Proudly supported by The University of Canberra and The Institute for Governance & Policy Analysis.
The surprise swing in 'unconscious bias'
Q & A:
Are recruitment methods targeted at eliminating unconscious bias doing more harm than good? According to the findings of a recent report by Professor Michael Hiscox and the Behavioural Economics Team (BETA) in the Department of Prime Minister & Cabinet, efforts to assist the promotion of diversity in executive levels through the de-identification of applications may have unintended consequences. The study, Going blind to see more clearly: Unconscious bias in Australian Public Service shortlisting processes was the first of its kind, involving over 2,100 participants from 14 agencies in the Australian Public Service. And the findings surprised... well, everyone!
The study assessed whether women and minorities are discriminated against in the early stages of the recruitment process, and whether 'blind' or de-identified recruitment strategies had an impact on these outcomes. It found that because APS officers generally discriminated in favour of female and minority candidates, de-identification may set back efforts to promote diversity in senior management. So does this mean that blind reviewing doesn't work? Not exactly, say the BETA team.
Q: What motivated the ‘Going blind to see more clearly’ study?
The driver for this study was quite simple: women make up more than half of the APS, but at senior levels the proportion of women falls significantly. One possible reason for the lack of female representation at more senior levels is bias; that is, people making hiring and promotion decisions may be, consciously or unconsciously, assessing female candidates unfairly, making explicit or implicit assumptions that women are less able than male counterparts to perform senior management roles.
One possible reason for the lack of female representation at more senior levels is bias.
If this is the case, introducing de-identification into the recruitment process could be a solution that would make the process fairer and improve diversity. A number of APS agencies have been experimenting with de-identifying CVs as part of their recruitment processes. We wanted to find out the impact de-identification can have on the shortlisting phase of APS recruitment processes.
Q: The findings from this report are quite surprising. Were the team surprised by the results that participants were more likely to engage in positive discrimination towards female and minority candidates?
We were surprised by the results too. We thought de-identifying CVs could be a solution that would make the shortlisting process fairer and improve diversity. We see the results of the trial as evidence that recruiters in the APS are supporting diversity as an organisational priority with a very subtle form of positive discrimination in favour of female candidates.
Q: The report suggest that it may be more valuable to focus on other stages of recruitment instead of the initial review process. What methods of overcoming unconscious bias might be implemented in these stages?
There are other stages of recruitment where bias may be more pronounced. For example, job advertisements may be worded in ways that discourage (or encourage) applications from candidates of a specific gender. Making sure that job advertisements use gender-neutral language and emphasise only the key requirements of the role may help to eliminate any bias.
Making sure that job advertisements use gender-neutral language and emphasise only the key requirements of the role may help to eliminate any bias.
Job interviews and review panel discussions may also be structured and conducted in ways that make some candidates less (or more) likely to be favourable assessed compared with candidates of the opposite gender. Re-working the recruitment process to rely more on work-sample tests, using structured interviews, and involving multiple people in the recruitment process instead of relying on a small panel may all help to overcome unconscious bias in recruitment decisions.
Q: Would these policy lessons be applicable when recruiting non-executive positions?
These interventions appear likely to be applicable to all levels of recruitment. But we also know that context matters, and understanding the current workplace environment appears important in determining whether blind recruitment may be suitable.
Q: What about other sectors?
It would be really interesting to understand the impact that de-identifying CV’s has in other sectors outside of the APS. We recognise that every sector has a different status quo setting and internal culture so it’s therefore possible the results could be different.
Q: Where to from here?
The first step is to take this trial into the field and conduct it using a real recruitment process. We also think it's important to consider things like gendered language in adverts, and reducing bias at the interview stage and in the final selection of offers made to candidates.
The report, 'Going blind to see more clearly: Unconscious bias in Australian Public Service shortlisting processes', can be downloaded from the Department's website.