05/04 - October

BACKGROUND TO THE REVIEW

The HESA Student Record was collected for the first time in 1994/95. It replaced a number of different individualised student records that served the various parts of the HE sector before the 1992 Further and Higher Education Acts. For a small number of institutions it was the first regular individualised student record to be collected.

The original record specification was the result of much development and consultation during 1993 and 1994. A single record structure could not be agreed upon and the record specification ended up containing both Combined and Student/Module structures with the freedom for institutions to use either or both. From 1995/96 institutions in Wales were instructed by their funding council to only use the Student/Module structure.

The HESA Student Record covers all students registered at HE institutions, including those taught elsewhere under franchise and “Associate College” arrangements.

The record is designed to support specific requirements from different stakeholders in different parts of the UK and those with specific areas of interest such as heath and social care and teacher training. The HESA Student Record is collected once for use by many different stakeholders.

Data for FE students registered at English HE institutions is passed to the LSC (formerly FEFCE) and is used to support analysis and funding calculations. Many changes in the HESA Student Record have been the result of changes in data requirements for the FEFCE/LSC.

Changes to the specification were made for 1995/96 and 1996/97 and the first major review of the record took place in 1997 for implementation for the reporting period 1998/99. The record specification was relatively stable in the years immediately following the 1998/99 review, although by 2001/02 changes were being implemented on an annual basis – a pattern that has continued to the present day.

One consistency in the record since its inception has been its definition on the basis of registrations; a person registered on two programmes has two records in a collection. However, the assumption that has always underpinned standard analyses of the record has been that the difference between counts of registrations and counts of people is not statistically significant.

The concept of registration was further reinforced in 1998/99 with the introduction of the student instance (HIN). This mechanism facilitates the year-on-year linking of registration records and now forms a key element in the data quality assurance processes undertaken by HESA. The implementation of the HIN was arguably the most significant conceptual development in the history of the record.

The record specification has grown significantly over the years. New fields have been added and many existing fields have seen their code lists change over time. The table below summarises the growth of the record specification.

HESA Student Record Evolution

Patterns of Data Collection

The HESA Student Record collection has always been undertaken as a single, complete collection. Every relevant student has been returned in every collection, irrespective of whether they have been returned previously or not, and the collections have taken place at specific periods, which are not related to when students actually register and graduate.

The Student Record was collected twice a year from its inception until 2001-02. The end of the December collection was agreed in the Summer of 2002 and the reporting period 2002/03 was the first to run with a single end-of-year collection.

The end-of-year collection forms the basis of the subsequent destinations return (formerly FDS, now DLHE). In the first year of collection of the FDS return, it became apparent that some HEIs had under-reported their “qualifications awarded” information leading to a lack of records to be included in the destinations survey. For the end-of-year 1995/96 collection a new mechanism for the later return of results information was implemented. This involved the submission of the main student data to the normal mid September timetable followed by the submission of the results/leaving fields by mid November. These extra fields were added to the existing records and the final version of the database was delivered to a later timetable than in the previous collection.

The Late Return of Results (LRR) process caused a number of difficulties: the algorithm for matching LRR fields to the existing records relied on some fuzzy logic and often failed in cases where a person was studying on more than one programme. (The implementation of the HIN in 1998/99 alleviated this problem to an extent, although the LRR matching algorithm was never without an element of fuzzy logic.) The LRR process also created the possibility of making invalid records that had previously passed the Agency’s validation checks. This resulted in a “double-pass” process being developed whereby the LRR submission was checked (validated), slotted into the existing data and then the entire return was extracted from the database and passed back through the validation processes.

Over time the number of institutions that used the LRR mechanism steadily fell and the entire LRR process became less and less ‘late’. In the Summer of 2000 HEIs were consulted about the removal of the LRR option (circular 00/02) and with no objection from the sector, the LRR was removed from the record specification for 2000/01.

Data Quality

The implementation of the HESA Student Record in 1994/95 was the single most significant upheaval in HE student data collection in the UK and the fact that the sector achieved a full response is a major tribute to everybody involved in the work at that time.

Data quality assurance in that first year of collection was focused almost exclusively on a suite of simple validation checks. These checks ensured that fields contained valid entries and that the relationship between certain related fields was consistent in each record. These checks worked well and delivered all that was asked of them, though it quickly became apparent that many errors could not be trapped by this type of systematic check.

The 1995/96 end-of-year collection saw the introduction of additional quality assurance processes including validation checks between records (e.g. searching for duplicate records), validation checks between files (e.g. ensuring that the module codes used on the student record are consistent with those used on the module record) and a full range of soft checks based on a pack of check documentation that highlight key statistics, current priorities and information derived from the record as well as year-on-year comparisons.

Data quality was, and continues to be, the highest priority of HESA and of every Statutory Customer to whom the Agency delivers data.