Last week’s release of the Destination of Leavers from Higher Education (DLHE) UK Performance Indicators allows us a number of opportunities. It gives us a first glimpse of the trends that could underpin future TEF scores, and allows us to examine (in a robust, benchmarked, manner) trends linked to institutional performance.
What we are looking at here is the difference in institutional performance in graduate employment or further study, against a benchmark. This is complicated by the fact that the benchmark itself moves, based on the changing makeup of the student body at any given institution between cohorts.
Institutions with a high positive difference between their performance and the benchmark (those doing better than expected) are towards the top, whereas those performing below the benchmark are towards the bottom. You can choose to look at full- or part-time performance. Graphs on the other tabs show the changes in benchmark and indicator over the past year.
Moreso even than usual, there is a lot to be gained from playing with the filters in this graph. Whereas a first glance could cause mellifluous delight at a number of music colleges and abject despair at the University of Bolton, there is a lot of signal within the overall noise.
Points of interest
Consider, first of all, the Russell Group. Though in general benchmarks have remained near-enough static, performance against the benchmark for full-time graduates has worsened dramatically at King’s College and Imperial College. But Nottingham and Liverpool have seen a notable improvement.
There’s a fascinatingly mixed picture for small and specialist institutions too. Both the Royal College of Music and the Royal Northern College of Music substantially outperform their benchmark (RCM boast an incredible 100% of full-time graduates in employment or further study for the 2016-17 cohort). And there is a stunning improvement in performance at the Conservatoire for Dance and Drama. But at Rose Bruford and the Liverpool Institute of the Performing Arts there has been a substantial slip against the benchmark in the same time period.
Points of significance
You’ll probably remember significance flags from the TEF – and the UKPI datasets are where these flags come from. However a single flag (+ or -) is equivalent to a double flag (++ or –) in TEF. To be really robust, we should only really be looking at flagged performances as being a significant change in institutional impact on graduate outcomes.
I’ve included an option to filter by positive or negative flags, and we immediately see a much less crowded graph if we look at either. It’s still bad news for the University of Bolton, but great news for De Montfort, Coventry, West London and (especially) Guildhall. This latter group of institutions can confidently argue that they are significantly improving the employability of their graduates.
About the data
No matter how much a man loves HESA data, there’s something about a UK Performance Indicators release that quickens the pulse that little bit more. And it is all in the masterly way that benchmarks are used to look at the actual performance against the expected performance for each institution. The most recent iteration examines the DLHE data.
These benchmarks incorporate an understanding of the context within which graduates proceed to employment and further study. On an institutional basis, the benchmarks for DLHE-derived data incorporate the results of controlling for subject group, entry qualifications, and ethnicity.
Not all institutions will have data points for both years – this is linked to the way that HESA manage the release of data relating to small groups of individuals.
DLHE itself is a survey – with return rates generally encompassing a little over 80% of graduates from a given cohort, six months after graduation. It’s worth, at this point, reminding ourselves of how DLHE relates to other major graduate employment related releases.
|Name||Sample||Timing||Best used for...||Also used for...|
|Destinations of Leavers from Higher Education||80% of graduates from a given cohort, self-selecting survey respondents.||Six months after graduation||Showing variations in graduate experience shortly after the completion of study.||Rating institutions within TEF, Unistats.|
|Longitudinal Educational Outcomes||All graduates from a given cohort with PAYE tax records||1, 3, 5, and 10 years from graduation||Estimating the median historic graduate earnings premium for given institutions and subjects||Sam Gyimah’s app competition, Unistats.|
|Graduate Labour Market Survey||Weighted sample of all graduates between 21-30.||Variable||Understanding the current state of the graduate job market.||Actually, nothing that bad. Yet.|
Our usual caveats – of course – apply: near-immediate graduate destinations are a single datapoint and (even benchmarked) do not underpin any responsible opinion on the quality or otherwise of institutional provision. No one we respect would ever claim that such data represents a true or useful picture of the sector. But as government tends to like to use this sort of information, we thought you’d like to see it.