Statistical Analysis of Socio-Economic Factors as Predictors of Performance in a Brazilian Science Fair

Science fairs provide an opportunity for students to showcase their scientific endeavours and practice important skills. Additionally, judging and awarding are very important parts of the process, providing feedback and recognition to participants. To maintain fairness in the competition and avoid reinforcing systemic inequalities, it’s crucial to examine the current landscape of science fairs and identify potential areas for improvement. We explore this issue by analyzing the impact of socio-economic factors on students' performance in the Brazilian Fair of Young Scientists. Focusing on gender, race, English level, school type, and location, we use linear and logistic regression models to predict students’ scores and awards. The results suggest that school type and English level, features associated with access to resources, significantly influence success in the competition. Notably, the only criterion that did not show a significant difference in scores for these features was the one about the project idea, indicating that innovative research questions from underserved students may have been hindered by a lack of resources for project development. By addressing these disparities, science fair organizers and policymakers can foster more equitable opportunities and encourage the next generation of scientists, regardless of their background.