What did students get their parents for Christmas? Apparently, not Covid-19

In the charming way that we all became experts in epidemiology in 2020, everyone knew that there was loads of Covid-19 around universities late last year – so students coming home would likely infect everyone they came into contact with – driving the virus round the country.

It was for this reason that Michelle Donelan and Gavin Williamson invented the Student Travel Window, mass testing, and advice for students to avoid public transport and enclosed spaces where possible. The mass testing, arguably, sparked the English education sector’s increasingly magical belief in the efficacy of lateral flow testing to ensure setting specific safety.

Today saw the release of a more recent cut of TUNDRA (basically POLAR young participation data but removing some unusual school type outliers), covering participation between 2012 and 2016. It doesn’t exactly predict the proportion of HE students originally domiciled in each (MSOA) area, but it’s the best data we’ve got. And, as it happens, the PHE dashboard data offers us a week of Covid-19 case numbers and rates for MSOA starting on 25 December.

[Full screen]

There is no correlation between participation rates and case numbers (or case rates). R squared is about 0.06 – we’d only start thinking about a relationship between two variables at around 0.7. I’ve included an UTLA filter to let you narrow down to particular areas.

There was clearly a lot of Covid-19 in Essex at the time, but participation rates had nothing to do with it. Shenfiled and Hutton Mount (TUNDRA: 66.4 per cent) saw 1,448 cases per 100k that week – Bocking (TUNDRA: 22.5 per cent) saw 1,372.

Bonus plot

People like maps, so here’s a TUNDRA map (quintiles as colours) showing the location of HE campuses. If you stumble across a grey area (ha!) that’s because there weren’t enough people in that MSOA to safely calculate the rate.

[Full screen]