Early Prediction of Student Dropout and Performance in MOOCs using Higher Granularity Temporal Information. Ye, C. & Biswas, G. Journal of Learning Analytics, 1(3):169-172, 2014. ![Early Prediction of Student Dropout and Performance in MOOCs using Higher Granularity Temporal Information [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper abstract bibtex Our project is motivated by the early dropout and low completion rate problem in MOOCs. We have extended traditional features for MOOC analysis with richer and higher granularity information to make more accurate predictions of dropout and performance. The results show that finer-grained temporal information increases the predictive power in the early phases of the Pattern-Oriented Software Architectures (POSA) MOOC offered in summer 2013 by Vanderbilt University. As a next step, we plan to develop unsupervised learning methods with our extended feature set to define profiles that can be used for effective scaffolding and feedback. 1. MOTIVATION The popularity and large initial enrollments in Massive Online Open Courses (MOOCs), driven by their wide accessibility, relative openness, and the reputation of the instructors and institutions offering them has created significant interest in analyzing and characterizing student learning behaviours to support scaffolding in these systems (Brown, 2013; Matkin, 2013). A common pattern in MOOCs has been the large number of no shows, early dropout, and low completion rates (Clow, 2013; Hill, 2013). For example, in the POSA course offered by Vanderbilt University, there were 31,053 enrollees: only 6,953 students watched a video lecture in week 1; of these, only 1,699 students also took a quiz that week. At the end of the course, 1,051 students got a pass certificate and 592 obtained a distinction grade. These numbers motivate our research question: Can we derive accurate and reliable early predictors of student dropout and performance in MOOC environments? Early dropout predictions will provide a framework for developing scaffolding mechanisms in MOOCs that provide individualized guidance and small-group support, which should significantly increase retention rates. For example, avoiding the video-quizzes embedded in the lectures is a good indicator of early dropout. For students who tend to ignore the embedded quizzes, therefore, we may briefly explain the importance of formative assessments, provide feedback that links the quiz questions to the topic-related material in the video, and provide more explanation for the alternatives provided for the answer. A number of studies have been conducted on student retention in traditional settings, e.g., dropout analysis of university freshmen (Dekker, Pechenizkiy, & Vleeshouwers, 2009). MOOCs, in their current form, represent a different form of learning and target a greater diversity of learners as compared to
@article{
title = {Early Prediction of Student Dropout and Performance in MOOCs using Higher Granularity Temporal Information},
type = {article},
year = {2014},
keywords = {MOOCs,dropout,feature engineering,performance,prediction},
pages = {169-172},
volume = {1},
id = {f44400d5-89bb-311c-aae0-2300061bcdd8},
created = {2016-02-07T00:01:23.000Z},
file_attached = {true},
profile_id = {9c969e40-69e4-31d7-92aa-104e4ff537e8},
group_id = {a7fbd406-18f3-3b0b-bf98-38c617d1e04e},
last_modified = {2017-03-14T17:40:34.521Z},
read = {false},
starred = {false},
authored = {false},
confirmed = {false},
hidden = {false},
private_publication = {false},
abstract = {Our project is motivated by the early dropout and low completion rate problem in MOOCs. We have extended traditional features for MOOC analysis with richer and higher granularity information to make more accurate predictions of dropout and performance. The results show that finer-grained temporal information increases the predictive power in the early phases of the Pattern-Oriented Software Architectures (POSA) MOOC offered in summer 2013 by Vanderbilt University. As a next step, we plan to develop unsupervised learning methods with our extended feature set to define profiles that can be used for effective scaffolding and feedback. 