LEO and its limitations for learners

The Department for Education’s release of Longitudinal Educational Outcomes data in March – LEO – is an exciting development.

Linking HM Revenue and Customs (HMRC) data about graduates back to their subject at the university where they studied sounds, on the face of it, like a very promising dataset.

The analysis of the data, while experimental, could help us discover trends in salaries among graduates who studied different subjects, trends around the number of graduates who employed in related areas, and links between subject studied and the likelihood of employment or further study. For those developing higher education policy, and for providers, this is all useful information.

LEO is destined to be included on the UK’s Unistats website, which is designed to advise potential students about their undergraduate UK course choices – but is it helpful information for students?

Helpful for students?

Unistats – the official website for comparing UK undergraduate course data – was set up to ensure that prospective students had access to “robust, reliable and comparable information” to help them decide where to study. Statistics which are “experimental” are still undergoing evaluation and development and may not fall into that category.

The analysis also only covers graduates from undergraduate first degrees. Further education colleges outside England are excluded, as are graduates from postgraduate courses, thus limiting the information available to students.

The salary information tells you more about where the graduate has chosen to work, rather than their choice of university. Salaries – even for identical roles – are quite different in different areas of the country. Graduates choosing to work in London and in the south east of England are likely to gain a higher salary for like-for-like jobs than those in Cornwall, Cumbria or Ceredigion.

It’s about what you do, not what you studied

The data does not provide information on the sector in which graduates choose to work. Many graduates go on to work in professions that use the transferable skills developed during their time at university, and do not directly relate to their area of study. As such, interpretation of the data is very complex and has the potential to misdirect students.

For instance, graduates who choose to work in the public sector or for a charity may well be choosing a comparatively lower salary or a more lengthy career progression and longer salary trajectory than those who choose a different path.

No context

No contextual information, including benchmark data, is currently available to explain regional pay, cost-of-living or quality of life differences. Until appropriate benchmark data on pay levels has been developed, this data is incomplete.

As an extreme example, a certain graduate salary level in London and in the South East could get you one room in a shared house with additional costs for commuting, while, elsewhere, a lower salary might pay the rent for a whole house and a (healthier) walk to work.

Furthermore, there is little explanation of other factors that contribute to pay differences such as the gender pay gap, prior attainment, and student characteristics.

It doesn’t cover everyone…

As LEO is based on HMRC tax pay-as-you-earn data and the self-assessed earnings of self-employed graduates, some graduates will be excluded from the earnings figures. It also excludes earners below the tax threshold. Institutions with higher levels of creative arts courses are more likely to produce entrepreneurial graduates who work for themselves. These graduates are more likely to be below the tax threshold early on in their careers and excluded from the analysis. There is no account of the hours actually worked, and it excludes those working outside of the UK.

… Except it does cover everyone

Full-time and part-time graduates – whose median salaries and student profiles can be very different – are presented together. This could be misleading – many part-time students are already in employment when they study.

Out of date

The latest dataset is quite out of date: publishing the salaries of students who graduated more than 10 years ago might not give a helpful indication of salaries in the future, particularly for those just starting university.

Unintended consequences

If HE providers are to be judged – by potential students or policy-makers – on graduate salary levels, they are likely to want to encourage more of their graduates into higher-salaried sectors and areas to improve their performance in this “league table”. While good for salary levels, it could lead to prospective students believing that going to a certain university or studying a certain subject will get them a certain salary, and potentially play a role in producing disaffected graduates.

Action by universities to improve their position in LEO could lead to:

  • More graduates being encouraged to move to London and the south east of England for higher-paid employment, increasing the “overheating” already apparent in those areas.
  • Fewer graduates considering crucial posts in parts of the UK where salaries are lower.
  • Fewer graduates considering some essential lower-paid public sector services, meaning fewer choosing shortage NHS professions, teaching, and social work – professions that are already seeing major recruitment difficulties.
  • More graduates being encouraged into lower-paid salaried roles rather than pursuing creative talents and working for themselves.
  • More graduates being encouraged to work for large employers, where salary rates are benchmarked with others, rather than small and medium enterprises (SMEs), where substantial benefits can be gained by employing a single graduate.

The 2014 UK review of the provision of information about HE indicated that there were “limitations to the amount of information-processing that people can undertake when making a decision about whether to enter HE and which courses or university to attend”. More information doesn’t necessarily mean that people will be better informed to make better decisions.

In addition, it was suggested that “prospective students from lower socio-economic groups are more likely to choose universities based on their proximity to home, and are less likely to search for external information on universities”. This means that additional information for prospective students may not have the impact expected.

Make it official

LEO data can be useful, and it is always good that such data is in the public domain. However, before any attempt is made to promote LEO to students as a source of information for differentiating between HE providers, the analysis of the data needs to be official, not experimental.

The data was originally developed as an English analysis, and most of the analyses only include HE providers in England. It is important to ensure that this is a truly UK-wide analysis going forward.

The data also needs context, and should be accompanied by appropriate benchmarking against regional graduate salaries, and presented with additional contextual information.

The new Graduate Outcomes survey will provide the opportunity for some of that important contextual information to be captured.

3 responses to “LEO and its limitations for learners

  1. Sounds a lot like a fear of presenting students with an unbiased reflection of labour market outcomes from the degrees. Why is the HE sector so defensive about this?

  2. Salary is important information in any assessment of an outcome, but alone it doesn’t tell you much. To put salary in context, you need to know whether it was earned full- or part-time, where in the country, and in what sort of job role. Even data about those not in work needs careful handling – this is sometimes part of a meaningful life-choice based around caring for others, rather than an absence of work.

    When HESA publishes the first Graduate Outcomes data, we will have a data source that can help all stakeholders understand the patterns of outcomes from various kinds of study, by relating salary data (among other things) to the voices of half a million graduates annually.

    Mind you, even with a data source of that quality, I’d still be hesitant about predicting the future using it!

Leave a Reply