(250) 860-3866   (778) 753-0681

Professional Development Articles

Harvard Graduate School of Education: Measuring Implicit Bias in Schools

Sep 29, 2020 jjamison

Measuring Implicit Bias in Schools

A new study finds evidence that teachers’ implicit bias may lead to unequal student outcomes.
August 11, 2020
Colorful profiles

School leaders across the country are reflecting on the conversation about race and racism in America that came to a head this summer, proposing policy changes and professional development. A timely new study from Harvard Ph.D. student Mark Chin and his coauthors, University of Southern California professor David Quinn and Ph.D. student Tasminda Dhaliwal, and Harvard Ph.D. student Virginia Lovison, provides some of the first suggestive quantitative evidence to suggest teachers’ implicit biases may effect student outcomes, underscoring the necessity of these conversations.

“Identifying teachers’ implicit bias is difficult because measuring it is often not part of the review that school districts and school leaders do — it’s not part of the process of looking at teachers in the classroom,” says Chin. He notes that there’s been a lot of theoretical research on teachers’ biases, but quantitative evidence of the relationship between biases and outcomes hasn’t been available until recently. Data on biases from Project Implicit’s white-Black Implicit Associations Test, combined with county-level test score data from the Stanford Education Data Archive and disciplinary outcome data from the Office of Civil Rights provided the team of researchers with an opportunity to probe those connections between bias and outcomes.


Chin and his co-authors present two major findings. However, as this is one of the first studies to quantitatively examine the relationship between bias and student outcomes, stronger causal evidence is needed before drawing firm conclusions and drafting policy. “We can’t say biases are leading directly to disparities. There are potential factors we’re not capturing in our models that could explain the relationship,” he says.

  1. The implicit biases of teachers vary significantly by the race of the individual. Teachers of color were found to have lower levels of pro-white/anti-Black bias than white teachers, with Black teachers having the lowest levels of anti-Black bias.
    • Teachers with lower anti-Black bias tend to work in counties with more Black students. “This is a positive finding, given that you wouldn’t want teachers with strong anti-Black bias serving more Black students. Our results are also potentially showing an explanation for why Black students who have Black teachers have higher outcomes,” says Chin.
  2. Areas with stronger pro-white/anti-Black bias among teachers show larger gaps between test scores and in suspension rates for Black and white students. “The mechanism by which bias ends up influencing both sets of outcomes may be different, so I wasn’t convinced beforehand that we’d find a similar result,” said Chin. This finding comes after controlling for factors like differences in socioeconomic status that might influence test scores or discipline.


Some analyses suggest that smaller interventions to address implicit bias don’t actually result in long-term behavioral change, Chin says. “For an intervention to be useful, we would want to know that it actually leads to lasting changes in behaviors. Otherwise, the intervention may not be a useful investment for schools to spend money on.” 

But this study may underscore the efforts many districts have made to adopt hiring practices that increase the diversity of teachers and school leaders. “Hiring seems like a beneficial lever to pull, given that teachers of color have lower biases, and having more leaders of color in a school might challenge the structures that perpetuate these biases in schools and districts,” says Chin.

With a teaching force that is 75% white, however, it’s also important to focus on supporting white teachers in learning to recognize and monitor their own implicit biases. In the last few months, many policy recommendations have pointed towards implicit bias trainings to effect this change. And yet, as Chin notes, individuals’ biases are formed and perpetuated by systems. “I would argue what’s more important is to ensure the systems in a school, in a district, are set up so they don’t replicate symbols of systemic inequality,” he says. “In the long term, bias training only does so much if society doesn’t change too.”

Share this article: