2 https://www.researchprofessionalnews.com/rr-news-uk-research-councils-2021-8-black-academics-hit-out- at-ukri-s-lack-of-public-reply-to-open-letter/
control to the funding process in various ways4. Their decisions are pivotal. UKRI grant capture is a significant achievement and is typically career-changing for successful applicants. Invitations to participate in UKRI committees also provide opportunities for career advancement because they represent recognition of significant research experience, and seniority in one’s research field. It is important that committees are diverse, that their processes are inclusive, and that their collective decision-making is as free from bias as possible.
It is generally good practice for any large organisation to audit the composition of its own structures and check that processes are working as they should. It has been known for some time that ethnic minorities appear to be under-represented, and sometimes even completely absent, from some UKRI governance structures. In 2018 it was reported that just five out of 57 appointed RC board members were from Black or ethnic minority backgrounds, and that “AHRC and BBSRC fare particularly badly, each with a 100 per cent White British line-up”7. Individual RCs have occasionally published some figures relating to the ethnic composition of its committees8 but some commitments to do so have not been met. In 2016, BBSRC’s Action Plan for Equality, Diversity and Inclusion9 committed to “publish annually diversity data for applications to and membership of BBSRC Council, advisory and peer review bodies”, but these data don’t appear to have been published. Public commitments to transparency need to be met for the research community and others to have confidence that organisations are genuinely committed to change. Data releases relating to decision-making structures have been partial and inconsistent across RCs, and have typically placed all individuals that belong to ethnic minorities into one category.
4 https://www.ukri.org/wp-content/uploads/2021/03/UKRI-310321-Principles-of-Assessment-and-Decision- Making-V2.pdf
5 https://www.ukri.org/publications/equality-diversity-and-inclusion-strategy-draft-for-consultation/ukri- equality-diversity-and-inclusion-strategy-draft-for-consultation/
6 https://www.ukri.org/our-work/supporting-healthy-research-and-innovation-culture/equality-diversity-and- inclusion/diversity-data/
I record here my thanks to UKRI staff who have met and corresponded with me, worked with me to understand the details of my request, and have combined a large quantity of data across very disparate datasets in different organisations. This was a significant undertaking.
Data were provided for the following RCs:
Data from Innovate UK and the Science and Technology Facilities Council were omitted from FOI responses to avoid invocation of Section 12 of the Freedom of Information Act. Data pertaining to the Councils and Strategic Advisory bodies associated with RCs were not provided on the basis of exemption from Section 20(4). Nevertheless, it is important that UKRI finds ways in which to report this information transparently, particularly in relation to UKRI leadership roles.
Data relating to AHRC, BBSRC, ESRC, EPSRC, MRC and NERC committee membership are considered together in this report, given the broad similarities between their funding schemes, committee structures and operating procedures. Information about RE and the FLF Scheme are provided in the original FOI releases for transparency, but not considered further since one is cross-council and both differ in major respects from the other RCs.
The data fall into three categories:
(i) The numbers of committee places taken up by people in different ethnic categories, reported in two ways:
(a) Aggregating over ethnic minority categories so that data can be disaggregated annually over a five-year period, to reveal trends over time (table 1).
(b) Aggregating over time so that data can be disaggregated across ethnic categories to understand the representation of different ethnic minority categories (table 2)
(ii) The number of committee meetings at which there was no individual present that disclosed their membership of an ethnic minority, by year and RC (table 3).
(iii) Written committee building guidance issued to UKRI staff.
Aggregation and Disaggregation.
No analytical approach is complete unless it considers the heterogeneity of the populations it examines. The work here considers multiple ethnicities because previous work in other areas shows that exclusion impacts some ethnicities much more than others. The failure to consider ethnicities separately can mask large and significant impacts on particular populations. However, sub-dividing the data into increasingly refined sub-categories eventually runs the risk of revealing personal information about particular individuals, so an appropriate balance needs to be struck between aggregation and disaggregation.
It is important to recognise at the outset, rather than as a footnote, that the work presented here is still incomplete because it focuses on ethnicity without taking an intersectional perspective11. Many belong to multiple minorities and live at the intersections of ethnicity, gender, disability, sexuality, and socioeconomic status. They experience marginalisation that is worse, and qualitatively different (e.g. gendered racism), compared with those who do not. UKRI needs to continue the work started here by understanding their own data in an intersectional context12.
