Abstract: The aim of this work is to provide data-driven insights regarding the factors behind dropouts in Higher Education and their impact over time. To this end, we analyzed students' data collected by a Higher Education Institute over the last 11 years and we explored how socio-economic and academic changes may have impacted student dropouts and how these changes may have been reflected or captured by students' data. To analyze the data, we engineered features that may predict student dropouts on three dimensions: academic background, students' performance and students' effort. Then we carried out a correlation analysis to investigate the potential relationship between these features and dropouts, we performed a multivariate analysis of variance (MANOVA) to investigate whether the engineered features change significantly among student cohorts with different admission year and, finally, we carried out a regression analysis to confirm that the engineered features' impact on predicting dropouts changes over the years. The results suggest that the importance of features regarding the academic background of students (such as the students' prior experience with the academic institution), and the effort students make (for example, the number of days students spend on academic leave) may change over time. On the contrary, performance-based features (such as credit points and grades) do interact with time suggesting that performance measures are stable predictors of dropouts over time. On the basis of the findings, we argue that the performance of prediction models for assessing students at risk of dropping out of their studies can be affected by the age of data and we outline the possibility of including a forgetting factor for non-recent data in order to leverage their impact on prediction performance.