Data intelligence

The Open University under-reported the number of students in receipt of DSA in England and Wales in 2014/15. This is due to changes in the administration arrangements for DSA awards in 2014/15 which resulted in information not being available for all students in receipt of an award. They estimate that the return undercounts students in receipt of DSA in England and Wales by c.370 awards and 34 awards, respectively (applicable to table T7).

In 2011/12, the questions on the Destinations of Leavers from HE (DLHE) questionnaire were modified and as a result, the concept of activity was redefined for use in HESA publications. The DLHE population was also extended to include leavers with additional qualifications. Both of these factors had an impact on the employment indicators making data prior to 2011/12 not comparable with data from 2011/12 onwards; see the changes document for more details (applicable to table E1).

Rounding strategy

Due to the provisions of the Data Protection Act 1998 HESA implements a strategy in published and released tabulations designed to prevent the disclosure of personal information about any individual. This strategy involves rounding all numbers to the nearest multiple of 5 and suppressing percentages and averages based on small populations.

HESA Services Standard Rounding Methodology:

1. All numbers are rounded to the nearest multiple of 5
2. Any number lower than 2.5 is rounded to 0
3. Halves are always rounded upwards (e.g. 2.5 is rounded to 5)
4. Percentages based on fewer than 22.5 individuals are suppressed
5. Averages based on 7 or fewer individuals are suppressed
6. The above requirements apply to headcounts, FPE and FTE data
7. Financial data is not rounded

Total figures are also subject to this rounding methodology after calculation; so the sum of numbers in each row or column may not match the total shown. Suppressed values are normally represented as '..' in published tables (prior to 2006/07 these were represented as a blank value within the UK Performance Indicators.

Note: The suppression level within the UK Performance Indicators changed from 20 to 22.5 in 2011/12.

From 2011/12, data within the employment indicators have been suppressed and represented as a blank cell where the response rate for a HE provider is less than 85 per cent of the target response rate (68.0% for table E1a, 59.5% for tables E1b-E1d). Data for these HE providers have been excluded from all indicator and benchmark calculations. In 2010/11, data for HE providers in table E1 have been suppressed where response rates to the destinations of leavers survey was considered to be sufficiently low that their data may not be comparable with other HE providers.

From 2012/13, where data for a HE provider has been suppressed, they have been removed from all totals and benchmark calculations, see PITG paper 13/01. Also from 2012/13, in the event that a HE provider's data contains more than 50 per cent unknown values within the benchmarking factors, the benchmark has been suppressed and represented as a blank cell. This data has, however, been retained and included in totals and benchmark calculations for the rest of the sector. See PITG paper 13/03 and PITG minutes, 28 February 2013.

The low participation measure used in Tables T1, T2 and T3b is based on a UK-wide classification of areas into participation bands. From 2011/12 it uses the POLAR3 methodology which is based on a similar method to POLAR2 but uses more up to date information. As these two methods are not strictly comparable, the 2009/10 to 2013/14 indicators have been produced using both POLAR2 and POLAR3 to provide time series information. The POLAR2 data was no longer published as part of the UK Performance Indicators from 2014/15. The POLAR3 data was updated in July 2015 and this updated version was used in the creation of the Performance Indicators from 2014/15. The updated version includes new postcodes which have been added in the UK in recent times. More information on the POLAR methodologies can be found in the definitions document.

The relatively high (in UK terms) participation rate in Scotland coupled with the very high proportion of HE that occurs in FE colleges means that the figures for Scottish HE providers could, when viewed in isolation, misrepresent their contribution to widening participation. Therefore, low participation data has not been produced for HE providers in Scotland from 2007/08 (applicable to main tables T1, T2, T3b and supplementary tables SP3, SP6, SP7) and only for supplementary table SP1, for Scottish-domiciled students.

The Scottish Funding Council (SFC) produce their own indicators relating to young full-time Scottish domiciled undergraduate entrants, showing proportions from the most deprived datazones and also from social classes NS-SEC 4-7. These measures are produced and published independently of ‘Performance Indicators in HE in the UK’. The Performance Indicators Steering Group bears no responsibility for the SFC measures but has recognised that they may provide some contextual information for interested readers. The measures are available from the SFC website.