1. MOTIVATION The popularity and large initial enrollments in Massive Online Open Courses (MOOCs), driven by their wide accessibility, relative openness, and the reputation of the instructors and institutions offering them has created significant interest in analyzing and characterizing student learning behaviours to support scaffolding in these systems (Brown, 2013; Matkin, 2013). A common pattern in MOOCs has been the large number of no shows, early dropout, and low completion rates (Clow, 2013; Hill, 2013). For example, in the POSA course offered by Vanderbilt University, there were 31,053 enrollees: only 6,953 students watched a video lecture in week 1; of these, only 1,699 students also took a quiz that week. At the end of the course, 1,051 students got a pass certificate and 592 obtained a distinction grade. These numbers motivate our research question: Can we derive accurate and reliable early predictors of student dropout and performance in MOOC environments? Early dropout predictions will provide a framework for developing scaffolding mechanisms in MOOCs that provide individualized guidance and small-group support, which should significantly increase retention rates. For example, avoiding the video-quizzes embedded in the lectures is a good indicator of early dropout. For students who tend to ignore the embedded quizzes, therefore, we may briefly explain the importance of formative assessments, provide feedback that links the quiz questions to the topic-related material in the video, and provide more explanation for the alternatives provided for the answer. A number of studies have been conducted on student retention in traditional settings, e.g., dropout analysis of university freshmen (Dekker, Pechenizkiy, & Vleeshouwers, 2009). MOOCs, in their current form, represent a different form of learning and target a greater diversity of learners as compared to},
bibtype = {article},
author = {Ye, Cheng and Biswas, Gautam},
journal = {Journal of Learning Analytics},
number = {3}
}
Downloads: 0
{"_id":"Te8XZBHMKZx2Lf4nv","bibbaseid":"ye-biswas-earlypredictionofstudentdropoutandperformanceinmoocsusinghighergranularitytemporalinformation-2014","authorIDs":[],"author_short":["Ye, C.","Biswas, G."],"bibdata":{"title":"Early Prediction of Student Dropout and Performance in MOOCs using Higher Granularity Temporal Information","type":"article","year":"2014","keywords":"MOOCs,dropout,feature engineering,performance,prediction","pages":"169-172","volume":"1","id":"f44400d5-89bb-311c-aae0-2300061bcdd8","created":"2016-02-07T00:01:23.000Z","file_attached":"true","profile_id":"9c969e40-69e4-31d7-92aa-104e4ff537e8","group_id":"a7fbd406-18f3-3b0b-bf98-38c617d1e04e","last_modified":"2017-03-14T17:40:34.521Z","read":false,"starred":false,"authored":false,"confirmed":false,"hidden":false,"private_publication":false,"abstract":"Our project is motivated by the early dropout and low completion rate problem in MOOCs. We have extended traditional features for MOOC analysis with richer and higher granularity information to make more accurate predictions of dropout and performance. The results show that finer-grained temporal information increases the predictive power in the early phases of the Pattern-Oriented Software Architectures (POSA) MOOC offered in summer 2013 by Vanderbilt University. As a next step, we plan to develop unsupervised learning methods with our extended feature set to define profiles that can be used for effective scaffolding and feedback. 1. MOTIVATION The popularity and large initial enrollments in Massive Online Open Courses (MOOCs), driven by their wide accessibility, relative openness, and the reputation of the instructors and institutions offering them has created significant interest in analyzing and characterizing student learning behaviours to support scaffolding in these systems (Brown, 2013; Matkin, 2013). A common pattern in MOOCs has been the large number of no shows, early dropout, and low completion rates (Clow, 2013; Hill, 2013). For example, in the POSA course offered by Vanderbilt University, there were 31,053 enrollees: only 6,953 students watched a video lecture in week 1; of these, only 1,699 students also took a quiz that week. At the end of the course, 1,051 students got a pass certificate and 592 obtained a distinction grade. These numbers motivate our research question: Can we derive accurate and reliable early predictors of student dropout and performance in MOOC environments? Early dropout predictions will provide a framework for developing scaffolding mechanisms in MOOCs that provide individualized guidance and small-group support, which should significantly increase retention rates. For example, avoiding the video-quizzes embedded in the lectures is a good indicator of early dropout. For students who tend to ignore the embedded quizzes, therefore, we may briefly explain the importance of formative assessments, provide feedback that links the quiz questions to the topic-related material in the video, and provide more explanation for the alternatives provided for the answer. A number of studies have been conducted on student retention in traditional