There is an important caveat. Although the data reported in the FOIs show that ethnic minorities who self-reported their ethnicities were under-represented or even completely absent from some processes (see below), there might have been additional participants who do belong to ethnic minority categories, but whose ethnicity is not identified and fall into ‘Unknown’ or ‘Not disclosed’ categories. Although the actual numbers of ethnic minority individuals is likely to be higher, it is not clear whether or not the percentages of participants in different ethnic categories would change significantly.
(i) A recently commissioned report for the Royal Society13 14 (Joyce and Tetlow, 2020) shows that in 2018/2019, 18.7% of STEM academic staff were from ethnic minority groups, and this varied considerably across disciplines ranging from 7.5% (Veterinary
12 https://www.aacu.org/publications-research/periodicals/if-not-now-when-promise-stem-intersectionality- twenty-first
14 https://royalsociety.org/-/media/policy/Publications/2021/trends-ethnic-minorities-stem/Ethnicity-STEM- data-for-students-and-academic-staff-in-higher-education.pdf?la=en- GB&hash=22B252EFA4A87B0D869BE288F7EF724F
Science) to 33.1% (Engineering and Technology). Table 1 shows that across six RCs and the five-year period considered, 8% of committee places were taken up by committee members who disclosed their membership of an ethnic minority. There was considerable variation across research councils and years (range: 3% [NERC, 2015/2016] to 13% [BBSRC, 2017/2018]). In one extreme case, there were so few (1-4 individuals; [AHRC, 2015/2016]) that the data were suppressed for data protection purposes. In most RCs, the general trend has been a small but steady increase over time. UKRI will need to work to routinely examine and publish such data, and establish an appropriate baseline against which to compare, and indeed this is specified as a priority in the UKRI draft Equality, Diversity and Inclusion strategy15.
Table 1: Numbers and proportions of committee places taken up, by ethnicity and by year. Data supplied in UKRI FOI2020/00242. Composition of committees by Research Council, over financial years 2015/2016 to 2019/2020 (1 April and ending 31 March). UKRI advises that members in each category have been rounded to the nearest value of five and their respective percentage. Percentages have been rounded to zero decimal places and may not sum to exactly 100%. Null values (..) arise because data are suppressed in line with section 40(2) of the FOIA, and reflect one to four individuals. Note that data were compiled on a per-meeting basis, so if the same member has sat on more than one committee meeting per year, or across years, their data has been included for each meeting. The 'All Years' column aggregates data across the five years. White cells are original data, as supplied in FOI. Grey cells are additional calculations applied to the original data.
(ii) The Royal Society report mentioned above shows that that in 2018/2019, 13.2% of STEM academic staff were Asian and 1.7% were Black. Table 2 shows data aggregated across years and so allows disaggregation across ethnic minority categories, revealing evidence of significant under-representation. Individuals who disclosed their ethnicity as Asian were under-represented across RCs, ranging from 3% (NERC) to 7% (EPSRC). Individuals who
15 “Priority 4A: advancing how EDI data are collected, used and analysed”, in https://www.ukri.org/publications/equality-diversity-and-inclusion-strategy-draft-for-consultation/ukri- equality-diversity-and-inclusion-strategy-draft-for-consultation/
disclosed their ethnicity as Black16 represented 2.1% for ESRC, but were only 0.34% to 0.63% for all other RCs in which Black individuals participated. One of the most striking findings is that there were no committee members who disclosed their ethnicity as Black over a five- year period on any BBSRC committees.17 For comparison, what would these percentages look like if individuals from eligible populations were selected into committees without bias (randomly)? UKRI will need to calculate appropriate baselines using its information on committee and population sizes, and interpret comparisons as part of its work. Whatever the estimated baseline figures, in an unbiased committee building process, it seems unlikely that out of 1525 BBSRC committee places none would be filled by individuals who disclosed their ethnicity as Black over a five year period.
