STUDENT RETENTION 

IMPROVES STUDENT RETENTION THROUGH A DROPOUT RISK PREDICTION MODEL FOR EACH INDIVIDUAL STUDENT 

It is always sad and painful to see students abandon their studies, especially if they were the first generation of their family to enter higher education. 

Is a lack of detail concerning the factors that affect the dropout risk of each student a problem for your institution? Is there a feeling of disappointment that more students do not successfully complete their degrees? 

What it does?

  Our specialized uRetention software enables estimation of dropout

risk for each student at a higher education institution at any point in time, even on the first day of class. In addition, it enables identification of those factors that are driving an increase in that risk for each student, along with their relative importance.  

The predictive model generated by uRetention is specific to each institution.

 It is created using machine learning algorithms that review historical data for the institution and its students, such as the courses included in each study plan and students’ academic performance and dropout rate. The software processes each new piece of data added and updates its parameters, improving its predictive value. 

Each institution can choose between a general predictive model for all of its students, or
‘per-segment predictive models’

that use historical data and predictions for the various campuses, degree programs, or other segments. Per-segment models can have greater predictive value, especially when students exhibit different behaviors in different segments. 

uRetention generates reports that identify all of those students with a high dropout risk.

In addition, it generates a personalized profile of each student, complete with characteristics, dropout risk, specific risk factors, and other information.  

Benefits

uRetention offers considerable benefits for higher education institutions: 

Provides rapid and automated estimation of the risk

that a student will abandon their studies. This releases teams responsible for student success 
+
from the burden of data processing tasks and enables them to redirect their efforts towards the formulation and implementation of actions to improve retention. 

Centralizes all information relating to dropout risk

for each student: data may be spread across several different systems,
+
as they relate to different and varied subjects, such as academic performance, personal and family information, fee payment history, etc. 

Generates highly precise estimates

especially when sufficient historical data are available for the generation of the predictive model.
+
This is all thanks to our advanced algorithms, and predictions can be put to the test by entering the relevant data from a previous period into our dropout risk estimation model and comparing the estimates produced with the dropouts that occurred. 

The combined power of the features detailed above result in the most important benefit of all: uRetention provides an effective contribution to improving student retention. Availability of relevant and opportune information concerning dropout risk enables the design, implementation and monitoring of action to reduce student dropout, globally and at the individual student level.  

How does it work?

Use of uRetention involves the following main steps: 

The institution provides a set of data to be loaded into the uRetention platform: 

  • Basic data: degree programs, campuses, academic periods, teaching formats, etc. 
  • Study plan data: the sequence of courses within a given study plan. 
  • Academic records, specifically data concerning the academic history of each student: course enrolment, courses completed, grades obtained, etc. 

To this are added all of the other data sets available within the institution that are of acceptable quality and which may be relevant to dropout (degree of relevance is calculated upon generation of the model in step 2):  

  • Related data: partial grades, grants, fee payment history, age, sex, marital status, secondary education grade average, socio-economic level, hometown elsewhere, etc. 

Once the data have been loaded, uRetention generates a quality report, highlighting data that are insufficiently complete or consistent to be processed by the algorithms.

Working alongside the uPlanner team, the institution decides whether to use a single predictive model for all students or specific models for each campus, degree program, teaching format, or other segment, based on the historical data available for each. 
Generally speaking, if the abundance and quality of data is sufficient, per-segment models tend to provide a better representation of the behaviors of a certain group of students and, as such, deliver more precise estimates of dropout risk. 

Using the predictive model (general or per-segment), uRetention estimates the likelihood of each student abandoning their studies. Based on these values, the software generates a report identifying those students whose dropout risk is above a certain threshold predefined by the institution. 

uRetention can also generate group reports, identifying those degree programs or cohorts whose average dropout risk is above a certain value.  

To accompany its estimates, uRetention generates a personalized profile for each student in which it details their characteristics, their dropout risk (and evolution thereof), their specific list of associated factors (each with its relative weighting), and a log of the steps taken by the institution to mitigate the student’s dropout risk.

Institutions use the data generated by the predictive model to formulate and implement actions aimed at mitigating its students’ dropout risk. As new estimates can be generated at any point in time, uRetention is a highly effective tool with which to monitor the effectiveness of the steps taken by the institution. 

Our Clients

More than 70 educational institutions trust u-planner solutions, ranging from small and medium institutions up to the most prestigious and highly ranked universities.