Ep. 157: Segregation Research with Brian Kisida

February 4, 2020

Brian Kisida, assistant professor at the University of Missouri, discusses his co-authored report, When is a School Segregated? Making Sense of Segregation 65 Years After Brown v. Board of Education.

Drew Catt: Hello. I’m Drew Catt, EdChoice’s director of state research and special projects. Today I’m in the studio to speak with Brian Kisida. Brian is an assistant professor in the Truman School of Public Affairs at the University of Missouri. And, along with Tomas Maranas and Matt Chingos of the Urban Institute, is coauthor of the report, When is a School Segregated? Making Sense of Segregation 65 Years After Brown v. Board of Education. Welcome and thanks for joining us today, Brian.

Brian Kisida: Hi. Thanks for having me.

Drew Catt: All right, well let’s jump right in. Would you start by telling us about this research and what inspired it, and what ultimately were you hoping to learn?

Brian Kisida: Sure, so like I think many research projects, it was kind of just inspired by dissatisfaction with the current state of research on this area. And this is a long issue that we’ve been dealing with in the United States when we’d call a school segregated. And for a hundred years or more, whether or not a school was segregated was a matter of a law. It was a legal definition because we had de jure segregation in this country and there was sort of no ambiguity what segregation meant. And post Brown v. Board when we started making some attempts to integrate our schools, no longer did we have such clear or 100% minority schools and 100% white schools. It ended up being more of a matter of degrees of segregation.

But our language around this has not really changed. And since Brown v. Board of Education for 65 years, we’ve still used this kind of binary language to apply the designation of segregation to schools. And so we’re kind of dissatisfied with this binary measure. Sort of like a school is segregated or isn’t segregated when really it’s a matter of degrees. There’s also problems with kind of using these types of measures because typically what they incorporate is some sort of arbitrary cutoff. It might be if there’s 90% minority students in a community or in a school, I mean, that researchers have said, “OK, this is a segregated school.” But we have increasingly diverse population of students. It’s now a majority of students in the United States identify as ethnic minorities. So, as communities increasingly look more and more diverse, there’s a mismatch between that and labeling a school as being segregated using these sort of arbitrary measures.

Drew Catt: Yeah, that’s fascinating. And would you mind, so you talked about the de jure segregation. Would you mind touching on the difference between absolute versus relative measures of segregation and kind of getting more into those degrees of segregation that you talked about?

Brian Kisida: Sure. Right. That’s a great segue from sort of like where we were having issues with the binary measures. So, when we talked about these binary measures of segregation, we’re talking about what falls into this category of an absolute measure. And in the segregation literature, there’s really two types of ways of thinking about segregation.

There’s absolute measures and there are relative measures. So, an absolute measure, like I’ve mentioned, might be this sort of measure that says, “OK, 90% minority, therefore this is a segregated school.” There’s no sort of distinction between a school that’s 90% versus a school that’s 99%. There is no consideration of what type of neighborhood that school might be in or what kind of city that school might be in. Perhaps it’s as we see in many neighborhoods and cities in this country, perhaps it’s in a high minority neighborhood or city. And the fact that there is a lot of minority students there only reflects that. And so this is one of the problems with these absolute measures. They’re also difficult to use over time because over time we’ve increasingly had more diversity in America. So, there’s a plethora of studies out there that have said segregation has been increasing over the last 20 or 30 years. Well, that’s only true if we look at absolute measures. And part of the problem there as well, we have an increasingly diverse population. So, that actually makes it look like schools are more segregated using these absolute measures. And then there’s also a problem comparing across different regions and geographical units because if I’m comparing a school in Idaho to a school in Southern California, clearly there’s a different population of students that those schools can draw from. But those absolute measures wouldn’t pick that up.

Segue that to this idea of a relative measure. So, a relative measure of segregation looks at how a population of students in a community is distributed among the schools in that community. And so right off the bat that’s taking into account what the community looks like and then asking how are the students distributed across this community? And so we have measures like the dissimilarity index, the variance ratio index, these relative measures that allow us to take into account these other aspects. These become much more useful when it comes to comparing schools across time because when we’re looking at this relative measure, we’re able to adjust for the underlying population of students. And one of the facts in the literature out there is when we use relative measures, schools are not more segregated today than they were 20 or 30 years ago. In fact, segregation seems to have decreased a little bit.

