Predicting academic performance: executive functions in university environments
This study aims to investigate the relationship between academic performance (measured by students’ grades) and executive function (EF) skills, a set of high-level cognitive skills that enable individuals to regulate their thoughts, emotions, and actions in goal-directed behavior. Executive function...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2025 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/445085 |
| Acceso en línea: | https://hdl.handle.net/2117/445085 |
| Access Level: | acceso abierto |
| Palabra clave: | Academic achievement Executive functions (Neuropsychology) Cognitive psychology Rendiment escolar Funció executiva (Neuropsicologia) Psicologia de la cognició Àrees temàtiques de la UPC::Matemàtiques i estadística |
| Sumario: | This study aims to investigate the relationship between academic performance (measured by students’ grades) and executive function (EF) skills, a set of high-level cognitive skills that enable individuals to regulate their thoughts, emotions, and actions in goal-directed behavior. Executive functions include cognitive domains such as inhibition (the ability to suppress impulsive responses), attentional control (the capacity to maintain focus), working memory (the temporary storage and manipulation of information), cognitive flexibility (the ability to shift between tasks or strategies), and sustained attention (the capacity to maintain focus over time). These components were assessed through standardized and computerized cognitive tasks (programmed in PsychoPy and deployed via Pavlovia) administered to students of the Escola Tecnica Superior d’Enginyeria Industrial de Barcelona (ETSEIB, UPC). The collected data consisted of behavioral performance metrics (e.g., reaction time and accuracy on the tasks) and academic grades, and were analyzed through three types of predictive models: Linear Regression, Random Forest, and Artificial Neural Networks (ANNs). Principal Component Analysis (PCA) was used to reduce dimensionality and improve model interpretability. Results showed that linear models, especially when combined with PCA, outperformed more complex approaches. Variables from the Wisconsin Card Sorting Task and the Stroop Task, measuring cognitive flexibility and attentional control, were the most predictive of academic grades. In contrast, working memory measures showed weaker associations. Despite initial setbacks, the research confirms that EF components are meaningful predictors of academic outcomes in analytical subjects such as statistics and data science. The findings emphasize the importance of aligning model complexity with data quality and offer insights for future studies in educational data mining. |
|---|