I predict that over the next couple of years a lot will be said, written, and argued about subject classification in higher education.
The sector will sidle up to the new HECoS subject classification system with some trepidation. Change is always tricky, but in this case the pros far outweigh the cons; Andy Youell has explained the need for a new system and introduced HECoS in his earlier article on Wonkhe. I was a member of the Cetis and Higher Education Data and Information Improvement Programme (HEDIIP) team that developed HECoS, and so offer a few thoughts on the practicalities of using it and the associated Common Aggregation Hierarchy (CAH).
HECoS is a similar size and similar granularity to JACS; in fact, it has just over 1,000 terms to JACS’ over 1,500. Some of the terms are the same as those in JACS and have the same definitions. Some JACS terms have been dropped, and new terms have been added, reflecting the development of course provision over the years.
But the most obvious difference from JACS is that HECoS is not a hierarchical classification system; it’s a flat list of subjects of study. Clearly the sector needs a mechanism for grouping together data about subjects of study. In JACS this is provided by its hierarchy, and this caused a tension between a node used to classify a course (“this is a history course”, “this is a medieval history course”) and a node used to group data by subject (“this is data about history courses, including both general or unspecified sub-divisions and specific sub-divisions, such as medieval history”).
In the new system, the grouping of data about subjects is not through HECoS, but through the Common Aggregation Hierarchy (CAH), so that decisions about how we group subjects together – typically for statistical analysis – is separate from how we classify records by subject. The CAH is called “common” because the design team, driven by extensive consultation throughout the sector, envisaged that most HECoS users would want a common framework for grouping subjects, rather than several bespoke ones. In practice, this means that those classifying with HECoS can focus much more on allocating the correct subject term, rather than on where the subject lies in a hierarchy. Discussions about subject groupings can take place in relation to the CAH instead of the classification system itself.
For example, you can classify your BA Honours in History as “history” in HECoS, rather than arguing about whether it’s history by period or history by topic, as under JACS, which lacks a general history term. If you need to group the data together, you can use the CAH “history” grouping to aggregate all the data on various historical periods and topics, including general history courses.
HECoS and the CAH may be useful in discussions about Longitudinal Education Outcomes (LEO) data in the future. HECoS has updated definitions of the meaning of subjects of study in many cases, and a completely new mapping to a new aggregation framework. Some of the difficulties in how we define “subject of study”, and how we might usefully aggregate subjects may be either resolved or at least become more transparent, so that issues can be addressed more clearly. Analysis of TEF or LEO is always limited by the accuracy of classification, which will be rather better using HECoS compared with JACS. Analysis also rests on the way in which subjects are grouped, clearly separated when grouping HECoS subjects into the CAH.
In principle, a subject term in HECoS has been included where its definition and scope is sufficient and comprehensive enough to allow relatively simple classification, where one subject term can be reliably distinguished from other subject terms. This makes practical classification with HECoS easier than with JACS, because ambiguities have been removed. For classification of courses offered by universities and colleges, we looked carefully at course provision and have included more terms to cover the whole breadth of a subject or subject area as described by the provider, for example: ‘history’ or ‘law’ or ‘humanities’. This principle should result in greater consistency of classification across the sector than was achieved with JACS.
Already, some are talking about HECoS codes and how to remember them. I would like to instigate a rule that whenever anyone mentions a HECoS *code* when classifying, they have to write out 100 times “I will not use HECoS codes when I classify things”. HECoS codes are unmemorable, and this is entirely deliberate. Using modern interfaces, we expect that classifiers will use the labels for the subjects, not the codes. Let’s face it, ‘history’ is more memorable than ‘100302’, and it’s not going to aggregate to ‘1003’ or anything like that.
As an aid to the re-classification effort, there are mappings from JACS3 to HECoS. However, a word of warning here: it’s important that users don’t just automatically assign mapped terms. This will set in stone any imprecise, shoe-horned (or wrong) coding from JACS, whereas HECoS may well have a correct term for you to assign. With better classification comes better data analysis!
All of the above is not to say that HECoS and the CAH are perfect. No doubt the extensive usage that both will be subjected to in the coming years will reveal room for improvement. We should all find that the governance structures to be put in place by HESA will give the sector new opportunities to shape the future development of these vital parts of our higher education information infrastructure.
So, as you start to address the move from JACS to HECoS, consider carefully the process of re-classification. It’s an opportunity to make sure your subject of study data is of high quality, so that internal analysis and planning is better, and external data returns provide a more accurate picture. Of course, you are not alone in this task. Cetis wonks, and others with experience in this field, can help advise and streamline your effort.