TNE taxonomies

11 December 2015
Rod Bristow

Dan Cook

Head of Data Policy and Development
Higher Education Statistics Agency

​​The Higher Education Statistics Agency (HESA) collects data about the UK higher education sector’s transnational education (TNE) through the aggregate offshore record (AOR) in institution’s annual data returns. However, commentators often suggest that the AOR does not capture TNE operating models in ways that fit the evolution of offshore programme delivery. Dan Cook takes a critical look at the opportunities for change, and asks how the agency might improve data on one of the most central elements of the UK’s TNE data collection: its categorisation of the ‘type’ of TNE being delivered. 

In a recent HEGlobal commentary, Peter Dickinson points out that the HESA Aggregate Offshore record (AOR), while being one of the world’s foremost TNE data collections, still offers ‘limited information about the types and modes of delivery’. When the AOR was first designed, experts had not agreed a standard terminology for TNE: words like ‘franchising’, ‘articulation’ and ‘branch campus’ were used to indicate different ‘types’ of TNE but precise definitions did not exist. Inevitably, HESA’s TYPE coding frame attempted to cover an abundance of activity with very basic identifiers.

The situation has not advanced far today. Iterations of research have created new taxonomies for TNE ‘types’, but precise and broadly-agreed definitions still elude us. Although there has been some progress, the limitations of the HESA taxonomy have been reached:

“The categorisation of transnational education used in the Aggregate Offshore Record is not useful for understanding patterns of transnational education provision.”

BIS (2014), The Value of TNE to the UK

Unfortunately, there is no ‘off-the-shelf’ solution, new thinking is needed:

“Institutions would appear to be supportive of the development of a more systematic nomenclature and categorisation of the types and modes of transnational education to enhance sector-level data collection and improve comparability across the sector. Work is therefore required to improve the TYPE field, to reduce burden and increase the relevance of data. This demands the engagement of those who produce and use these data, as well as HESA.”

BIS (2014), The Value of TNE to the UK


What are the options? 

Several possible taxonomies for TNE exist: examples include ‘Trans-national Education and Higher Education Institutions: Exploring Patterns of HE Institutional Activity​’, ‘The Value of TNE to the UK’ ; and the Quality Assurance Agency’s (QAA)  Chapter B10 in the Quality Code for Higher Education. However, when these controlled vocabularies for TNE are set side-by-side, superficial similarities give way to inconsistencies in detail. For instance, the term ‘franchise’ is defined subtly differently by each of these schemes (see table below). Similarly, terminology used for branch campuses neither agrees between the three schemes, nor with the C-BERT branch campus listing, often recognised as the most significant global study of branch campuses.

Because of these differences, a simplified taxonomy for ‘types’ of TNE cannot be created by combining the different terminologies. Rather than boiling down to a revised, simple ‘meta-standard’, an extensive and confusing list would be required to capture the variations in terminology, a point not lost on HEGlobal, whose glossary of terms is illustrative and ‘not exhaustive.’

HE data management professionals tell HESA that it is difficult to apply the current HESA taxonomy to their provision, and in addition, many HE providers use their own in-house terminology, adding further complexity. Since substantial background knowledge is already required to inform judgements about categorisation, expanding our list of ‘types’ is unlikely to improve data quality. Quite the opposite. We need to go back to fundamentals when describing ‘types’ of TNE, and work out exactly what it is we are trying to measure.


Underlying taxonomic problem

As the range, scale and diversity of TNE has grown, we have reached the limits of the descriptive power of current taxonomies. The principal reason for this confusion is that the concept of TNE ‘type’ conflates many separate issues. These include:

• Student registration status

• Campus/location of instruction

• Awarding body (possibly including issues of 3rd party professional accreditation)

• Curriculum design responsibilities

• Teaching responsibilities of the reporting institution

• Teaching responsibilities of the collaborating partner

These six dimensions of TNE cannot be measured on a single scale, as they are about quite different ideas. If we conflate them, we require a huge number of categories to cope with the variations in meaning; if we instead take a reductionist approach and limit categories to a manageable number, we create overlaps and/or gaps where our meaning is not properly captured. By combining multiple discrete factors simultaneously in a single measure, no TNE taxonomy can represent the underlying activity adequately.


Rethinking TNE taxonomy

We could make progress by measuring separately the various elements of which TNE ‘types’ are comprised. Through asking a greater number of simpler questions, the production of the data could also be less burdensome, and future-proofed. We could replace laborious self-categorisation against contested terms, with straightforward questions. Some initial ideas – by no means set – include those in Table 1 below.

Table 1: Rethinking TNE taxonomy​

Turning new base data back into TNE ‘types’

An approach like this shouldn’t lose the consistency and comparability of our present taxonomies. Existing TNE categories could be ‘reverse-engineered’ as analytical derivations as long as the necessary basic data are present. For instance, a ‘franchise’ could be identified (for each available definition discussed above) through valid answers to each new question, as set out in Table 2. In this table, the numbers indicate the valid responses required to identify a ‘franchise’ according to each of the three definitions. Numbers indicate the valid responses (as numbered above against items a) to f)) that could be used to construct each of the three definitions of ‘franchise’ from base data. This example demonstrates both current inconsistency in terminology, and the potential usefulness of the proposed scheme.

Table 2: Correlation of proposed approach with three definitions of 'franchise' 

Our team


Measuring transnational education: changes to the TNE data landscape

10 May 2021
Eduardo Ramos, Head of TNE and Griff Ryan, Transnational Education Projects Officer share insights on the new data on UUKi's Scale of TNE interactive webpage.

You can’t create virtual exchange overnight: what we have learnt from moving international experiences online

20 April 2021
Leo Smith, Head of Global Mobility at De Montfort University shares reflections from moving mobility experiences online during the pandemic.