So, the way that you choose these measures matters quite a bit. Maybe the biggest sort of problem with the relative measures is they’re not as user friendly. It’s a little bit more difficult to understand conceptually than a binary measure of this is a segregated school and this is not a segregated school. But most importantly there are measures that apply to school systems. When I use a relative measure, I’m saying, “What is the relative segregation level in the city of Los Angeles or the city of Indianapolis?” But it’s not a measure that applies to schools because it’s a relative measure. There’s no sort of way to break it down to the school level. And I think this is a problem, I think, conceptually for us, for people wanting to understand this because we’re used to thinking about schools being segregated, not school systems being segregated. So, this is the difficulty in the relative versus absolute distinction in measuring school segregation.

Drew Catt: I’m glad that you brought up the dissimilarity index or the DI, which is something that I’ve seen used in previous research and read about in other reports. But I was fascinated by the segregation contribution index that you all used in the report. So, would you mind explaining to our listeners what this is and why it’s integral to this research?

Brian Kisida: Sure. That’s a perfect segue from what I was discussing with shortcomings with relative measures. Like I said, we would prefer relative measures because of the qualities that they have taking into account the underlying population, but they don’t apply to schools. And we wanted to produce something that we thought would be helpful to policy makers. And if you tell a policy maker, “Your district or your town is segregated,” there’s not a whole lot they can do with that. They really want to be able to think about what is happening at the school level.

So, what we did with the segregation contribution index is we took a relative measure—the dissimilarity index in this case—and we decomposed it to the school level. So, what we’re actually doing with the segregation contribution index is we’re not saying, “Is a school segregated or not,” which we think is overly simplistic. But we’re saying, “How much does a school contribute to segregation in that system?”

Most schools aren’t going to contribute a lot. A school is going to only contribute 1% or 2% to the total system level of segregation. We’re going to set segregation in a system to be 100%. there is a definite amount of segregation that exists in a community and we desegregate it to the school level, which brings us back to that more intuitive way of thinking about things. So, people can think about, OK, where are the schools that are segregated? But in this case, they’re going to be able to answer the question, “Which schools are contributing the most to segregation in my community?” It has properties that are, I think, still better than the absolute binary way of thinking about things because it’s a continuous measure. We’re going to have some schools that contribute 2%, 3%, 5% and we’re going to have some schools that contribute less than 1%. And our hope is that by doing this, policymakers are actually equipped with the tool that they can then use to proactively remedy segregation.

Drew Catt: So, you talked about how different sectors a little bit, or at least kind of started to lead into discussion of the results. So, you all did look at different sectors of schooling that were included in your study. So, what sectors did you focus on and how did the different results by sector shake out?

Brian Kisida: Sure. So, we look at traditional public schools, we’ve looked at charter schools and we actually include private schools with, to my knowledge is something that’s been pretty much ignored in the segregation literature largely because I think there’s just a belief that they’re kind of outside of the system and there’s not a whole lot of policy levers to deal with that. But when we sort of look at every, we do this for all the school communities in the country and we take a view from 30,000 feet. It’s like, OK, well because public schools serve the majority of students in this country and tend to be larger, on average 81% of K–5 students are in public schools. Public schools account for 76% of the segregation in our school contribution index. So, a little bit less than their actual share of students.

Charter schools also track pretty closely to their percentage of students. We estimate in our sample they’re serving about 8% of the students and they’re accounting for only about 9% of the segregation. But it’s private schools where we find that those numbers actually get out of whack in the other direction where charter schools, even though they’re only around 11% of the K–5 population, in our sample they’re contributing 15% to segregation. So, just like on average, their share of segregation is higher than their share of students. And so this is something that I think is important to shed a light on.

We actually do have a second step of this in our paper where we look at sectors in a different way. So, we actually take our school contribution index, which is just kind of like a raw measure of where is segregation attributable to, but we break it down into neighborhood characteristics as well. And so the way that we define this as that we define a neighborhood for a school and we ask, well, look, we know that there’s this constraint that schools have that they can only draw from students within some reasonable area before transportation issues would be too large. So, we define this neighborhood that’s this reasonable travel distance and we look at the demographics of that neighborhood. This is kind of like an addition that we do to the school contribution index. We ask, “Well, given your neighborhood, are you an integrating school or a segregating school? Meaning are you more likely to look like your neighborhood in a way that makes you look more integrated judging by the system as a whole? Or are you actually increasing segregation relative to your neighborhood?” It’s a bit complicated I think probably just to describe over the phone, but anybody can access the report at the Urban Institute website. And I think we explain it a little bit better.

