In 2015, a study showing
how implicit bias affects teacher expectations by Seth Gershenson and two colleagues made waves throughout the profession. Gershenson, an associate professor of public policy at American University, has continued to examine the effects of implicit bias and examined how role models can support students to succeed. He talked with Education Post about his research and how the teaching profession is now trying to address implicit bias and promote empathy among teachers.
How do you like your coffee? I drink coffee every morning and some afternoons. My current favorite is the pour-over. There’s a little contraption you put over the coffee pot and it drips in. I had a pour-over coffee in a coffee shop and I thought I could do it at home. I now have a Coffee Gator—it’s a coffeepot with a filter. It’s cheaper than a fancy coffee machine.
How did you get interested in studying teacher expectations? In economics, we think about how people make decisions. Teachers are in a really unique position to share their expectations with students and shape the students’ own expectations. That process might be especially important for students who aren’t talking about college outside of school. I just think back to my own high school and middle school days, [pullquote]having some teachers I felt like made an effort and cared—I was excited to go to their classrooms.[/pullquote] I was lucky to have a dad who made it seem like college was a normal thing to pursue. I was in a conversation with my coworker, Nicholas Papageorge, who had been a Teach For America (TFA) teacher in New York City. He saw the differences among teachers and how some had low expectations. He could see how those kids would internalize those expectations. We started asking questions: Could we find data on expectations? Can we singly identify the effect of those expectations? That’s where the hard work began, and it led to the research we’re talking about today.
Your research has shown that Black teachers are much more likely to see their students as college-bound than White teachers. How did you figure this out? Part of the reason there hasn’t been a tremendous amount of work in this area is because there hasn’t been a lot of data. There is one really good source—the
Education Longitudinal Study of 2002. For every student, two different teachers—in math and reading—were asked how much education they expected that student to complete. We looked for disagreements about whether that student would go to college. Of course, people are going to disagree, because they have different information. Those differences could be almost random. But if there were systemic patterns in those differences, then we could look at those disagreements and predict them.
Are there systematic patterns in different predictions about college among the students in the survey? There’s a pretty big average difference between the White teachers’ assessments and the Black teachers’ assessments. There’s no other confounding factors when the two teachers are evaluating the same student at the exact same time. There’s this discrepancy between how far White teachers and Black teachers expect their Black students to go in school.
What can we do to increase teacher expectations for all students? One long-term implication of the work is to diversify the teacher workforce, which currently is highly White and female. Building up that pipeline will take many years. What can we do in the short-run? More than you might think. I think that’s where the cutting-edge research is happening: in designing effective interventions, how we do teacher training and professional development. This is absolutely not condemning teachers. A lot of this is unconscious bias and people don’t even know it is happening. One area where I think there is low-hanging fruit is training to counter implicit bias.
[pullquote position="right"]There is good and bad implicit bias training.[/pullquote] Bad ones say explicitly, “don’t be biased.” Here, that approach is not good—it leads to a boomerang effect. In the short-term people start thinking about it, but it’s so exhausting they go back to their original place and even revert farther. We need well-designed interventions that avoid that. We need multiple steps. It’s still a light touch. It could happen in a workshop a couple of hours long. It starts by talking about what implicit bias is, that most people have it, and avoids shaming. Then it goes on to talking about constructive ways to make better decisions and interact more constructively with people who are different from you. Increase trust, interact with people, practice in low-stakes situations. TFA includes some direct training on the power of expectations. There’s some evidence that that training changed TFA grads’ racial biases and expectations. I think there’s real potential for reducing implicit bias through these kinds of workshops. The other big component is empathy. For teachers this is really useful—to think about what the world looks like from your student’s eyes, especially if your student is coming from a different background. In one study of five California middle schools, an empathy workshop
significantly reduced suspensions and increased students’ performance. There are some really effective workshops that are light-touch and easy to implement, and we’re developing a research base on them.
Where is your research taking you next? The
long-run effects of having a same-race teacher. We used the
Tennessee STAR data on large- and small-class assignments, which also randomly assigned students to teachers. We leveraged this experiment to ask: What’s the effect of randomly being assigned to a Black or White teacher? Black students in grades kindergarten through third who are assigned to a Black teacher are almost 15 percent more likely to take the ACT or SAT in high school. These are persistent, big effects, that have life-changing results. Now we’re trying to understand why those effects are happening. This builds on the expectations work. Preliminary results suggest [pullquote]having a same-race teacher in the early grades changes students’ effort level[/pullquote] a little bit. Students become a little more persistent. The other thing it does is related to expectations and role models. In eighth grade, students report trusting teachers more and being more willing to talk to teachers about problems. It’s building trust and engagement with school and teachers. I hope this will influence how colleges and universities train teachers and help teachers build relationships outside of the content they teach.