Graduate earnings data on Unistats from the Longitudinal Education Outcomes (LEO) data

Experimental statistics: 
These statistics are published as experimental statistics. This means they are still being evaluated and are not yet fully developed.

This page provides more information about the graduate income data we have published on the Unistats website.

This data is drawn from the Longitudinal Education Outcomes (LEO) data and is published as experimental statistics.

Experimental statistics are new official statistics which are undergoing evaluation.

They are published in order to involve users and stakeholders in the assessment of their suitability.

We would like to hear what people think of the graduate income data on Unistats. In particular, we would like to know how useful prospective students find it, and what they think about how it is presented on the site.

We will consider all comments as part of our longer-term plans for using the data.

To give feedback, email [email protected].

The Longitudinal Education Outcomes (LEO) data refers to a dataset of education records joined to tax and benefits data. This shows whether graduates were employed and how much they were paid.

It also shows if they studied on another course after their first.

See more about the data

LEO data for specific subjects offered at each provider is produced by the OfS and published by the Higher Education Statistics Agency (HESA) in the Unistats open dataset. This data is also published on the Discover Uni website. The latest data was released in September 2019.

The data shows the typical annual earnings range three years after the student has graduated. It is based on UK students who qualified from an undergraduate course in the academic year 2011-12 or 2012-13 and were in sustained paid employment in the 2015-16 or 2016-17 tax years.

For instance, if someone qualified in June 2012 the data would show earnings from April 2015 to March 2016. If they were also studying during that tax year they are excluded because employment was probably not their main activity. If people worked less than a full year we estimate what their earnings would have been if they had worked all year.

LEO is based on the Higher Education Statistics Agency’s subject groupings. So it shows the most detailed level of the subject grouping available. Sometimes there is only one course in that subject grouping at a provider, but sometimes there is more than one course. We are not trying to show data at course level because the number of people at course level is usually too small for reliable analysis.

In each subject, the data looks at the range of earnings for all students. We then show the median earnings and an interquartile range. The median is the middle value when all earnings are ordered numerically. The interquartile range shows the middle 50 per cent of values (25 per cent either side of the median). All figures are rounded to the nearest £500.

The LEO data provides earnings information for most graduates for many years after they finished their course.

LEO data on Unistats replaces the Longitudinal Destinations of Leavers from Higher Education survey. The latter only used a relatively small sample of graduates. This meant that data could not be shown for individual providers.

But LEO has information on everyone in paid work in the UK. So we can show data for nearly all subjects at each English provider.

The survey also relies on graduates giving accurate and unbiased answers about their earnings and employment. These factors do not affect LEO.

The LEO data is based on HMRC’s tax records about the amount individuals earned in the UK during the year. The data does not include whether a graduate is working full- or part-time, or the details of job they are doing, as these are not required by the HMRC.

This means certain types of workers may have a lower income recorded in LEO. The tax data for part-time workers, for example, does not indicate that they are working part-time.

This means that if many people from a particular course worked part-time, the earnings may look much lower than for a course where everyone worked full-time.

The average earnings illustrate typical earnings for graduates in that subject.

As with all income data, many factors may affect what someone earns besides their undergraduate course: their ability, background, chosen career, and where they are employed, among other things.

It is important to remember that the data from small numbers of people are less likely to represent typical earnings than data from larger numbers. Also, past trends in earning potential are not always a reliable guide to the future.

We have noted that earnings data displayed on Unistats may relate to graduates from more than one course. For example, where there are fewer than 15 graduates who studied Chemical Engineering at a provider, their earnings data may be combined with graduates from other engineering courses. If these different courses have very different earnings, this will not be visible to the user.

The following groups are not included:

  • People with earnings from self-employment are not included. We are not able to calculate what their earnings would be if they worked for the full year, or remove those who only worked for a small part of the year. We will continue to explore possible ways of presenting data for those in self-employment.
  • People who work abroad are not included because they are not in the UK tax data.
  • International students are not included, because a high proportion of them will not stay in the UK after graduating.
  • Those undertaking voluntary work or unpaid internships will not be included because they are not earning.
  • We only show LEO data for higher education providers in England, excluding those who were not submitting student data before 2012-13. This is due to different legislative and policy positions in the other countries of the UK.

Next steps

This data will be updated to use more recent tax data as that becomes available.

Describe your experience of using this website

Improve experience feedback
* *

Thank you for your feedback