Psychologist logo
Government and politics

Five minutes with… David Halpern

Ella Rhodes poses the questions to the Chief Executive and Board Director of the Behavioural Insights Team.

05 April 2017

The Behavioural Insight's Team, a social purpose company set up by the UK government to apply behavioural science to policy, has recently released a new hiring tool, Applied. The platform aims to remove the bias inherent to hiring by removing the use of CVs and rather asks potential employees questions related to a specific role; their answers to these questions are anonymised and rated by people within a given company or organisation. Ella Rhodes spoke to Dr David Halpern, Chief Executive of the team, psychology graduate and one of those behind the creation of Applied [see also tinyurl.com/halpern0611].

Where did Applied emerge from?
The Behavioural Insights Team was created to identify certain problems which have a certain especially strong behavioural or cognitive elements to them and see if we can figure out solutions. Sometimes that takes us to legislation or trials, but in some cases we think the answer is a product or service. Applied was one of those issues. We were drawn into it by the previous and current Prime Ministers around issues of discrimination and disadvantage. The Behavioural Insights Team was also growing very rapidly as an organisation and having to do a lot of recruitment; we were intrigued by the question of 'What would good look like?' from a hiring perspective.

What are some of the biases and weaknesses in the hiring process?
I used to lecture on this and was struck by the low predictive validity of most of the selection techniques which are used. The conventional unstructured interviews rarely achieve acceptable predictive validity, and then there's lots of more specific concerns around gender or race discrimination. When people are doing selection they think they're weighting many different dimensions, but in fact they often collapse on one or two dimensions and converge on an answer about a candidate much too rapidly. Tools that are better at prediction are those which incorporate work samples, an actual sample that's behaviourally close to what the actual job will be.

How does Applied work and how is it different?
A key element of the Applied platform is it does something called chunking. So if you want to ask a series of questions about whether someone will be a good journalist for The Psychologist magazine, for example, you need to ask what would be the questions that are close to work samples? What are the questions and skills you'd want to know about? Once you have the answers to those questions they are bundled anonymously and you have groups of raters who rate each answer separately. They don't see only one person's answers… they see a string of answers to the same question and rate how good, or not, each one is. If you handle your first question in an interview really well or really badly that can overshadow the rest of the interview, but chunking can remove that problem.

How have you looked into the effectiveness of the technique?
We realised we had to run this as a trial for ourselves and see if it performed better. This for me was a key bit of evidence, and whether we progressed it depended on the results of that. We were doing a recruitment exercise to the Behavioural Insights Team with about 800 applicants; we ran two different processes, one using the then-prototype Applied platform to see how it would perform, the other was a conventional CV sift. The candidates then go to a series of very detailed interviews and other processes, and the question is, which group of candidates did better? Applied performed much better than a conventional CV sift. We brought through candidates who wouldn't have made the cut based on their CVs alone, and went on to appoint them on the basis of using Applied.

What does the future hold for the Behavioural Insights Team?
We continue to put a steady stream of new results out into the field and we're moving beyond the nuts-and-bolts issues, like what's the most appropriate letter to make someone pay their tax on time, to much bigger, so-called wicked issues. The BIT has also been shortlisted to win the MacArthur Prize of $100 million working with David Miliband's group (the International Rescue Committee) and Sesame Street, in developing educational content for refugee children and their parents and caregivers in Iraq, Jordan, Lebanon and Syria. Hopefully other psychologists are encouraged that psychology is finding its way into policy in a way it hasn't done before.

To find out more about Applied see tinyurl.com/kbatd3p