Table 2: Numbers (rows 1 to 7) and proportions (rows 8-14) of committee places taken up by members who disclosed their membership of one of five ethnic categories (Asian, Black, Mixed, Other or White), or did not disclose their ethnicity ('Unknown' or 'Not disclosed'). Data were supplied in UKRI FOI2021/00376, and were the same data provided in UKRI FOI2020/00242, aggregated over the five-year period. UKRI advise that members in each category have been rounded to the nearest value of five and their respective percentage proportions rounded to whole values, so proportions may not add up to 100%. They further advise that the same individual may contribute data to more than one committee meeting, and therefore might contribute more than once to each cell. Information relating to less than five individuals was withheld as personal information under section 40(2) of the FOIA. Where there are no members within a category, a count of zero is included. Grey cells are additional calculations applied to the original data.
(iii) In response to FOI1, UKRI reported the fraction of UKRI committee meetings at which no individual attended who disclosed their membership of an ethnic minority. This measure reveals the relationship between actual participation in UKRI processes and the ethnicity of decision-makers. It relates to committee meetings, not individuals, and asks about the numbers absent from processes rather than those present. Hence, there are no data rounding requirements, and there are no omissions required to protect individual identities where numbers in some groups may be very low. Table 3 shows that a large fraction of committee meetings contained no committee members who disclosed their membership of an ethnic minority. Across RCs and years, this amounted to 51%. There
16 Table 2, rows 8-14, column heading for Black ethnicity; percentages re-expressed to two decimal places.
17 UKRI have advised that members in each category were rounded to the nearest value of five, and their respective proportions were rounded to whole values. Those relating to less than five individuals were treated as as personal information and so withheld under section 40(2) of the FOIA. However, where there were actually no members within a category, a count of zero was used. Any ‘zeros’ in rows 1 to 7 of Table 2 are not the result of rounding non-zero values. This reflects true zero values.
was considerable variation across STEM RCs. MRC reported the lowest proportions of such meetings, indicating higher levels of inclusivity (7% in 2015/2016 to 22% in 2017/2018).
NERC appeared to have the highest proportions of such meetings, and by this measure the least inclusivity, ranging from 63% in 2018/2019 and 2019/2020, to 74% in 2015/2016. By comparison, the non-STEM AHRC had the highest proportions of all, between 44% in 2019/20 and 92% in 2015/2016).
Table 3: Numbers and proportions of committee meetings with no individuals who disclosed their membership of an ethnic minority. Data supplied in UKRI FOI2020/00242. White cells are original data, as supplied in FOI. Grey cells are additional calculations applied to the original data.
These findings prompted requests in FOI2 for more detailed information about committee building processes (information supplied, see point (iv) below). Data on the number of committee invitations to individuals in different ethnic groups was also requested, but only AHRC was able to supply this information because it is not held by the other RCs. This is not considered further.
When interpreting data in Table 3, it is important to form a view on what number of meetings, if any, could be expected to have no attendance from people who disclosed membership of an ethnic minority, based on the numbers of eligible individuals in the various ethnic categories. UKRI will need to form a view of this as part of its work to become a more inclusive organisation. However, by way of example, in 2019/2020, EPSRC ran 103 committee meetings (Table 3) with 840 meeting participants (Table 1). In that year, 12% of participants (Table 1) and 15% of the College of Peer Reviewers18 (905 individuals eligible for committee participation) identified themselves as members ethnic minorities. In this example, UKRI should consider whether one would expect 50% of EPSRC meetings to have no participation from those who disclosed their membership of ethnic minority groups, if members of every ethnic group have an equal probability of being invited to each meeting.
(iv) Is the low participation of ethnic minorities on committees related to the ways in which they are invited to participate? FOI2 requested UKRI to provide “any written guidance issued to UKRI staff about the process of selecting and inviting individuals in the FOI [2020/00242] to committee meetings”, so that one can better understand the basis of committee building. These were provided (see attached). In committee-building guidance, the ESRC was found to be the only RC to mention the need to consider ethnicity when building committees. Guidance to AHRC, BBSRC, ESRC, EPSRC and NERC staff mentions the need to balance gender, but does not mention ethnicity. The MRC invites all members to attend every meeting as a matter of routine, so no guidance is issued.
