ABSTRACT
The individual behavior of human beings is susceptible to influences
from their peers. It is known that contact between individuals,
both direct and indirect, can foster or inhibit a considerable range
of human characteristics and behaviors. This study aims to evaluate
the impact of the social context of undergraduate students on their
academic performance, understanding "social context"as the average
performance of classmates socially close to each student. Our
methodology involves reconstructing the underlying social network
of a class computationally, using data gathered from a questionnaire
applied to the students of the class, and testing the hypothesis that
the change in the grade of a student can be accurately modeled
as a linear function of the differences between the student’s grade
and the mean of their peers’ grades. The results show that defining
a student’s social circle as the community they belong to instead
of their set of neighbors allows for the construction of statistical
models with significatively higher predictive potential