But look, the bottom line when you do this, is that most schools actually look like their neighborhoods. And other people have looked at this research and said, “Yeah, I mean, you know, 90% of schools look just almost just like their neighborhoods.” When we look at public schools, they especially tend to look like their neighborhoods. It makes sense because public schools have boundaries that are drawn to collect the students that are in that area. So, public schools tend to look very much like their neighborhoods.

Charter schools actually also look very much like their neighborhoods, but they diverge more. So, a charter school is equally likely to have more minority students than the neighborhood would suggest, or fewer minority students than the neighborhoods would suggest. But the story here is that it’s symmetrical. And this also makes sense. Charter schools are drawing from a larger boundary. They’re not constrained by school attendance boundaries, so it’s more likely that they’re going to diverge somewhat from their neighbor hood.

And then private schools actually diverge from their neighborhoods in ways that are different than public schools and charter schools. They diverge from their neighborhoods in an asymmetrical way. The bottom line is that they tend to be whiter. And when we look at this instead of just sort of like where is this happening, when we look at urban schools, schools in urban areas, private schools and charter schools and traditional public schools all tend to diverge from their neighborhoods in ways that are fairly neutral. Actually, when we look at the neighborhoods with low black and Hispanic representation that we find that private schools, actually 81% of them tend to have fewer black and Hispanic students than the surrounding neighborhoods, which we find to be quite striking result.

Drew Catt: Yeah, and I’m sure that it matters when you’re using different definitions of what a neighborhood is. It’s like I’m on the board of a school here in town and you know, getting ready to roll out incoming freshman class and we’re just really talking about, OK, is the neighborhood just the district or being an area where both inter- and intra-district transfers are very commonplace? OK, are we looking at just a five-minute drive time, 10-minute drive time? What’s the actual catchment area of the school when you have all of these various methods of choosing schools and accessing those schools? So, I was glad that you did group schools into more than one geographic level or neighborhood. Looking at the district boundaries that do include those zip code catchment areas, but then also the metropolitan areas and looking at the county level. You touched on like some of the private schools are different like here, especially in some of the rural areas, some of the private schools in the rural areas just happen to be situated right on the county line. And I’ve seen some charter schools do the same thing. So, would you mind explaining like why—and I think you alluded to it a fair amount—but why you looked at these different geographic levels and what differences kind of exist in the results when doing it those three different ways?

Brian Kisida: Sure, absolutely. And let me be clear and I apologize. I think this is a difficult topic. There’s so many terms and a lot to wrap your head around here, which is why we named the report Making Sense Out of Segregation. It’s difficult to make sense out of this.

So, we try to define our terms carefully. So, when we talk about neighborhoods, we’re talking about something very local. And then when we talk about the broader things like the district or the County or the metropolitan area, we call that the system. And so earlier when I alluded to differences based on neighborhood characteristics, we’re talking about something like a two- to three-mile radius around the school. And that’s where we find that private schools tend to not look like their neighborhoods. And we’re saying like in a very local way.

The broader issue when we talk about school systems, you’re absolutely right. It’s also important and it’s going to matter in the grand scheme of how this shakes out. It doesn’t matter a lot, but it ends up mattering in some important ways. So, when we build the basic school contribution index, we do this at the district level, the county level and the metropolitan area level. And that’s in order typically from smallest to largest. And the basic findings, even this sort of like the descriptive statistics that I mentioned in the beginning are that private schools tend to account for more segregation than their share. And charter schools do but just slightly. And public schools look pretty neutral. But those are based off of the county level of estimates. Those results also hold if we define the school system as the district.

But when we define the school system as the metropolitan area, the results actually do change slightly. Traditional public schools, that result doesn’t really change. But actually the fine on charter schools actually flips. So, this is maybe difficult to understand. But one way to think about this is that within districts or counties, controlling for school size and neighborhood composition, charters are accounting for slightly more segregation than traditional public schools within that same district. But when we expand the definition of the school system to be a metro area, the opposite is true. So, when we do this, of course, one of the things that’s changing when we look at the metro area is that the yardstick for what we consider segregation changes because the average demographics within the districts are not the same as the average demographics within metro areas. And so when we changed this to Metro area and sort of like the sign, the charter school ends up actually contributing to integration. Not in a causal way, this is just descriptive research. But what the result is telling us is that charters tend to look more like the metro areas in which they reside. Whereas traditional public schools tend to look more like the districts in which they reside.

And again, that intuitively makes sense. The traditional public schools are confined to enrolling schools within a certain district. So, not surprisingly, they look more like the district. Charter schools can attract students from multiple districts. So, when we change the yardstick be the metro level area demographics, charter schools tend to look more like the metro area.

