Modeling the quantification of engineering students’ academic performance and its association to dropout rates

• Autores: Sandra Maria da Conceição Pinheiro, Karla Oliveira Esquerre, Márcio A. F. Martins, Roseline Oliveira
• Localización: The International journal of engineering education, ISSN-e 0949-149X, Vol. 36, no. 1 (Parte A), 2020, págs. 201-212
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • The academic performance of most engineering students has been unsatisfactory in math and physics courses. This work proposesthe construction of a measurement for evaluating students’ academic performance based on grades and numbers of failures,associating this performance measurement to dropout percentages. This performance measurement proposed in this study aims toidentify and track students who perform poorly in the initial semesters in order to monitor them during the program. Theperformances of 1622 students in math and physics courses in the first two years of engineering programs were analyzed. Daytimeprograms analyzed were: Civil Engineering, Electrical Engineering, Mechanical Engineering, Mining Engineering, ChemicalEngineering, and Sanitary and Environmental Engineering. Evening programs were: Production Engineering, ComputerEngineering, and Control and Automation Engineering. Descriptive analyses of the data, Spearman correlation tests, Mann-Whitney tests and Poisson regression models were performed. Results obtained showed an association between the proposedperformance measurement and the students’ entrance and exit forms in the programs. It was found that the majority of studentsperformance below median in mathematics and physics courses. There was an inversely proportional relationship between theperformance measurement and dropout levels, and higher risks for dropout in the first two performance quartiles, which are thelowest. The analysis based on the Generalized Linear Model, using Poisson regression, presented consistent and statisticallysignificant estimates of relative risk. These analyses indicate that students with the lowest performances in the Analytical Geometry,Calculus I, Calculus II, and General and Experimental Physics I E courses are twice as likely to drop out of engineering courseswhen compared to students with higher performances.