How well can completion of online courses be predicted using binary logistic regression?
2016 (English)In: Proceedings of IRIS39, Information Systems Research Seminar in Scandinavia, Ljungskile, August 7-10, 2016 / [ed] Pareto, Lena, Svensson, Lars, Lundin, Johan, Lundh Snis, Ulrika Lundh Snis, 2016, 1-12 p.Conference paper (Other academic)
This article uses binary logistic regression to create models for predicting course performance. The data used is the data-trail left by students activities on a discussion forum while attending an online course. The purpose of the study is to evalute how well models based on binary logistic regression can be used to predict course completion.Three sets of data was used for this. One set collected at the end of the course, one collected after 75% of the course and one set collected after half the course. The result of the study says that it's possible to design models with an accuracy of between 70% and 80% using these methods, regardless of what time is used.
Place, publisher, year, edition, pages
2016. 1-12 p.
Online education, statistics, course completion, online forums, educational data mining
Information Systems, Social aspects
Research subject SOCIAL SCIENCE, Informatics
IdentifiersURN: urn:nbn:se:hv:diva-10385ISBN: 978-91-87531-38-5 (USB)OAI: oai:DiVA.org:hv-10385DiVA: diva2:1059250
IRIS39, Information Systems Research Seminar in Scandinavia, Ljungskile, August 7-10, 2016