Associations between characteristics of students

About the ABCS data

There are seven ABCS measures, which encompass all stages of the student lifecycle.

For each of these measures, we have used statistical modelling to calculate predicted outcome rates for each of the student groups in our modelling data. We have then used these rates to separate groups of students holding a set of characteristics into quintiles.

Those with the lowest modelled rates will be in the lowest quintile and those with the highest will be in the highest quintile.

If you require any additional information or have any comments or suggestions please contact us.


ABCS are measures that can be used, alongside others, to identify and target groups of students who are underrepresented in their access to higher education or who experience lower continuation, completion or progression rates than other student groups.

No. ABCS measure the likely outcomes for groups of students based on a set of characteristics.

We do not recommend doing this. Each measure is tailored according to the characteristics associated with positive outcomes at the stage of the student lifecycle to which it relates. Also, many of the characteristics used to define ABCS groups are either unavailable or less meaningful for postgraduate students.

This means that each measure should normally only be applied to their relevant stage of the student lifecycle, based on the populations used to construct the measure.

For example, ABCS part-time continuation quintiles should be applied to undergraduate students, in the entrant year of their course, studying part-time.

We have not included apprenticeship students when constructing ABCS measures. This is because there are not enough of these students to create a standalone measure for apprenticeships, and we did not want them to influence the measure for either full-time or part-time students.

But it is possible to assign ABCS quintiles to apprenticeship students using the full-time measures. Overall, we consider this to be the best match for apprenticeship students, considering the outcomes and census dates of these students.

We have explained this in more detail in the addendum we have published to our analysis of responses to our consultation on constructing student outcome and experience indicators for use in OfS regulation.

If you are assigning quintiles to apprenticeship students, the population should be matched to the stage of the lifecycle for that measure. For example, ABCS full-time continuation quintiles can be applied to apprenticeship students in the entrant year of their course.

ABCS measures could be used in different ways:

  • Providers may wish to apply ABCS quintiles to their own students. This will help them to understand their students’ outcomes and experiences. It will also identify any groups who have, historically, been less likely to achieve positive outcomes.
  • Outreach practitioners can explore which of the ABCS access quintiles prospective students are likely to fall into. This means they can be used to identify and support groups less likely to access higher education.

Please note, the measures can be used for other purposes too.

The main reason is data availability. The data sources for the measures vary, and so some characteristics are available for some measures and not for others.

This is also why the access measure reports ethnicity in a different way – because the data source collects it in a different way.

Availability also accounts for differences between the full-time and part-time versions of a measure.

The number of part-time students is much lower than the number of full-time students. So where there are challenges with data availability, these are amplified in the part-time population.

To obtain the quintile for a single individual, choose attributes for each characteristic you know about the individual.

If you don’t know the attribute for some characteristics, then leave the dropdown as ‘—any [characteristic]—'. The derived quintile will reflect the one calculated by the weighted average approach as described in the methodology document.  

Please note that the more characteristics that have unknown attribute values, the less confidence can be placed on the derived quintile. So they should be used with caution in these cases.

In some cases where one or more characteristics are set to ‘—any [characteristics]—’, the dashboard will not display any results. This indicates that it is not possible to calculate a quintile distribution or a derived quintile for this combination using the weighted average approach.

This is because there are no students in the original modelling data in any of the combinations represented by the attributes which comprise the missing characteristic(s) in question. Therefore there is no quintile distribution from which to derive a quintile.

Take this example: say there were no students in the original modelling data with FSM eligibility either ‘Has ever been eligible for FSM’ or ‘Never been eligible for FSM’ (assuming all the other attribute values are known and are the same for these two student groups). This means there will be no students in either of these groups and consequently, even though we can predict a quintile for such groups, when FSM eligibility is set to ’—Any FSM --’, the weighted average for these two student groups will be zero too.

It is not practical to find the quintiles for large numbers of individuals using the dashboards, as this would require a lot of manual interaction.

To obtain quintiles for multiple individuals, you would need to use the data downloads and look up their quintiles based on the individuals’ characteristics. You will need to ensure that the characteristics match in terms of the attributes.

If any attribute groups have been combined in the modelling data (for example, small groups such as the Gypsy or Traveller ethnic group are often combined with a larger ethnic group for ABCS modelling), then any individuals with this attribute should have their attribute value changed to match the larger group. This will allow the individual to be matched with the ABCS data.

Information about which attributes have been combined in this way can be found in the report associated with each ABCS measure.

Although each year we receive an additional cohort of student and pupil data, we do not plan to update the ABCS measures annually.

Instead, we will use some broad principles to work out if we need to update a measure, as outlined in the methodology document.

Last updated 30 September 2022
30 September 2022
Information updated in view of general updates and additions to the measures.
13 October 2021
Updated for latest ABCS report
26 November 2020
Information updated to reflect the latest report and updates to the data dashboards.

Describe your experience of using this website

Improve experience feedback
* *

Thank you for your feedback