Today UK Research and Innovation (UKRI) has published its regular report on the demographic characteristics of the applicants and awardees for our funding. The results continue to highlight both under-representation of people from particular demographic groups among our applicant pool, and frustratingly persistent bias in award rates against people from many of these under-represented groups.
We, like others, are running experiments with the peer review process to investigate and mitigate the causes of these award rate biases. Many of these experiments are designed to test for direct bias against people from particular demographic groups. Most of the evidence to date suggests that this is only part of the problem.
So what other sources of bias might there be?
Of particular concern is the evidence for bias against novelty. One of the most convincing studies exploring this issue involved an analysis of the contributions and careers of 1.2 million US doctoral recipients from 1975 to 2015.
There are two particularly important results from this work:
- people from a group under-represented in a particular field are more likely to make novel contributions than people from a majority group, for example, in nursing, men are more likely to make novel contributions than women
- the novel contributions made by people from minority groups are less likely to be taken up and rewarded than novel contributions made by people from majority groups
Diversity introduces novelty
These are striking results. The first result speaks to the importance of diverse views in research and innovation. This is sometimes referred to as ‘diversity of thought’ and is often considered to be independent of legally protected characteristics. This study clearly shows that inclusion of people from under-represented demographic groups is relevant in the context of providing novel research ideas.
The second result speaks to bias in the system against people from groups under-represented within any particular discipline. The analysis shows that even this can be partly accounted for by the fact that minority groups were found to introduce more extreme novelty.
Novelty was assessed in the analysis by looking for new linkages between previously unconnected terms. Minority groups were found to be more likely to connect more distant terms, and this was less likely to be taken up and recognised by others, regardless of the identity of the contributor.
Creating research and innovation environments that value diversity and creative disagreements
Excellent research and innovation require open-minded exploration of the unknown. And they require researchers to engage constructively with divergent views.
This is why diversity and research culture matter so much. We need to create research and innovation environments where creative disagreement and different ways of looking at things are harnessed to drive progress. In short, we need to include and value difference.
A research and innovation system that is bias against novel ideas is clearly not functioning as effectively as it should. This imperative must inform how we respond to the bias in our system.
The typical response to bias against under-represented groups is to narrow the assessment criteria to focus on those considered objective, such as the number and quality of papers someone has published. This inherently reduces diversity, for example by excluding people who have followed unusual career paths.
To realise the full potential of research and innovation, we must double down on our efforts to support different ideas, different approaches and different research questions. We must value difference, which inherently involves valuing novelty.
