Last year, we teamed up with DataKind UK to spend a whole weekend delving into the mountain of high-quality data produced by our learning platform. We explored a series of questions relating to pupil engagement on our platform, with particular focus on the differences between those entitled to Pupil Premium funding and their peers. DataKind organised a team of volunteer data scientists who used a wide range of analytical methods to explore these questions, and brought their findings together in a report. Here we take a look back at some of the key takeaways…
Backing up what we already know
The findings reinforced what we already know about parental engagement and about the challenges of supporting those families who face the greatest barriers. The data confirms that past engagement is a strong predictor of future engagement, and those entitled to Pupil Premium support are on the median less likely to leave feedback for an activity. It is however important to note that for some activities there is no completion gap.
This suggests that we can work to narrow the overall gap by continuing to design activities that are fully accessible and inclusive for children eligible for the Pupil Premium.
The researchers also found that the order of the activities makes a difference in getting pupils to engage with the activity. Activities get less engagement the further to the right (on desktop) or lower down (on mobile) they appear on screen, which is likely a result of both the layout and the activity design. This is something we will need to consider closely when designing activities and the user experience on our platform.
New things we learned
The importance of lower targets
One of the key new findings that came from the data dive relates to the reading targets set by schools. Previously the standard reading target was five days a week, but the data shows that schools with a target of three days have a higher proportion of pupils reading at least once a week.
The data also suggests that having a lower target narrows the gap between children entitled to Pupil Premium funding and their peers.
As a result of these findings, we now recommend that schools consider choosing a three-day reading target (instead of a five-day default) and have made changes to the platform to nudge teachers towards a lower target.
Fridays are better than Mondays
The data suggests that the start day of the reading log matters. It appears that more Pupil Premium children meet their reading target when the start day is Friday, rather than Monday.
Because of this, we now encourage teachers to think about which start day will work better for their families and then make an active choice, rather than just picking Monday as default.
The impact of reading on Maths & English
Having a reading log, and more importantly, using it, correlates with a higher completion rate of our Maths and English topics.
We know that past completion rate is a strong predictor of future engagement, and it seems likely that the reading log may break down some of the barriers to getting started with other learning activities. This is likely due to the relative simplicity of logging reading, as well as a greater degree of parental buy-in for reading at home, compared with other home learning activities.
It appears that the reading log is not only a useful tool in its own right, but also a key lever to help drive engagement with other home learning. Thanks to this new understanding, schools will be encouraged to get families started with logging their reading, in order to begin building learning habits and reduce the barriers to completing other activities.
Food for thought - future research
The volunteers identified several areas for further research with some intriguing initial findings. For example, pupil age appears to influence engagement, with older and younger pupils responding differently. Teacher actions also play a role: their messages improve completion rates, particularly for older pupils who can read them independently. Future studies could examine how teachers use messages and whether targeting them at pupils or parents has differing effects.
The type of resources also matters – collaborative activities seem to encourage higher completion rates than parent-led ones. More detailed analysis of activity tags could offer additional clarity. Finally, an exploratory analysis of pupil and parent comments using natural language processing highlighted differences in sentiment, suggesting opportunities for deeper investigation. Volunteers also demonstrated how an interactive dashboard could support analysis and prioritisation.
Thank you, DataKind!
A huge thank you to DataKind for all their hard work producing the data dive, and for bringing together all the analysis into a single wrap up report. Thank you also to all the data scientists who volunteered their time to work with us. Your findings are already helping us to improve our platform and to better help the children and families who face the greatest barriers.
