This research proposes analyzing the impact of local, national, and international migration on the academic performance of tenth- and eleventh-grade high school students in mathematics, economics, and political science at a college in Medellín, Colombia. The significant increase in migrant students, particularly international migrants from Venezuela and internal migrants from other regions of Colombia, highlights the importance of comparing their academic performance with that of native students. This comparison is grounded in an educational inclusion and cultural diversity approach, aligning with the global objective of equitable and accessible education.

We examine the relationship between academic performance and multiple student attributes using statistical techniques. Our methodology combines the machine learning method of random forest with student characterization surveys to capture the social and cultural conditions that shape the educational process. Our analysis is based on a sample of 150 students and considers the following variables: socioeconomic background (e.g., family size, parental education and employment, economic resources, housing conditions, and socioeconomic stratum); school context and mobility (e.g., neighborhood, commuting time, displacement status, distance to school, and transportation); family and social environment (e.g., guardian type, family support, quality of relationships, and exposure to discrimination or public safety issues); and individual characteristics (e.g., gender, age, health and mental health status, study time, extracurricular activities, substance use, absenteeism, and academic projection). Academic performance in mathematics and economics serves as the main outcome variable. We expect our findings to inform the design of inclusive pedagogical strategies that enhance educational quality and foster equity in diverse, mobile student contexts.

Keywords: Migration, academic performance, mathematics, economics and politics sciences, educational inclusion, machine learning.