UKRI was also requested to provide the “median number of invitations to committee meetings issued to individuals over the period in the last FOI”. It was hoped that this would bring to light how frequently individuals in the different ethnic minority groups are invited to attend. Although AHRC were able to provide this data, it could not be provided by any of the STEM RCs. In all such cases, the information has not been held. UKRI have reported that in some RCs, processes are under review. Also, NERC started collecting these data from February 2021, but data collected between then and the FOI release are not provided.
The findings above show evidence of under-representation that has been widespread across RCs and long-standing over a period of at least five years. Yet, it appears to be the case that neither UKRI, nor its predecessor RCUK (formed in 2002), have reported the data, or any coherent and comprehensive analyses. There is also no publicly available, UKRI-wide, evidence-based action plan that aims to rectify the problems revealed in the data (although there are some notable new actions19 20 taken by RCs). Approaches to monitor and tackle under-representation appear to be coordinated at the level of individual RCs, rather than being driven by strategic decisions in UKRI.
The following recommendations are made:
21 https://www.ukri.org/wp-content/uploads/2021/10/UKRI-071021-AddressingUnder- representationUpdate07Oct.pdf
decision-makers, that increases diversity and inclusion, and uses evidence to check that actions have impact.
6. Acknowledgements and Endorsements
I’m grateful to Professor Jackie Hunter CBE (Former Chief Executive, BBSRC)22, the British Neuroscience Association (BNA)23 and to Dr Arun Verma (intersectionality, inclusion and impact specialist) for various roles in helping to shape the report and for endorsing the content. I’m also grateful to Dr Addy Adelaine for wide-ranging and helpful discussions about equity, diversity and inclusion in UK research.
I am Professor of Neuroscience at Royal Holloway, University of London. Prior to this I completed my doctoral and early career training at University College London and the University of Oxford. My research expertise includes brain organisation, the brain mechanisms of learning and decision-making, and the use of functional neuroimaging methods. My research evaluation experience over some years includes core membership of BBSRC Committee A and various international evaluation committees. I have successfully won grant funding from BBSRC. I have governance roles in organisations concerned with STEM (Council Member, Parliamentary and Scientific Committee; Trustee for Research Policy, BNA). I also have experience with equality, diversity and inclusion in STEM and academia (Governance Committee member, Race Equality Charter; Chair of Departmental Athena SWAN Self Assessment Team; nearly a decade of continuous work on Royal Holloway Race Equality Charter self-assessment teams). I also bring to this exercise my identity and experience as an Asian man in academia engaged in STEM research.
22 Letter of Support from Professor Jackie Hunter CBE (Former Chief Executive, BBSRC), appended below.
23 I’m grateful to Joseph Clift (Head of Policy and Campaigns, BNA), Dr Anne Cooke (Chief Executive, BNA), and Prof. Richard Henson (President, BNA) for their guidance and involvement; www.bna.org.uk.
LETTER OF SUPPORT
Professor Jackie Hunter CBE PhD Hon DSc FBPharmacolS, FMedSci, FRSB, FZSL
Former Chief Executive, BBSRC
Ethnic Minority Participation in UKRI Processes: Evidence from Six UKRI Research Councils over Five Years.
I would like to strongly support this analysis and the recommendations made.
When I was CE of the BBSRC, very early on in my tenure, I initiated more detailed analyses on the make up of the BBSRC’s funding statistics and also initiated an equality and diversity strategy for the BBSRC. Subsequently I was then asked to work across all the research councils to deliver a cross council equality and diversity strategy which launched in 2016.
Critical to any initiative of this nature is the provision of data and its use to guide policy and interventions. This evidence report demonstrates the importance of gathering and publishing data with regard to ethnicity. I am sad that it has taken freedom of information requests to obtain the information, rather than it being already published, but am glad that the information existed and was supplied.
It shows a clear need to increase participation from ethnic minorities at all levels of research council processes – from ensuring a proportional application rate through to peer review and research council committee demographics.
I hope the Committee considers the recommendations at the end of the report and finds ways to ensure they are implemented.
Professor Jackie Hunter CBE PhD Hon DSc FBPharmacolS, FMedSci, FRSB, FZSL
Former Chief Executive, BBSRC