Uma avaliação sobre o desempenho de alunos do IFPB em olimpíadas de informática

Abstract

<p>In order to recruit and train students to participate in Olympiads of Informatics (OI), the Olympic Programming Project group of IFPB Campina Grande Campus (POP-CG) has started to perform OI simulations. For this purpose, an online judge was used to create programming competitions, ForCode (https://github.com/ LADOSSIFPB/POP-Judge). Mining students’ performance data (number of submissions and hits, programming language used) from the code repository generated by ForCode may reveal what programming skills should be improved. Associating this information with holistic data about students (eg.: schooling, socioeconomic status, and so on) may indicate which factors impact (positively or negatively) on the performance of these students. In this article, we joined performance data of 47 students participating in the OI simulation, conducted in November 2015, to holistic data collected from the IFPB academic management system: the year of schooling. Then, we evaluated statistically the relation between the performance’s simulated OI with year of schooling. The results point out that, regardless of programming language used, the “trial and error” learning strategy is the one that prevails among the students with the best performance. When considering schooling, the average score of the students in the 2nd year is 15% higher than the average of the students in the 1st year, while the average score of the students in the 4th year is 78% higher than those of the students of the 2nd year 105% higher than those obtained in the 1st year</p>