Dropping out of first-year college students is a very common problem in universities in general, but especially in engineering and degrees in the STEM field. As a public university, we are committed to use resources as efficiently as possible, so it’s important to avoid dropouts as much as possible and improve student learning and performance. In particular, there are schools where the dropout rate is significant and it is absolutely necessary to carry out a project to study the problem, find the causes and apply tools to reduce it.

Dropping out can be due to various causes related to, for example, lack of motivation of students, entrance marks, previous studies, etc. But the search for strategies to solve the problem can only be done after a systematic analysis of these causes. To go even further, the goal is to quantify the probability of college drop-out in advance by means of indicators so that preventive measures can be implemented to reduce it.

A study was carried out in pilot groups of several degrees at the Polytechnic School of Engineering of Manresa, a college of Universitat Politècnica de Catalunya · BarcelonaTech (UPC), in the framework of a teaching improvement project promoted by the UPC. Statistical research tools and new technologies were used to classify the causes, typify them and determine good parameters that allow to transform causes in quantitative variables. The main result is the construction of a mathematical model leading to the definition of risk indices. An experimental model was tested on former data, and the aim is to be tested again next academic year.

As additional result, student support and student cooperation were proven to be critical to the analysis and possibly also play a key role for resolution strategies. Also, specific apps could be implemented to help students to evaluate their personal risk in order to make their own decisions.

In particular, given the restrictions and incidences dues to covid, it is even more important to design measures to reduce dropout and encourage student learning and performance.