And look, I think this is kind of like mixed evidence that whether you’re an opponent or an advocate of charter schools, you may pick which statistic that you think is right. But I think it’s a nuance story. And we actually have another paper where we do provide causal estimates of the effect of charter schools on school segregation. And in that paper, which tries to use a causal framework, we actually find results that look like these same descriptive statistics in that if we look at the district level, charter schools tend to look like they’re increasing segregation. But then when we look at the metro level, they seem to be decreasing segregation. And the mechanism here seems to be that they’re pulling schools from multiple districts and that can have an integrative effect. Kind of like a not-well-known fact about school segregation is that it’s actually two-thirds of segregation in this country by some estimates occur between districts, not within districts. Districts tend to be more homogenous and this is the way it’s shaken out through years of residential segregation. But between districts, largely white districts right next to largely minority districts, that’s where we see most of the segregation in the U.S. So, it’s not surprising that schools that are drawing from multiple districts might have that effect.

All that said, while the effects for private schools, the amount that they contribute is smaller when we look at the metro area, they are still significantly higher than the other schools even when we use the metro area.

Drew Catt: Yeah. I’m glad you brought up the difference across district lines. I actually did an event here in Indy back in February 2019 partnering with EdBuild, looking at the differences across [inaudible 00:19:47] boundary lines, and then EdBuild kind of launched everything off looking at their national reports, looking at the differences district to district. And it’s fascinating looking at some of the maps that they have on their website. Just the stark differences that can exist between neighboring districts.

Brian Kisida: Right. I agree. It’s starting to get a little bit more traction and I actually applaud EdBuild for a lot of the work that they’ve done on this. They’ve definitely contributed to the conversation.

Drew Catt: Yeah. I think this comes down to something that I’ve been trying to talk about here internally at EdChoice and that’s how do we define community? Which I think some of this research that we’re talking about kind of gets at that. Like is the community just the area around the school? Is the community everything within the district? But especially talking about charter schools, it sounds like the community might be just the greater area in which the charter school exists.

Brian Kisida: We struggled with it as well. I mean it’s why in all the work that we’ve done on this, we typically have defined it three or four different ways because there is no set definition. And obviously the definition can change in so many ways. I mean distance doesn’t quite do it because you could be in urban areas versus rural areas. Distance doesn’t really well-define a community. I think it’s a difficult thing.

Conceptually, I try to think about community being something along the lines of where a school might reasonably be able to draw students from without bringing up insurmountable transportation issues. Because I think at the end of the day what we’re trying to do with this project is we’re trying to say like to policymakers and people interested in this, “Where can we actually affect change?”

So, moving away from this idea that we should be demonizing one school sector or another school sector or just sort of like wringing our hands about a problem that seems insurmountable. We want to break this down into actual bits of information that are small enough but yet meaningful enough that we could do something.

So, one of the hopes here is that with the school contribution index, you’re actually able to say like, “OK, you can define the community this way or you can define the community that way.” It probably isn’t going to affect much in the long run. But here’s the schools that are most contributing to segregation in that community. And now you know, and now you can look at our second modification of this where we bring in the neighborhood aspects and we can say like, “Oh well look, here’s a school that is contributing a lot to segregation. And is even contributing more to segregation than we would expect giving its local neighborhood.” There’s an opportunity for integration. There is an opportunity for integration that isn’t going to require massive, unrealistic resources in transportation and busing issues. But maybe just right there, there are ways that we could make the system more integrated perhaps from like just the perspective of like here’s the low hanging fruit.

Drew Catt: Yeah. And it’d be fascinating to have someone kind of replicate this research. Not looking at even all private schools, but specifically in a given state schools participating in a private school choice program. And also I think it would be great, I don’t know if you already have plans to do this, but whenever that 2020 census data comes out—which I personally have a lot of potential projects built around—if that would change any of the research since there would be better block level data than necessarily the 2010 census block.

Brian Kisida: Yeah, I guess I hadn’t quite started thinking about that yet. But yes, you’re right. The new census data will be coming soon. We’re using all common core data for most part on this, for our measures of demographics. And then I think to your other point, like I agree there are so many research questions that we could ask about this. Specifically the one that you mentioned I’m sure would be important to know if there are these differences in terms of schools that are participating in the voucher program versus ones that aren’t. We are going to be making all of the data publicly available. So, this is something that luckily lots of people should be able to get their hands into and answer these types of important questions.

Drew Catt: Yeah. That level of transparency and that availability is I’m sure going to be very loved and welcomed by the research community. So, I don’t normally ask researchers to do this, but I am curious, Brian, if there was one talking point you want private school choice advocates to have around this research, what would it be?

