Research intelligence - Gender bias hides, even in open minds

Beyond the ken: research has shown that male and female academics have a gender bias of which they may be unaware

Researchers - their skills honed over years of training in the scientific method - often see themselves as less susceptible to bias than the public. “I judge people solely on merit” is a common response by senior academics asked about their hiring practices. But that idea was discredited in September when a US study published in the Proceedings of the National Academy of Sciences showed that in academia, gender bias is implicit.

In the study, Yale University researchers asked 1 scientists at six universities to review identical CVs purporting to belong to senior undergraduate students that had been randomly assigned male or female names.

The researchers found that in considering the applicants for a laboratory manager position, staff consistently judged male candidates to be more competent and deserving of an extra $4,000 (£2,475) pay on average. They were also more willing to provide male applicants with mentoring and were more likely to hire them.

Women in the study were just as likely as men to make these judgements, and scientists responded no better than control groups.

The research is the latest in a long line of studies to show that despite legislation and greater awareness about the importance of equality, it is not just practical issues such as childcare that continue to prevent women from reaching the top echelons of academia: gender bias is unconscious and pervasive.

Some institutions are beginning to realise the importance of tackling unconscious attitudes.

Durham University, the universities of Manchester, Northumbria and Nottingham, and University College London are among a number of institutions offering training in how to recognise and avoid unconscious bias.

Training usually begins by convincing participants of the fallibility of their own judgement - such as demonstrating how the mind can be tricked into seeing identical lines as being of different lengths, said Binna Kandola, co-founder of Pearn Kandola, a business psychology consultancy that carries out such training for universities and companies.

Participants are then presented with the hard evidence that exists about unconscious bias, such as the Yale study and the long-running Implicit Association Test.

The test, devised by Anthony Greenwald, professor of psychology at the University of Washington, uses the speed with which people categorise words to test for unconscious associations.

From analysis of data gathered since 1998, the test shows that the strength of such biases - often reinforced by social cues - means that even women trained in the physical sciences often demonstrate some level of association of “male with science” and “female with liberal arts”.

Equality to the fore

According to Professor Kandola, bringing equality issues out of the minutiae of human resources procedures and into the realm of personal attitudes is critical in order to tackle the biases that hinder women’s ascent to the top. Once staff accept that bias affects every decision they make, they can look to monitor and challenge each other’s behaviour, especially during the hiring and promotion process.

However, Professor Kandola accepted that doing so could be difficult in academic “fiefdoms” where heads of department want to run their teams without interference.

“I understand that point of view, but there is a degree of arrogance in it,” he said. “One of the things I get told [about academia] is that people don’t get feedback regularly enough, that the doors are closed.”

Claire Warwick, professor of digital humanities and head of the department of information studies at UCL, recently took part in such training at her university. She was “surprised and taken aback” by the extent of the bias it uncovered.

“Academics like to think we’re more objective than other people, that we approach data and arguments with an open mind and therefore people with an open mind. But everyone can suffer from unconscious bias,” she said.

Indeed, researchers’ insistence on their objectivity can make them more resistant to such arguments, which can result in training sessions descending into “pantomime”, Professor Kandola added.

“You come across people who say: ‘I’m very logical and rational, I don’t form first impressions,’ even when presented with the data. I find you have to work that little bit harder with academics,” he said.

Although the training at UCL was not mandatory, as a head of department Professor Warwick was “strongly encouraged” to go along.

To avoid a sense of creating extra layers of bureaucracy, universities could wrap such training into existing courses on, for example, recruitment and selection, said Claire Herbert, senior policy adviser at the Equality Challenge Unit, an organisation that works to further equality and diversity in higher education.

“People often become quite self-reflective and don’t see it as a burden. But with anything, there are always going to be people who are resistant to it, so it’s partly about style, the way it is delivered, and targeting the right people at the right time.”

A recommendation to consider introducing such training in higher education is among the guidance on the impact of unconscious bias in recruitment and selection that the Equality Challenge Unit plans to publish in 2013.

Other recommendations will include collecting and examining recruitment and selection data to investigate where bias is happening - whether in race, gender or other areas - and making the decision-making paper trail as transparent and robust as possible.

“When staff are recruited, rational, evidence-based reasons should be documented to show why the successful candidate was chosen…over the others,” Ms Herbert said.

“While this does not prevent bias coming into play, it helps mitigate it and encourages decision-makers to be more reflective and thorough.”

Where possible, anonymous shortlisting should also be introduced, she said, although she conceded that this works only with professional and support staff, as academics can be identified from the published papers they must submit.

Getting interview panels to “peer-review” each other, with members raising concerns if they feel colleagues are treating candidates differently (for example, by using vastly different body language or tone of voice), is another way of avoiding bias, Ms Herbert added.

In the eyes of Professor Warwick, anyone who sits on an appointment panel should be aware of the possible effect of their unconscious bias - and how even the way in which panels prepare to tackle the issue can have an impact.

Data show that even a difference in phrasing - for instance, saying “we must try to avoid unconscious bias” rather than the negative “we must be careful not to be biased” - makes a difference, she added.

“I would barely believe it unless I’d had it explained to me and knew it had been tested. Even just knowing it helps, and having heard it I can’t forget it,” she said.

Professor Kandola hopes that the growing weight of evidence in this area will help to “reboot” a push for equality and diversity that he believes has stalled - something evident in the fact that just 17 of the UK’s 115 university leaders and fewer than a fifth of its professors are women.

“Some people think [equality] is cracked and don’t talk about it any more,” he said. “But just because there are policies in place, it doesn’t mean it’s sorted.”

elizabeth.gibney@tsleducation.com.