Since the protests over the police killing of George Floyd, companies are looking for ways to combat racial bias and discrimination at work. The stats are jarring. In one study, hiring managers were 74% more likely to hire candidates with white-sounding names when their resumes were identical. Despite research showing that teams with more diversity perform better than more homogeneous teams, Black and Latino workers remain underrepresented in STEM.
Just 3.1% of American tech workers are Black, and Silicon Valley is just 3% Black. And things may actually be getting worse. Around 1% of tech entrepreneurs in Silicon Valley are Black and Black Americans are severely underrepresented in positions of leadership at influential technology companies. In 2017 Fast Company reported there were fewer Black women in tech than there were in 2007.
Nearly every Fortune 500 company offers some variety of diversity training, yet diverse workers still face bias and discrimination. Referred to as implicit bias training, unconscious bias training, and antibias training, these classes are supposed to reduce bias against certain groups, such as women and BIPOC, by making people aware of their biases.
Diversity training is often the go-to after a diversity PR disaster. For example, after a Starbucks barista stopped a Black man from using their restroom without buying something, the company closed 8,000 stores for a half day to put 175,000 workers through diversity training.
Lately we’ve been delving into how to build tech ethically, how to hire inclusively, and how tech can stand for justice. We’re also deciding which diversity, equity, and inclusion initiatives to invest in.
This three-part series will investigate the evidence for and against diversity training, why most diversity training fails, and what companies can do to effectively increase their diversity, equity, and inclusion.
The idea behind diversity training comes from a team of social psychologists at the University of Washington and Yale. They developed the Implicit Association Test (IAT), which is meant to measure “the unconscious roots of prejudice” that affects 90-95% of people. The idea was that once people became aware of their unconscious biases, they’d work toward correcting them.
Hundreds of studies show that bias training doesn’t produce its intended effects.
“Diversity training is likely the most expensive, and least effective, diversity program around,” Frank Dobbin, professor of sociology at Harvard, and Alexandra Kalev, associate professor of sociology and anthropology at Tel Aviv University, wrote for Anthropology Now.
Let’s dig a little deeper into what the studies say.
Dobbin and Kalev go on in their paper to cite numerous studies and meta-analyses that find little evidence that training reduces bias. Even when anti-bias training improved how well people understood or how people felt about diversity, it rarely changed their behavior.
However, the researchers also found a lack of evidence to support the idea that diversity training reliably changes hearts and minds. When they looked at 426 studies where participants took the IAT test before and after anti-bias training, the meta-analysis found the training barely impacted unconscious bias and altered explicit bias even less.
One comparison of 17 interventions found that eight reduced white people’s unconscious bias against Black people. A follow-up experiment tested those eight interventions and one fake intervention. The fact that all nine seemed to work indicates that participants may be learning how to game the bias test rather than unlearning their biases. Furthermore, the effects dissipated within days.
A recent article in The Lancet referred to a recent meta-analysis showing diversity training to be largely ineffective at reducing inequality.
The IAT itself has a reproducibility problem. “Multiple studies have found IAT results are not strongly reproducible in an individual: one day you might have a ‘moderate’ bias and the next day you might come out in the clear,” The Guardian Science Correspondent Hannah Devlin wrote. It may also exacerbate bias. Devlin points to an experiment where aboriginal Canadians reported feeling less valued by white people who took a race-based version of the IAT than white people who took a different version.
However, there are some researchers who take the opposite view. “We believe that pessimism is premature,” Alex Lindsey, Eden King, Ashley Membere, and Ho Kwan Cheung write for HBR. “A recent meta-analysis of over 40 years of diversity training evaluations showed that diversity training can work, especially when it targets awareness and skill development and occurs over a significant period of time.”
Especially in light of recent events, Clockwise and many other companies are looking for ways to combat racial (and other types of) bias and discrimination at work. While diversity training is a popular and intuitive place to start, the research shows starting and stopping there isn’t likely to measurably increase diversity, equity, and inclusion.
In the next post in the series, we’ll examine why diversity training usually fails to achieve its goals. Then in part three, we’ll discuss some ways to actually make progress at your workplace.