Methodology

1. Sources

To investigate student academic performance and engagement, we worked with a student performance dataset from the UC Irvine Machine Learning Repository. This dataset compiles student demographic, familial, and educational background information from two secondary schools in Portugal—Gabriel Pereira (GP) and Mousinho da Silveira (MS). It includes key factors such as school type, student sex, age, home environment, and parental education and occupation. By analyzing these variables, we aimed to explore the relationships between external influences and student success, particularly in Mathematics and Portuguese, the two core subjects recorded in the dataset.

To further contextualize academic performance, we examined family-related factors such as parental marital status, family size, and parental support for education. These variables offered insight into how family structures might shape student outcomes. 

Before interpreting our data, we critically assessed potential limitations in its collection, structure, and scope. We considered how the dataset’s reliance on numeric indicators reinforces a grade-centric approach to academic success, potentially obscuring other meaningful dimensions of student learning and development. 

To complement the dataset, we reviewed peer-reviewed literature on educational inequality, the impact of socioeconomic status on learning outcomes, and broader systemic factors influencing student success. This provided a foundation for understanding how academic performance is shaped not only by individual effort but also by institutional and social conditions. By synthesizing these sources, we aimed to construct a more comprehensive narrative about the complexities of student achievement beyond the limitations of the dataset. More details on our sources and critical approach can be found in our annotated bibliography.