Skip to main content

Common Aggregation Hierarchy - about

The Common Aggregation Hierarchy (CAH) provides a standardised hierarchical aggregation of HECoS codes suitable for the majority of uses.

The CAH has been developed to provide standard groupings that can be applied to both HECoS and JACS subjects allowing for consistent analysis across both coding frames. These aggregations provide some continuity in subject analysis through the transition from JACS to the new scheme and to other schemes such as those used by QAA. This will improve consistency across the sector.

The CAH has been designed to act as a bridge between HECoS and JACS. These are, however, two distinct coding frames, with HECoS having been developed in part to allow for more robust coding to address inconsistencies in coding using the JACS framework, and to remove limitations imposed by the JACS structure. As such, analysis between HECoS and JACS using this Common Aggregation Hierarchy will only provide indicative comparisons for time series analysis.

On this page: SpecificationDevelopmentDownloadQuality

Specification of CAH

The CAH is a comprehensive aggregation of the entirety of HECoS at each of three hierarchical levels or tiers, where a “parent” group at a higher level of CAH always comprises the full set of HECoS codes represented by related “child” groups of CAH below it.

CAH groups are numbered with a series of two-digit integers representing a group of HECoS codes at a level, with the code at each level separated by a dash. Each unique CAH group code is referred to as a node.

CAH tier

Node coding format (where nn is a 2-digit integer)

1

CAHnn

2

CAHnn-nn

3

CAHnn-nn-nn

Each node also has a label which serves to represent groups of subjects intelligibly and with rigour. Subject groupings (represented as nodes in the CAH) are not the same as subjects (represented by HECoS codes) even in cases where a CAH label is identical with a HECoS term.

Where a higher-tier node is the parent of a single node at the next lower tier, they must share a title.

Where a higher-tier node is the parent of more than one node at the next lower tier, they may not share the same title.

In version 1.2 of the CAH, level 1 comprises 23 groups; level 2 comprises 35 groups, and; level 3 comprises 167 groups.

Development of CAH versions

Development of the CAH is subject to rules in the HECoS Management Guide. In summer 2018 further research undertaken by the OfS (and DfE) identified areas where intended users of the CAH found groupings to be confusing, and as a result developed a proposal for amending the CAH. HESA worked closely with users of the CAH and Version 1.3 was created and approved by the Data Landscape Steering Group. Within Version 1.3 an additional sheet highlights the changes that have been made between each version.

Version 1.2 of the CAH will be deprecated on the 31st July 2021. HESA is working with public bodies that are using CAH to support their work, as longer timescales are required to make changes to support continued use of CAH for student funding and security. We are working with organisations to determine when CAH 1.3 will be adopted in their processes and we will update this page shortly.

Download the CAH

Download the Common Aggregation Hierarchy (version 1.2)

Download the Common Aggregation Hierarchy (version 1.3.4)

A simplified presentation of the CAH is also available, showing the relationship between the nodes at each tier in a table:

View Common Aggregation Hierarchy - list

Quality characteristics of CAH

In preparation for the release of the HESA 2019/20 Student Statistical Bulletin, we investigated the quality characteristics of the CAH. We compared 2019/20 data with data from 2018/19. We derived the CAH groups for each dataset, to look at changes in how students were distributed in each year. Our preliminary analysis revealed that a small overall average change in the distribution of students masks wider underlying variations. Many individual CAH level 1 and CAH level 3 groups show more substantial changes in the numbers of students allocated to them. These differences appear to reflect underlying changes to courses, as well as movement from the many ‘others in…’ categories within JACS to the precisely defined subjects in HECoS. We determined that these inconsistencies meant we should not use CAH v1.2 for presentation of our initial outputs containing time series. We will undertake further investigation and publish a more detailed analysis shortly.