While many of the data structures that operate across higher education are the preserve of the hardcore data experts, subject coding – the Joint Academic Classification System (JACS) – seems to have a life and prominence that extends beyond the usual sector analysts, wonks and nerds. Yet all of that is about to change, as JACS makes way for a new acronym: HECoS.
A brief history of JACS
JACS was created in the mid-90s by HESA and UCAS, who had decided to merge two existing subject coding systems – the Standard Classification of Academic Subjects (SCAS) and HESACode – to create a system that both organisations could use. The name of the new system originally came from the names of the four members of the joint HESA and UCAS project team though a more appropriate title to fit the acronym was found before the system was launched.
JACS is a hierarchical system with codes made up of a letter followed by three numbers. The hierarchical logic was built into the codes themselves – so for example, if the F group is Physical Sciences, F100 is Chemistry, F140 is Environmental Chemistry and F141 is Marine Chemistry. The codes are familiar to thousands, largely due to UCAS (and its predecessor organisations) using subject codes as a basis for the course identifiers that applicants write on their application forms.
The need for a new coding system
The hierarchical nature of the JACS system, while seeming like a good idea at the time, has been one of its major weaknesses. It restricted the number of major groups to 21 (vowels were avoided in the first version) as well as the number of subjects that could exist below each branch of the hierarchy. This resulted in many complex subject areas being shoe-horned into the structure, and some subjects being pushed down the hierarchy to give them enough space.
Although both HESA and UCAS used JACS, differences in the approach to implementation meant that subject data published by the two organisations was never quite comparable. Furthermore, a number of data collectors didn’t use JACS because it didn’t fit their requirements particularly well. JACS underwent two major revisions and by the time version three was launched, there was no space left in the hierarchy for any further development. When the HEDIIP programme started to tackle the problem of a standard data language that all data processors could use, addressing the issue of subject coding was a high priority.
Introducing HECoS
The replacement for JACS is called the HE Classification of Subjects (HECoS). It contains a similar number of codes to JACS and defines subjects at a similar level of granularity. However, unlike JACS, HECoS is a simple list of subjects with no inherent hierarchy. This removes the artificial restrictions that came with the structure of JACS and it also provides more scope for the system to evolve as new subjects emerge. HECoS codes are simply six numeric digits that have no inherent meaning.
HECoS is a key element of the broader work to develop a common data language that can be adopted by data collectors across the sector and thus standardise and rationalise the myriad of data flows that currently exist. HECoS is being implemented for the academic year 2019/2020 and this means the first HESA data collections that will run under the HESA Data Futures programme and the UCAS application cycle for 2019/2020. Preparations for that UCAS cycle are already well underway and institutions are (or, at least, should be) re-coding their courses in preparation for this. There is even a #HECoSCodeOfTheDay on Twitter.
Subject grouping and the Common Aggregation Hierarchy
Although HECoS is not hierarchical, we still need a way of grouping subjects together for analysis and publications – and if we are going to have a standard subject coding system, we ought to have a standard approach to grouping subjects that can be used for routine outputs. So, hot on the heels of HECoS, we have developed the Common Aggregation Hierarchy – or CAH (which I declare can be pronounced “car” – pronunciation matters here at Wonkhe). The CAH provides three levels of aggregation, allowing HECoS data to be presented in 23, 35 or 167 subject groups.
The creation of the CAH has been a major piece of work, involving all the major data collectors, HE providers and consultations with the team that is developing the subject-level TEF. We have also developed a mapping from JACS to CAH which can form the basis of time-series analyses that span the JACS-era and the HECoS-era.
I sometimes wonder if I’m more excited than I should be about a new subject coding system, but the implementation of HECoS is a huge milestone in the journey to create a more joined-up, more user-friendly HE data landscape.
Subject coding is not the sole preserve of the data-nerds. Anybody with an interest in HE needs to be able to distinguish Dentistry from Divinity and Digital Media, and if we can do that in a consistent and well-managed way then our data will be more useful and less burdensome.
Will this affect HEFCE’s proposal to use HESA codes in REF?