Application of the classification tree model in predicting learner dropout behaviour in open and distance learning

AbstractThis paper demonstrates the meaningful application of learning analytics for determining dropout predictors in the context of open and distance learning in a large developing country. The study was conducted at the Directorate of Distance Education at the University of North Bengal, West Bengal, India. This study employed a quantitative research design using a data mining approach to examine the predictive relationship between pre-entry demographic variables of learners with their dropout behaviour. Demographic and academic variables of learners, such as gender, marital and employment status, subject chosen, social status, age and income status were taken as independent or explanatory variables for predicting the response variables. Data analysis showed that the pattern of learner attrition is strongly biased towards a relatively disadvantaged category of learners, namely married and employed learners and those belonging to a higher age group. It also indicated that employed men or married women are more likely to leave due to factors such as pregnancy or relocation, and that remoteness of location of residence contributed to a high dropout rate. The results of this study provide important input for counsellors and faculty members to advise learners for best possible completion options.Keywords: data miningclassification treelearning analyticsopen and distance learningdropout predictor Notes1. Category: The social status of the learner. Scheduled Caste (SC), Scheduled Tribe (ST) and Other Backward Classes (OBC) are considered to be marginalised and vulnerable sections of the society.2. Income Status of the household is captured by the variable Below Poverty Line (BPL). The State Government identifies BPL households based upon the criteria of annual household income and these households are given financial assistance in several areas including education.3. SPSS version 19.0 by SPSS Inc. is used to analyse the data and to develop the classification tree model.