ACCURACY OF LINEAR MODELS PREDICTING THE ACADEMIC PERFORMANCE OF EDUCATIONAL INNOVATION PROJECT STUDENTS

In previously reported studies a linear regression model was obtained where the marks of educational innovation project students from the Valencia University were considered as the explained variable and different factors related to it were employed as explicative variables. Although the observed model showed good performance for the considered set of student marks, its coefficient of determination (R-2) value displayed that it was not significantly representative for the whole students group. Indeed it could be appreciated how it was representative for a subset of the considered group. Now we aim to validate the results obtained for different year classes. Employing a similar methodology we look for significant differences among the obtained models and their cause supposing they exist. Finally, in case these results could not be validated, the characteristics of students' subsets will be studied in order to obtain more reliable models.