« The Real Impact Of Cutting Medicaid -- Just When We Need It Most | Main | Economic School Integration Update »

May 10, 2011

The Path to Economically Integrated Schools: A Response to Matt Yglesias

Richard Kahlenberg

Last October, The Century Foundation released a study by Heather Schwartz of the RAND corporation, finding that public housing students in Montgomery County, Maryland who had a chance to live in low-poverty neighborhoods and attend low poverty elementary schools performed far better than similar students in higher-poverty neighborhoods and schools.  Over time, students in low-poverty neighborhoods and schools cut the large initial gap with middle-class students by half in math and by one-third in reading.  The students performed at 0.4 of a standard deviation better than comparable low-income students in higher poverty neighborhoods and schools. Such large positive effects are rarely found in educational research.  The study’s findings were featured on the front page of the Washington Post, and, more recently, in a column by Bob Herbert for the New York Times.

The Schwartz study was remarkable not only because it found such big effects from a progressive housing program that promoted economic neighborhood and school integration, but also because it found that the students outperformed colleagues in higher poverty schools (designated as “red zone” schools) that spent substantially more resources on good programs, like reduced class size, improved professional development for teachers, and extended learning time.  Moreover, because families were randomly assigned by lottery to public housing in Montgomery County, the superior performance of students in low poverty neighborhoods and schools could not be dismissed as a reflection of self-selection related to higher levels of family motivation.   

Schwartz found that the best returns came from students attending schools with up to 20% low-income populations with more modest gains up to 35% low-income.  Some commentators have seized on this finding to suggest that economic integration will be difficult to implement on a widespread basis, given demographic constraints nationally.  The Center for American Progress’s Matt Yglesias, a smart observer of a variety of policy issues, noted in his blog yesterday: Given that 21% of American children live in households below the poverty line, and even more fit the standard of “low-income” (those eligible for subsidized meals, from families making up to 185% of the poverty line) “it would be literally impossible for all children to go to schools where fewer than one-fifth of the kids come from lunch discount families.”  If 20% were a magical figure nationwide, this would indeed present a major problem for advocates of socioeconomic integration.

But a long line of studies conducted over many years put the negative effects of concentrated poverty at 50% and even 75% low-income. For example, a 1992 U.S. Department of Education study of 20,000 8th grade students found the negative effect of concentrated poverty kicked in at the 50% low-income level, suggesting policies that allow students to avoid such high levels of concentrated poverty could be beneficial.  Likewise, a 1993 Congressionally-authorized study found that “the poverty level of certain schools placed disadvantaged students in double jeopardy.  School poverty depresses the scores of all students in schools where at least half of the students are eligible for subsidized lunch and seriously depresses the scores when over 75 percent of students live in low-income households.”

Moreover, Schwartz was careful in her study to discourage a broad interpretation suggesting that  20% low-income – or even 35% low income – was a “tipping point” for schools.  The problem with drawing that conclusion from the Montgomery County study is that the county includes very few truly high poverty schools.  Only 5% of elementary schools in the study had  more than 60% of students from low-income families.

On pp. 20-21 of the report, Schwartz notes that while students in up to 35% low-income schools do no better than those in 35-60% low income schools in Montgomery County, the study did not have sufficient numbers of high poverty schools to determine whether students in 35-60% schools might in fact do better than those in 60-100% low income schools. She writes: "Given the lack of truly high poverty schools in this sample, this study does not suggest that 35 percent school poverty is a tipping point, after which low-income students no longer benefit from socioeconomic integration."

In sum, the Montgomery County research suggests economic integration is a very powerful lever for raising the academic achievement of low-income students, particularly when they are in schools with less than 20% low-income students (compared with 35-60% low income students.)  But a long line of research has also suggested positive benefits of avoiding schools having more than 50% low-income populations.  In a country whose student population is still predominantly middle-class, we ought to find ways to allow more students of all backgrounds to attend strong, majority middle class schools.


TrackBack URL for this entry:

Listed below are links to weblogs that reference The Path to Economically Integrated Schools: A Response to Matt Yglesias:



How about high income students? Is there reason to think that presence of 35%, or even 20% of low income students in school does not negatively affect scores of all other students?

As Yglesias correctly observes, the percentage of "low-income" students in this country is very high, not the least because traditionally low-income minority groups such as Hispanics have much higher birth rates than whites. Your statement that "the student population is still predominantly middle-class" may still hold in the Midwest, but, say, in California, the percentage of low-income students (those receiving free/reduced lunches) among all K-12 students is - get this - 55%! How do you expect to make the case for integration there?

Richard Kahlenberg

Good question about California. I have two responses. First, even in a state with more than 50% of students from low-income families, it is possible to integrate a very large subset of schools, many more than are integrated today. For example, in Chicago, which is 85% low income, the goal is not to make all schools represent the district wide average, but to begin by integrating a subset of schools, which will strengthen the system and entice more middle class families to move from private to public schools. Second, take steps to reduce poverty -- which is, of course, important in its own right, and will also make socioeconomic school integration more feasible. We shouldn't assume that the percentage of low income students is unalterable.


I'm still not convinced that integration at the proposed scale is not a zero sum game (in other words, that it does not raise scores of disadvantaged students but simultaneously lower scores of high-income students by the same amount). And it's not just me: Many middle-class and particularly upper middle class whites are strongly opposed to the idea of integration (as numerous busing incidents in this country show) and they tend to begin pulling their children out of public schools, or to steer away from such neighborhoods when house-shopping, when the percentage of subsidized lunches gets above 20% or so. Not because they are all racist, but because they believe that low subsidized lunches mean better education.

For the same reason, too aggressive integration policies ultimately shoot you in the foot by creating white flight from the public education system. SF is a nice example in California: city officials tried to integrate things, and ended up with a city where median family income is $86k, but 57% of all students in the public school district come from low-income families.

Taking steps to reduce poverty is always a laudable goal, but poverty with regard to its impact on education may only be a derived indicator. The primary indicator is probably parent IQ or at least parent education level. There's nothing we can do about parent IQ (other than discouraging low-IQ parents from having children - which would create a massive public relations scandal if anyone dared to propose it in public). Average parent education can be improved in the long run, but it takes decades to make any changes.

Brian Beabout

For pretty convincing debunking of the "zero sum" conception of student achievement, see Dan Condron's recent piece in Ed. Researcher analyzing PISA data.

It's here: http://edr.sagepub.com/content/40/2/47.abstract

He finds that in affluent societies more integrated schools (economically) not only produced higher average achievement but higher percentages of VERY HIGH achievers and fewer VERY LOW achievers.

The comments to this entry are closed.