Joint funding bodies’
Review of research assessment

RA Review update

At its meeting on 17 December the RA review steering group discussed two models of research assessment.

Model 1

Model 1 is a subject based expert review model. It can be understood as a radical redesign of the RAE. The principal differences between the proposed model and the existing process are as follows:

• there would be many fewer units of assessment, with appropriate access to sub-panels and specialist reviewers, some internationally based
• there would be greater freedom for subject communities to determine the way in which they wish to be assessed
• panels would be asked to separately consider a number of aspects of research excellence (e.g. excellence in basic research; excellence in applicable research; development of young researchers). Panels would score each separately, possibly aggregating their scores to produce a single grade
• grades would either be normalised or replaced by rankings
• institutions would submit all staff whose contracts oblige them to do research.

Model 2

Model 2 is an institution-based partially metricated assessment model. This would abolish units of assessment, departing completely from the disciplinary basis of the existing RAE. Under model 2 an institution’s research would be assessed using two separate processes:

• the institution’s research performance would be assessed using a basket of discipline specific metrics
• additionally, institutions would be asked to provide an account of their leading research units. These will, in many cases, be interdisciplinary or multidisciplinary. They will often be inter-institutional collaborations and may well involve partners outside HE. This information would be reviewed by international experts and the best centres, networks and partnerships identified.

The review team is now examining each of these models. In addition it is working on a model which combines elements of both.

This ‘model 3’ would consist of three distinct strands:

• a full research quality assessment similar to that envisaged in model 1
• a light touch monitoring process to guarantee the quality of research undertaken in departments or institutions not participating in the full assessment
• a process for identifying the best research centres, partnerships and networks similar to that envisaged in model 2.

Less research intensive institutions might be encouraged to submit only a selection of units for full quality assessment. They would also have the option of making a single submission covering their research activity.

It is inevitable that, in the course of developing these models, the thinking of the steering group and the review team will continue to evolve. This update should not, therefore, be read as a preview of the group's final report.

The steering group will discuss the models further at its next meeting on 13 February.