settings, e.g., dropout analysis of university freshmen (Dekker, Pechenizkiy, & Vleeshouwers, 2009). MOOCs, in their current form, represent a different form of learning and target a greater diversity of learners as compared to","bibtype":"article","author":"Ye, Cheng and Biswas, Gautam","journal":"Journal of Learning Analytics","number":"3","bibtex":"@article{\n title = {Early Prediction of Student Dropout and Performance in MOOCs using Higher Granularity Temporal Information},\n type = {article},\n year = {2014},\n keywords = {MOOCs,dropout,feature engineering,performance,prediction},\n pages = {169-172},\n volume = {1},\n id = {f44400d5-89bb-311c-aae0-2300061bcdd8},\n created = {2016-02-07T00:01:23.000Z},\n file_attached = {true},\n profile_id = {9c969e40-69e4-31d7-92aa-104e4ff537e8},\n group_id = {a7fbd406-18f3-3b0b-bf98-38c617d1e04e},\n last_modified = {2017-03-14T17:40:34.521Z},\n read = {false},\n starred = {false},\n authored = {false},\n confirmed = {false},\n hidden = {false},\n private_publication = {false},\n abstract = {Our project is motivated by the early dropout and low completion rate problem in MOOCs. We have extended traditional features for MOOC analysis with richer and higher granularity information to make more accurate predictions of dropout and performance. The results show that finer-grained temporal information increases the predictive power in the early phases of the Pattern-Oriented Software Architectures (POSA) MOOC offered in summer 2013 by Vanderbilt University. As a next step, we plan to develop unsupervised learning methods with our extended feature set to define profiles that can be used for effective scaffolding and feedback. 1. MOTIVATION The popularity and large initial enrollments in Massive Online Open Courses (MOOCs), driven by their wide accessibility, relative openness, and the reputation of the instructors and institutions offering them has created significant interest in analyzing and characterizing student learning behaviours to support scaffolding in these systems (Brown, 2013; Matkin, 2013). A common pattern in MOOCs has been the large number of no shows, early dropout, and low completion rates (Clow, 2013; Hill, 2013). For example, in the POSA course offered by Vanderbilt University, there were 31,053 enrollees: only 6,953 students watched a video lecture in week 1; of these, only 1,699 students also took a quiz that week. At the end of the course, 1,051 students got a pass certificate and 592 obtained a distinction grade. These numbers motivate our research question: Can we derive accurate and reliable early predictors of student dropout and performance in MOOC environments? Early dropout predictions will provide a framework for developing scaffolding mechanisms in MOOCs that provide individualized guidance and small-group support, which should significantly increase retention rates. For example, avoiding the video-quizzes embedded in the lectures is a good indicator of early dropout. For students who tend to ignore the embedded quizzes, therefore, we may briefly explain the importance of formative assessments, provide feedback that links the quiz questions to the topic-related material in the video, and provide more explanation for the alternatives provided for the answer. A number of studies have been conducted on student retention in traditional settings, e.g., dropout analysis of university freshmen (Dekker, Pechenizkiy, & Vleeshouwers, 2009). MOOCs, in their current form, represent a different form of learning and target a greater diversity of learners as compared to},\n bibtype = {article},\n author = {Ye, Cheng and Biswas, Gautam},\n journal = {Journal of Learning Analytics},\n number = {3}\n}","author_short":["Ye, C.","Biswas, G."],"urls":{"Paper":"https://bibbase.org/service/mendeley/bfdabac2-d7f2-3c5b-aa7a-06431c0ae35e/file/6e929266-d446-ac1b-4e03-ca6315c48870/2014-Early_Prediction_of_Student_Dropout_and_Performance_in_MOOCs_using_Higher_Granularity_Temporal_Information.pdf.pdf"},"bibbaseid":"ye-biswas-earlypredictionofstudentdropoutandperformanceinmoocsusinghighergranularitytemporalinformation-2014","role":"author","keyword":["MOOCs","dropout","feature engineering","performance","prediction"],"downloads":0},"bibtype":"article","creationDate":"2020-02-06T23:44:30.828Z","downloads":0,"keywords":["moocs","dropout","feature engineering","performance","prediction"],"search_terms":["early","prediction","student","dropout","performance","moocs","using","higher","granularity","temporal","information","ye","biswas"],"title":"Early Prediction of Student Dropout and Performance in MOOCs using Higher Granularity Temporal Information","year":2014}