Improving representation with values-based recruitment
Kylie Norman and Dr Michelle Morris know a lot about the uses of data. So when all the data told them that there was significant underrepresentation of minority groups and people with protected characteristics in their own field, they and colleagues at Leeds Institute for Data Analytics (LIDA) were keen to work to change this.
Perhaps you’ve heard of positive action recruitment. Perhaps not.
But however blank your face right now, it’s worth understanding what it’s really all about – because misconceptions can mean it sometimes comes in for unfair criticism.
Positive action recruitment is a provision of the Equality Act 2010 that allows employers torn between two or more candidates of equal merit to take into consideration whether one has a protected characteristic that is underrepresented in the workforce.
In a nutshell, if you’re hiring two people with equally strong CVs and skills, you’re allowed – but not required – to choose one of them because there are few workers of their age, disability, gender, race, religion, sexual orientation or marital or maternity status in your industry.
It’s why, for example, a bank could legally decide to appoint a woman to a senior role normally performed by men over a man just as qualified as her.
So that’s the theory. But how on earth do you go about it in practice?
“LIDA is a forward-thinking institute with progressiveness and inclusion at its core, so we wanted to answer that question to not only help ourselves, but support others to achieve greater diversity across the board,” Kylie said.
“Thanks to multiple articles, blogs and academic papers on the subject, we were fairly confident that data science as a whole and our organisation specifically are not as representative as we could be.”
In fact, in 2021, only 4% of the total data scientists employed by LIDA’s Data Scientist Development Programme (DSDP) since it was founded in 2016 had been black.
As part of its values-based recruitment, the team at LIDA decided it was time to develop their own approach to positive action recruitment – and set an example that other departments across the University could replicate themselves.
LIDA is a forward-thinking institute with progressiveness and inclusion at its core.”
Reaching out to wider audiences
There may have been an evident representation issue at LIDA, but if there’s one thing the institute is, it’s collaborative.
“LIDA is a cross-faculty platform at Leeds, so all of us are employed by different departments and work in all kinds of areas,” Dr Morris, an Associate Professor in the School of Medicine, explained.
“It’s a melting pot of different disciplines and career stages,” Kylie, Senior Operations Co-ordinator of LIDA’s Consumer Data Research Centre, added. “What brings us together is our passion for using data to effect positive social impacts.”
But to create a case for positive action, LIDA would have to look beyond its own doors.
To go further together, Kylie, Michelle and their peers worked with funder DATA-CAN, the UK’s health data research hub for cancer; its partner, Health Data Research UK (HDR UK), had already demonstrated how positive action recruitment could succeed through the HDR UK Black Internship Programme launched in 2020.
With their support and links to new audiences, LIDA was able to ringfence one of the roles on the DSDP specifically for the recruitment of black data scientists – a move intended to encourage more applications from underrepresented candidates and play an equity-signalling role in tackling the underrepresentation within data science.
The team’s work was not without its challenges, however.
“There were so many discussions with stakeholders,” Kylie said.
“We quite rightly had to have lots of meetings with, among others, central representatives from HR, legal and Equality, Diversity and Inclusion at the University, because it was absolutely vital that we ensured compliance with the Equality Act 2010 at all times.”
Another hurdle was that, while wider industry issues of underrepresentation were clear, the granularity of data held by the HR team was insufficient to prove the same problem in the DSDP positive action case – meaning LIDA had to use data already amassed by the Office for National Statistics and HDR UK in support of its own programme.
The success of self-selection
Once LIDA’s case for positive action recruitment was approved in June 2021, the ringfenced role could be advertised.
Crucially, it was down to applicants to decide whether or not they wished to self-select for it specifically – and even more importantly, recruiters were kept entirely blind to this information during both shortlisting and interview stages, which progressed as before.
“I think that’s absolutely vital”, Kylie said.
“By enabling people to identify themselves as being from an underrepresented group while applying for a role that has been reserved for them, you effectively encourage people to apply for positions that they might not have previously related to.”
The results speak for themselves.
The 2021 recruitment campaign saw a 33% conversion of black applicants to data scientists – in stark contrast to that 4% black data scientists figure mentioned earlier.
Clearly, positive action is opening up a large pool of untapped talent for LIDA, just as was intended.
Yet it is only one of the values-based recruitment methods used for the DSDP; a diversity of graduates, from BSc to PhD, is employed per cohort from a range of disciplines, helping to foster a breadth of interests and innovation opportunities.
That relates to the DSDP’s mission statement, “data science for public good”, which candidates are asked to define in their own words during the interview process.
And for Michelle, it all comes back to diversity and representation.
“We’re all motivated in our work to do things to make a difference,” she explained.
“Data science is a very applied thing; we’re taking real-world data to solve real-world problems and share them as widely as we can with a range of different stakeholders straightaway.
“One clear way of doing that more effectively is to have a workforce that is more representative of the people whose problems we are trying to solve – so increasing diversity was and is a no-brainer.”
An act to follow
Solving the problem of underrepresentation is a common goal for many employers, not just at the University of Leeds.
To that end, LIDA has shared a guide to positive action recruitment that can help others follow its lead and encourage more people from minority groups to break the stereotypes and pursue their dreams.
Recommendations include building a provable case (as required by the law), gaining funder backing if possible, securing evidence of the underrepresentation you wish to address, gaining the approval of internal stakeholders, and agreeing specific wording for use in job descriptions and adverts.
The Strength in Diversity case study also highlights some of the challenges employers might encounter along the way, not least of which is the time involved in building a case for positive action.
For Dr Morris, publishing these insights for all to see is an act of data science for public good in itself.
“It’s great to have been at the forefront of doing this kind of recruitment first,” Michelle said.
“It was difficult to do it at the time last year, but now we know how to make it work, we can share those lessons with colleagues and people outside academia.
“Wherever you work, there’s nearly always an area that’s underrepresented in your discipline – and positive action recruitment is something you can do to tackle that.”
About Michelle
Dr Michelle Morris is Associate Professor of Nutrition and Lifestyle Analytics in the School of Medicine, based in the Leeds Institute for Data Analytics and Leeds Institute of Medical Research.
She is the Academic Chair of the Data Scientist Development Programme and leads its Academic Advisory Group, LIPAG.
She is an interdisciplinary researcher whose background spans health informatics, geography, nutritional epidemiology and health economics.
Michelle currently leads the Nutrition and Lifestyle Analytics team at Leeds, focusing on the use of new forms of ‘big’ and spatial data in health research.
About Kylie
Kylie Norman is Senior Operations Co-ordinator of the Consumer Data Research Centre within LIDA, as well as Programme Co-ordinator for the LIDA Data Scientist Development Programme.
She has managed the DSDP for five years now and specialises in delivering mentorship to its early career data scientists.
Her role also includes recruitment, building relationships with funders, partners and academics, as well as co-ordinating the Programme’s Academic Advisory Group, LIPAG.
About Indumini
Indumini Ranatunga was a LIDA data scientist in the sixth cohort of the DSDP (2021-22).
She produced some great work on using data to assess the efficacy of transport infrastructure in the North of England, as well as simulations of pedestrian travel patterns, and was one of Kylie’s mentees.
She now works as a Software Developer for the University of Manchester and is an active member of the LIDA DSDP alumni community.
Audio clips
Hear Kylie and Michelle explain why they think it’s vital to recruit from a wider pool of talent – and help others do the same:
Built with Shorthand