Families choosing schools for their kids can find themselves awash in information, from test scores and demographic data to local knowledge gleaned by talking to friends and family.
That information can feel critical for parents facing high-stakes schooling decisions. But it also may serve to entrench the segregation of schools by race and income. White families tend to avoid schools with many black students, research has shown, and low test scores can push those families away, too — scores that are also tightly correlated with student demographics.
New research suggests that providing parents with a different menu of information could nudge them to choose schools or districts they otherwise might not — potentially helping to create more integrated schools, which have been linked to better academic outcomes for students of color.
In an experiment designed by David Houston, a research fellow at Harvard, participants were presented with a hypothetical: they are moving to a new city and need to choose among different school districts in the area for their children.
All were provided with real data on race and average income for each district. Some also got information about how much students in a given school or district grow academically over time, a metric many researchers consider a better measure of school quality than proficiency rates, since schools don’t control the academic level of students when they arrive.
“Growth also has the nice advantage of being less tied to the racial and socioeconomic composition of the student body,” Houston said.
Other participants got information on students’ absolute test score performance instead, and still others received both growth and absolute achievement.
Houston’s theory was that, since most of the metropolitan areas had “high-growth” districts serving many low-income students or students of color, receiving growth data would encourage some people to consider districts that they otherwise wouldn’t.
He was right. Study participants chose districts that were lower-income and served more students of color when presented with data about districts’ academic growth.
Receiving growth data led participants to choose districts where an average of 36 percent of students were white; those who got only achievement data chose districts that were 43 percent white. The pattern was similar when looking at the districts’ share of low-income students.
The results are still quite limited, since the experiment didn’t involve parents making real decisions for their children. And the effects of having that growth data weren’t huge. But they were statistically significant and held across a variety of demographic groups, including participants who were white, high-income, and actually parents themselves. When Houston re-ran the study with new participants, the results were virtually the same.
“Small distinctions in the measurement of educational performance may be irrelevant in the face of demographic preferences,” Houston wrote, with Columbia’s Jeff Henig. “However, our results suggest some reason for optimism for an approach that could steer families towards high quality educational options that serve a wider range of students.”
Would these results hold in the real world?
Carla Shedd, a professor at the CUNY Graduate Center who has written about school choice and housing, isn’t so sure. She points out that families often conflate school quality with having an affluent, white student body, and that participants in the study might have thought that a high growth numbers indicate an influx of white students, even if the district is predominantly black and Hispanic.
“It’s an opening to think about how we present information to prospective parents,” she said. “But it doesn’t go as far as going into the school.”
Houston doesn’t disagree.
“Choices about where to send your child to school — those are both dramatically more constrained than what is present in the survey experiment,” he said. “That choice is also dramatically more complex than it appears in an abstracted survey context, with lots more information and much higher stakes.”
What the results do demonstrate, to him, is that it’s important to share growth data and explain to parents what it means.
“We should include measures of student growth in our efforts to communicate with the public about school and district quality — and we should put greater emphasis when we do include it.”
Other research suggests that demographics — and therefore biases — play a role in parents’ school choices. In one study of Washington D.C., white parents tended to prefer elementary schools where the majority of students are white; schools with more low-income students also tended to be less in-demand. School test scores and proximity also mattered, according to the study.
Affluent parents often make those choices through the housing market. Another D.C.-area study found that parents were willing to pay thousands of dollars more for homes pegged to schools with higher SAT scores. Other research has shown that even when neighborhoods become more integrated, schools don’t necessarily follow.
There’s less research on the role of growth ratings on school choices. One recent study found that, after New York City released how much individual teachers help students improve on state tests, housing prices increased around schools with many high-growth teachers — particularly in low-income neighborhoods. (The D.C. study found that higher growth ratings didn’t attract families to schools, though that may have been in large part because that information is not clearly communicated.)
The New York City research is consistent with the idea that growth data can encourage some families to reconsider schools they might have otherwise written off. At the same time, it raises the concern that higher demand for the best schools in low-income areas might push out some low-income students of color.
Still, Houston sees power in academic growth data to inform parents.
“When we inaccurately attribute differences in educational quality to school districts because of the students they serve rather than their effectiveness in serving those students, we shortchange both district and student,” he and Henig conclude.