Classification Tools to Assess Critical Thinking in Automotive Engineering Students

Abstract

Inadequate conflict resolution skills in automotive engineering students can have negative consequences in the workplace. The development of mathematical logical thinking can help students develop critical analysis skills, improve problem-solving ability, develop reasoning skills, and effective communication, enabling them to deal effectively with conflicts and find creative solutions. This research aims to identify predictors of problem-solving ability using classification algorithms. Methodology: In this study, three classification algo-rithms were applied and the KDD process was used to identify predictors of problem-solving ability. The data set includes 60 records of students from the automotive engineering program at Universidad Equinoccial in Quito, Ecuador, to whom three tools were applied: a sociodemographic card, a Shatnawi test related to mathematical logical thinking, and a Watson Glaser test on conflict resolution ability. Results: The best classification model is the K-nearest neighbors’ algorithm and its predictive ability is very good, with a true positive rate versus false positive rate AUC of 0.75, along with a good performance in classifying negative cases. The model can be improved with increased sampling, cross-validation, or hyper-parameter adjustment. Conclusion: Age and mathematical logical thinking are strongly associated with conflict resolution ability. In future research it is important to consider additional variables such as experience in problem-solving projects, technical knowledge and communication skills; to explore the use of more advanced machine learning algo-rhythms; to design specific educational interventions based on the development of mathematical logical thinking; or to compare conflict resolution ability between different engineering disciplines.