A Response to "We all want what’s best for our kids"

Brookings recently released a report on DC Public Schools. This report, "We all want what’s best for our kids", uses data from 10 years worth of posts in the DC Public and Public Charter Schools forum. My wife and I own and operate the DC Urban Moms and Dads website and are frequent participants in the DC Schools forum. Moreover, we have two sons who both attended a DC Public Charter School. One attended and one still attends DC Public Schools. Therefore, we know well both the DC public school system and the DC Schools forum.

Having read the report in detail, I believe that the research for this report was lazy, the analysis is flawed, and the conclusions are wrong. The entire report is based on flawed analysis -- word frequency analysis in which keywords are connected to school names -- a technique that does not take context into account. In fact, the report's examples show such context-based errors. The measures of school attention suffer from uncorrected bias due to school size and the uneven geographic sample representation. For instance, Alice Deal Middle School and Woodrow Wilson High School are two of the largest schools in the DCPS system and the in-bounds school for a large percentage of the posters. The finding that they are the two most discussed schools is exactly what would be expected.

Moreover, beyond the data and analysis issues, the key conclusions of the report are obvious and simplistic. The report's first and most important conclusion is that school selection is influenced by real estate prices and neighborhood segregation. Yet the data in the report is hardly required to make this conclusion. In fact, I hardly know anyone familiar with the DCPS system that does not agree with this statement. Because the report merely restates a problem that is widely acknowledged, even without discussing the quality of its data the report does not advance our thinking about the problem, its sources, or ways to fix it.

The most glaring failure of the report is its lack of recognition that practically all participants in the DC Schools forum who have children in DC public schools have chosen integrated schools. The report states that Alice Deal Middle School and Woodrow Wilson High School are the two most commonly discussed schools in the forum. Deal's demographics are 46% White, 28%, Black, and 16% Hispanic. Wilson's are 39% White, 29% Black, 22% Hispanic. There is no reasonable definition of "integrated" that does not describe these schools. The report does not mention the biggest way students segregate by socio-economic status, which is moving to the suburbs. If higher-SES parents want to avoid integrated schools, only a few scribbles in their checkbooks are needed for them to relocate to within the boundaries of Bethesda's Walt Whitman High with its 67% White, 4% Black student body. The report does not acknowledge that higher-SES families still in DCPS are still in the DC public school system, and often trying to do their best to help improve DCPS schools instead of leaving for often-much-more-segregated suburban schools.

The report repeatedly attributes to racial demographics what can easily be explained by other more obvious factors. The DC Schools forum (and to an extent the entire website) has always been most popular with parents in northwest DC and in parts of Ward 6. Therefore, there should be no surprise that schools in those areas get discussed the most. The report seems to miss this rather obvious fact that the amount of attention paid to schools is highly reflective of the self-selected user base. Users discuss what they know. The large number of users in northwest DC and Ward 6 means that posters in those areas can expect useful responses to questions about their neighborhood schools. Forum users in parts of the city in which we have fewer users may find little to no response to their queries simply because few of our users know anything about those neighborhood schools. The lack of discussion is a result of unfamiliarity, not the schools' demographics. Those outside of areas in which DC Urban Moms and Dads is popular may prefer to discuss their schools on neighborhood mailing lists and Facebook groups, also decreasing mentions of the schools in the forum for a reason unrelated to student demographics.

In another example, the report categorizes Woodrow Wilson as a "high-attention" school compared to Duke Ellington School of the Arts which is categorized as "low-attention". The report attributes the disparity of attention to demographics, ignoring that Wilson is the city's largest high school with 1,872 students while Ellington is a specialized arts school with 558 students. Of course Wilson would get more attention.

Moreover, the mere mention of a school's name is no real indication that the school is being promoted or favored. A thread about "Lack Of Eye Candy At Janney" (referring to dads) is 34 pages long. I'm not sure that helped increase interest in the school. Moreover, some of the "high-attention" schools gained high-attention because of scandals. The threads more likely scared people away from the schools rather than attracting them.

A similar misrepresentation of data occurs when the report discusses schools in the Brookland neighborhood. The report provides a graph showing which schools in Brookland get mentioned the most in the forum, concluding that schools with significant White demographics get discussed more than schools with fewer White students. What is missing from this conclusion is the fact that the most discussed schools are all charter schools which draw from the entire District for students while the less-discussed schools are all in-boundary and rely on their neighborhoods for students. As such, the pool of forum posters potentially interested in discussing the charter schools is much larger. This is simply not a valid comparison. To be accurate, the authors should have limited their study to only posters living within Brookland, but the data available to them did not allow this.

The basis of the report is word frequency analysis in which the presence or absence of a keyword is used to determine interest. This is a very primitive methodology replete with shortcomings. This practice removes all context from the discussion. For instance, the word "diversity" is listed as a popular keyword. Writing "no diversity" would also add to the count. But it is unclear what mentioning "diversity" means. If "diversity" is identified as a strength of a school, does that support segregation? The authors include "IB" as a synonym for "in-bound", apparently unaware that it is also used for "International Baccalaureate" and comes up often in reference to several popular schools. Further, the authors identify "in-bound" as the most common substantive term they analyzed. Significantly, the District's heavily discussed rezoning process occurred right in the middle of their dataset. Every DCPS-related publication was discussing "in-bound".

The lack of context shows itself in another example. In a paragraph beginning "Many school assignments are deemed unacceptable outcomes to DC Urban Moms participants; it is common to consider opting out of the District's school system entirely..." The report quotes a post, "Agree. I would apply [to] Janney and pay for private if I didn't get in." On the face of it, this post seems to support the authors' contention. But, if one actually reads the thread in which this is posted, the original poster is inbounds for Janney but currently has a child in a private preschool. The poster's question is what to do about pre-k 4 if she doesn't get a spot in the lottery. The recommendation that is quoted is not about leaving DCPS, but about not making school changes two years in a row (the OP's child is guaranteed a K spot at Janney). The thread almost entirely contradicts the premise it is being used to support. Far from threatening to leave DCPS, the poster is eager to get in to it. The authors spent four years on this study, but don't appear to have bothered reading the threads on which they were basing their conclusions.

The report is clear in its conclusion that the DC Schools forum conversations support school segregation. But the same conversations can be viewed differently. Privileged families who are unhappy with their inbound DCPS school have a number of choices. They can move, either within DC or to another jurisdiction. They can utilize the lottery and hope for acceptance to another DCPS or charter school, or they can choose a private school. By providing information, access to sometimes obscure data, advice on how to use the lottery and so on, the forum makes it easier and more likely for parents to choose a DC public school. The number of White students in DC's public schools is increasing. Not all of these are choosing predominately White schools, but rather are contributing to the integration of additional schools. A more serious and objective report would acknowledge how the DC Schools forum has contributed to desegregation. This is the irony of the report: it criticizes families who choose to remain in DC public schools and the forum that helps them do it.

DC Urban Moms and Dads posters are often optimistic about an integrated school experience, something in which most of them are gladly participating. Like all parents, they are attempting to find the best educational opportunities for their children. They are working within a system they did not invent and over which they have little control. The realities of race, class, and culture in DC public schools are nuanced and complex. They are deserving of serious study. A shoddy word frequency analysis completely devoid of context is not such a study. It does a disservice to the forum posters, to the authors who wrote it, and to Brookings who published it.