Prior to 2005/06 (T1a only) and 2006/07 (T1b-T1c), the low participation indicator was based on Super Profiles low participation, more details on the change in methodology between Super Profiles and POLAR can be found in changes to the postcode indicator.

A notable change from the 2014/15 UK Performance Indicators is the allocation of The Open University (OU) students to England, Wales, Scotland and Northern Ireland. Previously all OU enrolments and qualifications were counted within England, where the OU has its administrative centre. From this release onwards enrolments and qualifications registered at one of the OU’s national centres in Wales, Scotland and Northern Ireland will contribute to the totals of those countries where statistics are shown by country of HE provider.

For tables which include part-time students, a total row has been provided for the OU which has been produced consistently with previous published versions. Total rows for the OU will be produced for 3 years to enable the continuation of time series. From the 2019 publication, the OU will only be shown under the national centres.

For the 2001 census, a new classification, National Statistics - Socio-Economic Classification (NS-SEC), was developed to replace Social Class. It took into account new work patterns in the UK and the changes in education levels required for, and the status of, large numbers of occupations. This new classification was used for the social class Performance Indicator from 2002/03 and called the SEC indicator. More details on the differences between SEC and social class can be found in changes between 2001/02 and 2002/03.

For the 2008/09 academic year, UCAS changed the question that informs NS-SEC for the majority of applicants. The question reverted back to the original wording for 2009/10 applicants.

For applicants up to and including the 2007/08 academic year and for the 2009/10 academic year, UCAS asked:

"If you are under 21, please give the occupation of your parent, step-parent or guardian who earns the most. If he or she is retired or unemployed, give their most recent occupation. If you are 21 or over, please give your own occupation."

For applicants for 2008/09 entry, the question changed to:

"If you are in full-time education, please state the occupation of the highest-earning family member of the household in which you live. If he or she is retired or unemployed, give their most recent occupation. If you are not in full-time education, please state just your own occupation."

The change in question between 2007/08 and 2008/09 had an impact on the NS-SEC indicators, causing the proportion of students classified as ‘unknown' and those classified as falling into NS-SEC groups 4 to 7 to rise. Given these differences and the lack of any significant external changes to the system, it is safe to conclude that the change in question means that the NS-SEC data for 2008/09 is not comparable with that published previously and as a result, the 2008/09 NS-SEC data was published separately in tables T1ai, T1bi and T1ci and labelled as age-adjusted NS-SEC.

Although the question reverted back in 2009/10, there may still be a slight impact on the NS-SEC indicators resulting from applicants who applied using the 2008/09 form, but deferred entry until 2009/10. However, the 2009/10 data will be more comparable with that published up until 2007/08 than with the 2008/09 data. Therefore, NS-SEC time series data published within the summary excludes data for 2008/09.

The NS-SEC indicators will not be published in 2017 and subsequent publications of the UK Performance Indicators. For more details, please refer to the section on widening participation of under-represented groups.

Data for the University of Brighton have been suppressed in table T3d in the 2014/15 publication due to the incorrect coding of qualification obtained for a group of students (applicable to table T3d).

Data for the University of Plymouth was suppressed in table T3d in the 2013/14 publication due to the transfer of 50 students to City of Bristol College who are not included in the transfer figures as City of Bristol College are not within the UK Performance Indicators. As a result, their data has been excluded from table T4b in the 2014/15 publication (applicable to table T4b).

Data for the University of Brighton were suppressed in table T3d in the 2013/14 publication due to the incorrect coding of a group of students on courses delivered at a partner college as dormant, rather than having successfully completed. As a result, their data has been excluded from table T4b in the 2014/15 publication (applicable to table T4b).

Due to some students on first degree courses in 2013/14 being incorrectly recorded in 2014/15, the proportion with ‘not known’ outcome for the University of Stirling is higher than it should be. This in turn means that the other proportions will be lower than they should be (applicable to table T5).

From 2014/15, for new entrants, the last provider attended field (PREVINST) must contain a valid UKPRN or a valid generic code (see the HESA coding manual for details), rather than the historic UCAS, department and HESA school codes. These have been mapped to school type and grouped up to form the state school marker where appropriate. Where an unknown or invalid PREVINST code has been supplied, these students have been excluded from the calculations of the state school indicator. Due to the changes in the coding frame there is likely to an impact on the quality of the school type data.