Work in progress - predicting retention in engineering using an expanded scale of affective characteristics from incoming students

• Main Authors: Lin, J.J., Imbrie, P.K., Reid, K.J.
• Format: Conference Proceeding
• Language: English
• Subjects: Accuracy; Affective characteristics; Artificial neural networks; Data engineering; Educational institutions; Engineering education; first year retention; Logistics; Neural networks; predictive modeling; Predictive models; Teamwork; Testing
• ISSN: 0190-5848; 2377-634X
• DOI: 10.1109/FIE.2009.5350877

Earlier research published by the authors has demonstrated an improvement in prediction capability when incorporating nine affective characteristics into an artificial neural network retention model with eleven cognitive factors. Models developed previously have achieved moderate success with overall prediction accuracy above 70%. In this follow-up study, in order to develop new knowledge on relationships between other affective factors and student persistence, and further improve our capability to predict students' retention, five carefully selected affective characteristics are added to the existing retention model. These promising new affective factors are: goal orientation, implicit beliefs, intent to persist, social climate and self worth. New retention models based on logistic regression and neural networks are developed to identify the significant predictors among these new affective characteristics, and evaluate the overall predictive performance of new models incorporating them. The prediction accuracy results of models using only these new factors, as well as models including both new and existing factors are then compared with performance of previously published models. Upon completion of this project, confirmed significant predictors and their effects on predictive retention models will be reported. The potential engineering education applications based on these new findings will also be discussed.