Undirected Bipartite Networks as an Alternative Methodology to Probabilistic Exploration: Online Interaction and Academic Attainment in MOOC

Undirected Bipartite Networks as an Alternative Methodology to Probabilistic Exploration: Online Interaction and Academic Attainment in MOOC

Juan-Francisco Martínez-Cerdá (Universitat Oberta de Catalunya (UOC), Spain) and Joan Torrent-Sellens (Universitat Oberta de Catalunya (UOC), Spain)
DOI: 10.4018/978-1-5225-2814-2.ch005
OnDemand PDF Download:
List Price: $37.50


This chapter explores how graph analysis techniques are able to complement and speed up the process of learning analytics and probability theory. It uses a sample of 2,353 e-learners from six European countries (France, Germany, Greece, Poland, Portugal, and Spain), who were enrolled in their first year of open online courses offered by HarvardX and MITX. After controlling the variables for socio-demographics and online content interactions, the research reveals two main results relating student-content interactions and online behavior. First, a multiple binary logistic regression model tests that students who explore online chapters are more likely to be certified. Second, the authors propose an algorithm to generate an undirected bipartite network based on tabular data of student-content interactions (2,392 nodes, 25,883 edges, a visual representation based on modularity, degree and ForceAtlas2 layout); the graph shows a clear relationship between interactions with online chapters and chances of getting certified.
Chapter Preview

1. Introduction

During the last decades, the increase in the computer-based calculation capabilities and advances in data sciences (i.e. the integration of data sciences into educational environments), have enabled the development of research oriented to educational environments, especially through two approaches well described by Siemens and Baker (2012): i) educational data mining (EDM), aimed at statistical data analysis about students, settings, and automated findings; and ii) learning analytics (LA), oriented to understand the complexity of learning contexts by using mixed methodology for data analysis. In fact, the origin of EDM and the use of data mining in educational context comes from 1995 (Romero & Ventura, 2007; Baker & Yacef, 2009). Thus, LA has tried to produce knowledge about the whole learning ecosystem by analysing the huge number of interactions that exist between students, instructors, and content. In this way, many technologies and solutions have been used, such as business intelligence, web analytics, academic analytics, and action analytics (Elias, 2011). Recent research on using LA beyond traditional school environments have been conducted, such as in workplaces (Laat & Schreurs, 2013) and connecting EDM, LA, big data, and Massive Open Online Courses (MOOC) (Nisar, Fard, & Miller, 2013; Calvet Liñán & Juan Pérez, 2015).

On the other hand, since late last century international organizations have been drafting proposals aimed at the use of online education systems as tools for the qualification of workers through lifelong learning (Delors et al., 1996; Aceto, Borotis, Devine, & Fischer, 2014). This has led to a large increase in: i) use of new pedagogies and related technologies in open online courses, such as HarvardX and MITx (Anderson & McGreal, 2012), which have facilitated mobile and ubiquitous learning (Kinshuk, Hui-Wen, Sampson, & Chen, 2013); and ii) a wealth of educational data that have to be analysed according to several ethical principles (Ferguson, 2012; Slade & Prinsloo, 2013).

Therefore, LA should continue developing new interesting solutions on this scientific field. In this issue, note that new graph analysis techniques are being very useful for assessing of our daily scientific work (Brandes, Kenis, & Raab, 2006). These new approaches should go beyond the interesting proposals that have combined quantitative and qualitative methodologies in LA (Fournier, Kop, & Sitlia, 2011), and they should analyse behavior of online students and relationships between interaction with online resources and academic attainment in the context of MOOC. In this way, LA should be complemented with graph analytics methodologies, which have been useful for visualization MOOC data (Coffrin, Corrin, de Barba, & Kennedy, 2014), new frameworks (Satish et al., 2014), and interactive data repository for visualization (Rossi & Ahmed, 2015). Moreover, these new approaches and methodologies should be tested with traditional statistical techniques, helping to understand whether standard statistics could take advantage of them.

Complete Chapter List

Search this Book: