A Data Analysis of Student Success and Motivations in the BJCx MOOC

Yifat Amir

EECS Department
University of California, Berkeley
Technical Report No. UCB/EECS-2018-75
May 18, 2018

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-75.pdf

This technical report contains a detailed analysis of student motivations and success in the BJCx online course. Performance is quantified in the context of demographics, intentions, and motivations. Furthermore, a novel classification algorithm for categorizing students’ written goals by their overarching motivations is proposed. It utilizes natural language processing (NLP) methods such as the Topic Model and distributed vector embeddings for words. The results of its application to the BJCx dataset are used for further analysis and contextualization of student performance. Results show that performance and engagement with course material vary with motivations as well as with demographics. These results, along with the goal classifier, could be used in the future for personalization of the online learning experience to students’ goals or for the implementation of an automated intervention system to reduce student attrition. The analysis in this report is applied to BJCx, but the classification algorithm, methods of analysis, and many of the results can be generalized to other introductory computer science courses and to MOOCs in general.

Advisor: Dan Garcia


BibTeX citation:

@mastersthesis{Amir:EECS-2018-75,
    Author = {Amir, Yifat},
    Editor = {Garcia, Dan},
    Title = {A Data Analysis of Student Success and Motivations in the BJCx MOOC},
    School = {EECS Department, University of California, Berkeley},
    Year = {2018},
    Month = {May},
    URL = {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-75.html},
    Number = {UCB/EECS-2018-75},
    Abstract = {This technical report contains a detailed analysis of student motivations and success in the BJCx online course. Performance is quantified in the context of demographics, intentions, and motivations. Furthermore, a novel classification algorithm for categorizing students’ written goals by their overarching motivations is proposed. It utilizes natural language processing (NLP) methods such as the Topic Model and distributed vector embeddings for words. The results of its application to the BJCx dataset are used for further analysis and contextualization of student performance. Results show that performance and engagement with course material vary with motivations as well as with demographics. These results, along with the goal classifier, could be used in the future for personalization of the online learning experience to students’ goals or for the implementation of an automated intervention system to reduce student attrition. The analysis in this report is applied to BJCx, but the classification algorithm, methods of analysis, and many of the results can be generalized to other introductory computer science courses and to MOOCs in general.}
}

EndNote citation:

%0 Thesis
%A Amir, Yifat
%E Garcia, Dan
%T A Data Analysis of Student Success and Motivations in the BJCx MOOC
%I EECS Department, University of California, Berkeley
%D 2018
%8 May 18
%@ UCB/EECS-2018-75
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-75.html
%F Amir:EECS-2018-75