Prediction of student academic performance in an engineering dynamics course: development and validation of multivariate regression models

• Autores: Shaobo Huang, Ning Fang
• Localización: The International journal of engineering education, ISSN-e 0949-149X, Vol. 26, no. Extra 4, 2010, págs. 1008-1017
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • The present study aims to develop a validated set of multivariate regression models to predict student academic performance inEngineering Dynamics—a high-enrollment, high-impact, and core engineering course. The models include eight predictor/independentvariables that take into account student achievement before taking the course and student learning progression and achievement whiletaking the course. A total of 1,674 data points were collected from 186 undergraduate engineering students in two semesters. Fourmultivariate regression models were generated using different sample sizes of training datasets. The models were evaluated, validated,and compared using multiple criteria including R-square values, shrinkage, and prediction accuracy. The results show that the developedregression models have excellent predictability with 87–91% of the average prediction accuracy, and they have moderate predictability(46–66%) to generate good predictions (a good prediction is defined as a prediction that results in less than 10% of prediction error).