It’s a common accusation leveraged against schools of choice, and especially charter schools: They “skim” students who they perceive to be easier to teach and avoid those they perceive harder to educate. In this clever new study, analysts Peter Bergman and Isaac McFarlin conduct an experiment to test whether that’s true. They send emails from fictitious parents (sneaky!) to nearly 6500 schools of choice—that includes traditional public schools in areas with intra-district school choice and charter schools—in twenty-nine states and Washington, D.C., asking if any student is eligible to enroll in the school and, if so, how to apply.

More specifically, the experiment randomly assigned an attribute to the child (or no attribute). The control or baseline email read, “I am searching for schools for my [daughter/son]. Can anyone apply to your school? If so, can you tell me how to apply?” Then they added a sentence to indicate one of four other treatment conditions: 1) a special needs requirement (“Her IEP says she has to be taught in a separate room”); 2) bad behavior (“She gets in trouble a lot for behaving badly in school”); 3) poor grades (“She’s been getting bad grades”); and 4) good grades (“She has good grades and good attendance”). They also randomly varied four other variables: students’ implied race through a “black-sounding” or “Hispanic-sounding” name, the household structure (implying a one- or two-parent household—i.e., “My husband and I are searching…”) and the gender of both the parent and the student.

They did two experiments; the first included charter schools only, and the second sampled from the largest forty districts in twenty-nine states and Washington, D.C., with intra-district choice and charter schools. The results from both were similar. For the latter, they matched charter schools to the nearest traditional school with the same entry grade and within the same district boundaries. They also use a variety of demographic, achievement, and census tract data from multiple sources as controls.

The key finding is that all schools of choice (traditional and charter) are less likely to reply to messages signaling that the potential applicant has bad grades, poor behavior, or an IEP, compared to the baseline message, which had a response rate of 53 percent overall. The email indicating poor behavior decreased the response rates the most, a reduction of 7 percentage points. The email signaling good grades and attendance had no discernible impact. In terms of the other randomized elements, only the message that indicated a Hispanic-sounding name resulted in significantly lower response rates at 2 percentage points.

When breaking out traditional public schools located in areas of choice versus charter schools, analysts find that both generally responded at similar rates, with one exception: Traditional schools are more likely to respond to the IEP message—by 5.8 percentage points—than are charter schools. And high value-added charters are less likely to respond (by 10 percentage points).

Analysts hypothesize that the higher cost of educating special needs students for a single charter school versus one affiliated with a school district (or local education agency, a.k.a. LEA) may make a difference, so they examine that too. But they find no difference in responses relative to whether the charter is its own LEA or associated with an LEA. However, they do find that charter schools in states that reimburse schools or districts for a portion of their realized costs of serving these students are 7 percentage points more likely to respond to the IEP treatment message than charter schools in other states. This suggests that funding or lack thereof is a key barrier for charters in serving children with special needs. Analysts also posit that their particular IEP message, which signaled a student who required a restrictive environment (“Her IEP says she has to be taught in a separate room”) could be particularly costly and therefore may have impacted the results.

To summarize, both charter and traditional schools of choice are equally likely to balk when presented with the opportunity to enroll students who are lower performing or costlier to educate. In other words, it’s not at all clear which is the pot and which is the kettle.

An open enrollment policy is the right and noble way to go. But we also can’t ignore the reality that students come into school with any number of needs, interests, and challenges. Instead of pretending otherwise, let’s not only support schools in embracing all children, but commend those that excel in educating kids who don’t fit the mold—and incentivize more of them to open.

SOURCE: Peter Bergman and Isaac McFarlin, Jr., “Education for All? A Nationwide Audit Study of Schools of Choice,” National Bureau of Economic Research (December 2018).

Amber Northern is senior vice president for research at the Thomas B. Fordham Institute, where she supervises the Institute’s studies and research staff.  She has published in the areas of educational accountability, principal leadership, teacher quality, and academic standards, among others. Prior to joining Fordham, she served as senior study director at Westat. In that role, she provided evaluation services…

View Full Bio