Dados do Trabalho

Título

A BIVARIATE DISCRETE SURVIVAL MIXED MODEL TO EVALUATE PERFORMANCE OF STUDENTS IN CALCULUS I AND LINEAR ALGEBRA

Resumo

In this paper our interest relies on analyze the performance of undergraduate students in Calculus I and Linear Algebra and identify which characteristics may be associated with the chances of a student be approved in one of those disciplines. The database is from the University of Campinas, one of the top Public Research Universities in Brazil, with records of 9,130 students who enrolled at the university between 2009 and 2015 in courses such that Calculus I or Linear Algebra were required disciplines. The variable of interest is the number of times a student take each one of those course until he/she is approved, treating this as discrete survival variable, some students may take 7 or more time the discipline until be approved. We proposed a bivariate discrete survival model to analyses jointly the time until pass in both discipline taking account the strong correlation among them via a gaussian random component.

Palavras-chave

student performance, survival analysis, discrete time, censored data, mixed model