The civil service is one of 30 organisations trialling a new tool to help drive down bias in recruitment, as part of plans to move away from competency-based applications.
The online recruitment tool, Applied, which was launched this week, has been developed by the social purpose company the Behavioural Insights Team, which was spun out of the Cabinet Office in 2014.
It aims to make recruitment fairer by cutting out a number of biases that can affect the application process and going beyond anonymising – or “blinding” – applications.
Civil servants forced to "play the game" to make career progress, claims ex-minister Oliver Letwin
Former Treasury chief Lord Macpherson: don't leave Brexit to Whitehall's "gifted amateurs"
Civil service performance management shake-up will focus on “quality conversations”, says chief people officer
The cloud-based platform is now being used by 30 organisations, including Penguin, Cancer Research UK and Barnados, as well as the UK civil service.
Speaking at the launch event on Thursday night, government chief people officer Rupert McNeil said the civil service was “not getting the right people into the hopper” during recruitment and was relying on out-dated technology.
“[Recruitment] should be a huge, industrially-rigorous process, but if you applied the test that you’d apply to a process like procurement and looked at how, in many cases, recruitment is done, it’s pretty awful from a wastage point of view,” he said.
“We’re making all sorts of errors in dropping people out of the process at the wrong stage, so we want to [starting making] the best decisions.”
He said that the civil service had been using technology for recruitment that was “maybe 40 or 50 years out of date” and no longer fit for purpose. The technology behind the Applied platform offers a “fantastic” way of removing bias, McNeil said.
The platform requires applicants to fill in a series of questions that are based on work they might be expected to do in the role. The aim is to focus on what they can do, rather than what they have done in the past, or the sort of person that they are.
These are anonymised and randomised, and recruiters will review a set of answers to just one question – a process known as “chunking”, which is similar to the way teachers mark exam questions.
The idea is to reduce “halo bias”, where a good or bad answer to one question colours a reviewer’s opinion of all the questions, regardless of their individual merits.
It also allows a number of reviewers to independently assess sets of questions and then compare their results afterwards, rather than in a group setting where people tend to be influenced by their peer’s opinions.
The government’s use of Applied is part of wider plans to move away from competency-based reviews, which McNeil said were “very rigid”.
He said: “[The competency-based approach] has been used more rigidly than it was intended to be used, and to the exclusion of other data that can be used to inform decision-making, and actually to the exclusion of good reasonable judgement by people who are hiring managers.”
McNeil added: “People can get very good at filling in the competency-based forms, but that doesn’t actually tell you a lot about their ability to do the job.”
Applied success
Applied’s chief executive office Kate Glazebrook said that the BIT had tested the platform against the team’s standard recruitment approach of assessing CVs for a recent recruitment round, finding that they would have had to review three times as many CVs to identify the people who performed well at assessment.
Glazebrook added that anecdotal evidence from other users suggested that using Applied cut administration time by two-thirds, with a three times improvement in sourcing of diverse candidates.
The system also uses data gathered on applications to offer all applicants feedback, giving them their overall rankings for each question and information on their broad areas of strengths and weaknesses.
Speakers at the event also identified potential future uses for the system in ensuring that people were applying to the right jobs in the first place and taking the bias out of performance management and appraisals.
McNeil said that performance management was a place “where you get a lot of bias” and that it would “lend itself very much to this type of solution”.
However, McNeil stressed that, although there will be a lot of useful data generated, the new platform is not about automating the recruitment process, but about “making sure as much noise as possible is removed”.