Brian Kisida: I don’t know if it’s a single talking point, but I would say that it’s a single approach perhaps, which is these are issues that we shouldn’t sweep under the rug. These are issues that people from advocacy organizations on all sides should want to shine a light on so we can work to make it better. I think that confronting things like this, right? So look, I mean if you’re an advocate for private schools, you’re not pleased that they seem to be contributing more to segregation than traditional public schools. But I don’t think that the reaction to that should be anything other than how do we improve it? How do we make it better? How do we shine a light on exemplary models that can take this on and be proactive in finding solutions?

I think that the self-reflection that needs to happen in movements like this is more important than criticisms that occur across sectors and between sectors. And in my opinion, I don’t see enough of this. I think that this is in some ways wrapped up a little bit with the private school choice world being also very pro libertarian, but I don’t think that being libertarian has to mean that we’re agnostic on things that we should value. I understand that it means that those who adopt a more libertarian mindset towards schooling are suspicious of government coercion. That does not preclude private coercion.

There’s a case in Baltimore, I believe, of the private school who is excluded from the voucher program for not serving LGBTQ students. And the current administration joined the lawsuit on the side of the school. And even if we hold aside the legal aspect of this personally, I have no problem with excluding that school if they’re discriminatory. But even if we hold aside the legal principle here and we just think about the moral principle, or the values that the libertarian private school choice community voucher community thinks are important, I don’t see why it’s so rare that I see positions being taken on this where we call something that’s ******, ******.

Drew Catt: Here at EdChoice, we believe in school choice for all. That all families should have the opportunity to choose. But the other side of that is that all school should have the opportunity to choose. I’m not going to stand here and say that every single school has to participate in every single program. If their ideals don’t align, then so be it. That’s not necessarily the best for them.

Brian Kisida: Right. We don’t have to coerce them. But we can still say they’re ******.

Drew Catt: Yeah. I like to challenge everyone. Like, hey, all right. Yeah, there are some amazing public schools here. You’re right. There are also some terrible public schools. There are some amazing private schools. There are also some really ****** private schools. There are wonderful and terrible schools period.

Brian Kisida: Absolutely.

Drew Catt: It doesn’t matter what sector you’re talking about.

Brian Kisida: Right. Well, and just to bring it back to your general question. If you’re working from an advocacy perspective, that doesn’t mean giving your side a free pass. In fact, I think that the best way to work from an advocacy perspective is to hold your own side accountable.

Drew Catt: Back to that lovely word accountable.

Brian Kisida: That doesn’t have to be a government coercion form of accountability. But there’s lots of sunshine and shame that can move schools in a better direction. And I don’t think that we have to be agnostic on what’s better. Access is better.

Drew Catt: Yes, I wholeheartedly agree.

Brian Kisida: Access by socioeconomic status, racial status, lifestyle status, any of it.

Drew Catt: Choice for all.

Brian Kisida: I’m on board with that.

Drew Catt: Awesome. Brian, before we part, any last words? Any forthcoming research you’d like to plug other than what we’ve already discussed?

Brian Kisida: Sure. So, this school contribution index, actually the release of the report was kind of phase one. In 2020, we will put forth a data visualization and data exploration tool through the Urban Institute where you will be able to go and look up your school community. You’ll be able to look at how, we’re still developing it, but hopefully you’ll be able to easily search out Indianapolis or South Bend, Indiana, or whatever school community that you’re interested in and be able to look at which schools are contributing the most to segregation, whether they be public, private or charter. And our real hope is that for one, it would be great if people used this to do additional research. But our real hope is that policymakers, school leaders, school board members, superintendents, will be able to this as a tool to start proactively looking for ways that they can achieve greater integration in their school systems with just data that helps them identify where those opportunities exist.

Drew Catt: Awesome. That is so great to hear. Yeah, the data visualization team over at Urban does a wonderful job. So, really looking forward to seeing how those all look and playing around with the data myself.

Brian Kisida: Yes, me too.

Drew Catt: Awesome. So, thank you so much for joining today, Brian. And for always being forthright and speaking your mind. I’m very thankful for that and I look forward to seeing you soon.

Brian Kisida: Thank you. Thank you so much for having me. Appreciate it.

Drew Catt: Yeah. And to our listeners, be sure to subscribe to our podcast wherever you listen to them for more of our coverage of new school choice research, education reform policy chats and more. Thank you for listening and we’ll see you back soon. With more EdChoice Chats.