The importance of open diversity data
By Dan Hodges, Deputy Director of Analysis, UK Research and Innovation (UKRI)
I am passionate about using data to inform and improve everything we do at UKRI. Data analysis drives our work in key areas, from regional distribution of funding to measuring the impact of international collaboration. It is also key to our current effort to respond to the coronavirus pandemic, which has put the research and innovation community under an unprecedented level of strain.
This week, I am pleased to see us publish our first annual release of harmonised diversity data for funding applicants and recipients for the past five academic years. Research council diversity data has been published in the past, but this publication is the first of its kind, for several reasons:
- For the first time, data across all seven research councils is presented as a harmonised UKRI data set
- The publication includes new, previously unpublished, data on award values
- We are offering the material in a range of different formats to help the community access and analyse our data more easily
Making our diversity data openly available is crucial in our role as the major public funder of research and innovation in the UK.
We know that there is a lot of interest in this data in the community – this became clear last autumn when we responded to an initial data request from the Science and Technology Committee’s inquiry into the impact of science funding policy on equality, diversity, inclusion and accessibility. We welcomed the conversations this publication stimulated and look forward to continuing them – in even greater depth – following the current data release, as well as in the longer term.
Transparency is key when it comes to promoting equality, diversity and inclusion (EDI). While there are elements to celebrate in the current data release, we know that some of the findings are a cause for concern. For example, there are clear disparities in award rates and award values between male and female applicants, and between white and ethnic minority applicants.
We caution against drawing quick conclusions from these findings, without controlling for other factors and carrying out more advanced data analysis, which my team are continuing to work on.
At the same time, across UKRI we are committed to addressing disparities and inequalities where they are found. My team works closely with our policy colleagues to develop the best approach to tackle challenges and findings from our data continuously inform UKRI EDI policy. This has never been more important than during the current crisis. It is increasingly clear that the impact of the coronavirus pandemic will not be equal. We are looking into ways of monitoring our applications data to understand how different groups may be impacted and what we can do to mitigate adverse and disproportionate effects.
Data not only helps us identify problems, but also deliver solutions. It ultimately holds us accountable and helps track our progress. We will continue to publish findings from harmonised diversity data annually and, in the data narrative report (PDF, 600KB) have also announced further ambitions for our data work:
- Collect more data: We currently collect data on four diversity characteristics (age, gender, ethnicity and disability). In conversation with the community, including data specialists and regulators, we’re exploring the possibilities and benefits of collecting data on other protected characteristics. We’re also looking into adding additional gender categories and are preparing to release results by detailed ethnicity characteristics.
- Advanced data analysis: Work is underway on more advanced analytical projects related to our diversity data, such as intersectionality and regression analysis, to better understand trends. Within UKRI, we’re investigating ways of analysing EDI data within cross-cutting funds. We’re also preparing to publish and analyse our own employee diversity data.
- Working with the sector: We want to lead the public sector conversation about how to best collect and use diversity data to maximise benefit for all. For example, we want to investigate what is needed to increase disclosure rates for disability (which we believe may be consistently underreported). We will continue to consult on best practice with the EDI data community, e.g. when it comes to presenting our data through visualisations.
If you have feedback on our approach or would like to discuss any aspect of this data release, please contact us at email@example.com, noting in the subject line if your query is specifically related to our data analysis work.