were used along with the quantitative multilevel models. This study examined 504, 998 tweets posted by 72, 342 unique users by using #Edchat. Results: There were 150 topics discussed over the 6 years and multilevel random intercept regression model revealed that a specific topic discussed in the #Edchat VPLN is discussed longer when it has more tweets, rather than retweets, posted by a high number of different users along with moderate text, high or moderate mentions with more hashtags. Takeaways: The study developed an automated social media richness feature extraction framework that can be adapted for other theoretical applications in educational context. Emergent topics discussed in Twitter among #Edchat VPLN members for professional development were identified. It extends the social media richness theory for educational context and explore factors that affect an online professional learning activity in Twitter. Lay Description: What is already known about this topic: Professional Learning Communities have started to use both synchronous and asynchronous Web 2.0 and social media platforms where they may exchange information and develop a sense of belonging with others who share common interests. Synchronous online activity in social networking sites, such as Twitter, can be an effective, informal and a free way to convey information and create and develop personalized networks. Twitter offers immediacy, personalization, and support of networks to increase professional knowledge and the sense of membership. There have been a great number of hashtags in Twitter for educational context and those hashtags are used by teachers to find information and resources and gain new perspectives and ideas from their colleagues or experts. The richness of a social media post varies according to how it is structured and constructed. What this paper adds: By using #edchat in Twitter, educators discussed 150 topics in 6 years. The most discussed topics are Creating, changing school culture; Classroom management, teaching methods; Classroom settings, Educational technologies; Support – Needs; Students' subject skills and interest; and School environment. A specific topic stays longer when it has more tweets, rather than retweets, posted by a high number of different users. Topics that have moderate text, high or moderate mentions with more hashtags are discussed longer on Twitter. The duration of a topic can be changed according to the educators' behaviours in a synchronous online chat. Implications for practice and/or policy: Importance of social media for professional learning and how to sustain a topic for better learning and understanding. Developing and applying an automated computational discourse analysis social media learning. Applying media richness theory to understand the affordances of social media learning. Quantifying the influence of various factors on the discourse in social media learning.