Realise and @Cognassist pilot research on predicting dropout

Realise is one of the largest providers of apprenticeships and adult education programmes in England. They are passionate about continually driving improvements and reducing avoidable dropout. While some degree of dropout is inevitable, Realise wants to take a proactive approach to tackling an industry-wide issue and setting high standards within their organisation.

A new study from Realise and Cognassist aimed to identify learners at greater risk of dropping out based on cognitive assessment results and other individual factors.

Looking at historical data provided by Realise, Cognassist used internal statistical analysis to predict dropout rates and identify key factors that can affect dropout. Cognassist used regression-based statistical analysis on a sample of 223 Realise learners.

Although this work is preliminary, Cognassist was able to predict which learners were at a higher risk of dropping out with good accuracy. Learning difficulties, identified in a cognitive assessment, were significant risk factors for learners dropping out. The research also found that this risk may be increased by other factors relating to the circumstances of the learner and their area of study.

The Senior Research Scientist at Cognassist, William Brown, who ran the study commented on the findings, “We know learner retention is a challenge for the sector. This work is important progress towards understanding both who is at higher risk of dropping out and why they are at higher risk. This is important for our end goal of proactively identifying and supporting learners to prevent apprenticeship dropout. We are hugely grateful for the data collaboration from Realise, whose quality data collection and storage practices make this kind of analysis possible. We are looking to expand this research.”

Gregg Scott, Managing Director at Realise spoke about the impact for their organisation, “The new research developed by Cognassist is incredibly exciting. As a leading training provider, we have a responsibility to ensure that everyone is supported and given the right opportunities to progress.

“The data we have collected will allow our teams to have more meaningful conversations with apprentices and ultimately, do what’s right for our learners.

We’re proud to be the first organisation to take part and hope to see other providers follow suit, so that we can drive forward inclusivity and proactive improvement as a sector.”

Future applications of this analysis and a data science algorithm would give leading education providers like Realise a groundbreaking approach to identifying and supporting their learners. Knowing which learners are more likely to drop out from day one would have a significant impact on increasing retention and attainment rates. It would give providers the advantage when it comes to identifying learners who might otherwise fail to succeed.

Cognassist and Realise are hugely excited about this research and the role it could play in the ongoing commitment Realise has with its learners.