Interactive Map: NJ Racial Segregation & Student-Teacher Gaps

A previous analysis explored two ideas related to race in New Jersey. One, there is a much higher proportion of white teachers than white students in NJ, which means that students of color are unlikely to be taught by teachers of color. And two, the size of racial gaps between teachers and students varies greatly across schools. A small but growing body of evidence finds that mismatches and over matches are harmful (see the earlier analysis for some of that evidence).

Understanding the causes and correlates of the racial dispersion in schools may point to solutions to big gaps and over-matching. One relevant factor is racial dispersion in residence. Three sets of variables are rarely, if ever, explored together: student racial dispersion in schools, teacher racial dispersion in schools, and racial dispersion in housing. The interactive map below permits an exploration of these three variables and can be used to develop informed ideas about how space and location are connected to teacher-student racial gaps.

The map can be used to find a relevant school, maybe one you attended, one in your neighborhood or a school near where you work. Then, racial gaps between students and teachers can be explored. The data provide preliminary answers to questions like, are white students likely to be taught by teachers of color? Next, you can see how school racial makeup relates to residential racial patterns by examining the dispersion of the dots, which represent the number of residents. Finally, you can click on “show a mostly white school” to cycle through more than 100 schools that have greater than 90% white students and white teachers to explore schools and communities of whiteness. (The emphasis on white schools is explained more below.)

If neighborhoods were more integrated, you'd be less likely to find extreme teacher-student racial matches. For example, there were more than 450 NJ schools with all white teachers. Many of those schools had mostly white enrollment. These "schools of whiteness" often existed in "communities of whiteness." Take Monmouth Beach Elementary School in Monmouth Beach Boro. The 26 teachers in the school are white, as are all but one of 240 students. Intense racial segregation in so many schools may produce "cognitive deficits" in white children throughout New Jersey.

Elite students in segregated schools may develop what Elizabeth Anderson labels "cognitive deficits among elites." Anderson deftly describes "how segregation and group stereotypes generate knowledge deficits on the part of the advantaged, and especially within elites, and how their ignorance and incompetence put others at a disadvantage." White people in New Jersey who live in segregated neighborhoods and attend segregated schools are deprived of knowledge. Anderson writes, "To the extent that [elites] lead lives that are isolated from the lives of the disadvantaged and personally know few disadvantaged people, they are liable to be relatively ignorant of the problems the latter face in their lives and of the constraints within which the latter must cope with their problems." Children are also likely to develop harmful stereotypes, and distortions of reality, when they grow up in segregated spaces.

The white students in Monmouth Beach Elementary have been deprived of learning from students and teachers of color. Students who attend schools with more diverse staff seem more likely to develop diverse knowledge. Spatial integration of neighborhoods and social integration in institutions like schools may address the cognitive deficits among elite white students.

There is a bi-directional relationship between residential segregation and school segregation. Schools reflect residential segregation and residential segregation reflects policies and preferences about where families choose to, or are able to, send their children to school. Changing both residential and school segregation would likely effect the diversity of teachers in schools.

One interesting thing to note in the map is that although residential segregation is strong, the racial separation in schools is not merely a function of geographic separation. Put another way, school district boundaries are drawn in ways that reflect the racial preferences of folks in power. For example, although schools in southern Newark are mostly black, there are many white children who live just on the other side of the school district border in Union and Hillside. In cases like this, children would not have to travel far to create integrated schools. Although some people believe that it is "natural" that school segregation follow residential racial segregation, school district boundaries must be drawn in ways that express the racial preferences of policymakers. Black children in the south ward of Newark could attend more racially diverse schools merely by redrawing district boundaries -- and they wouldn't have to travel very far.

Background

I wanted to build a map that that could place student-teacher racial gaps in the context of racial segregation. The issue of racial mismatch between teachers and students seems connected to the racialized geographic dispersion of students and teachers. This map tries to make visible a connection between those two issues.

I also wanted to make it interactive to make personalized pictures of the schools and neighborhoods in which folks are interested. But don't focus only on a particular school! Racial segregation is both a within-city and across-the-state issue...so understanding the problems and coming up with good solutions needs to take into account both local and statewide issues.

Technical Details

The standard racial map tends to be the choropleth style. These maps will typically fill an entire region a single color corresponding to the proportion of a single race. In New Jersey, this can often be misleading because population density varies by urbanicity. The cities have far denser populations and far more people of color than other areas and a choropleth map often doesn't depict that well. A single color for a geographic unit in a choropleth would make it seem like the state was more white than it is.

The app was built with R, utilizing the leaflet package and Shiny.

Plotting all dots at once causes browsers to slow to crawl, if not crash because of lack of memory. I tried to find a nice balance between number of dots, graphic detail, and crash survival.

In addition, rounding occurs during dot plotting, so the visualization should be considered an estimate with approximate precision. The dots should give a fairly accurate representation of racial distribution in any given area you zoom to on the map.

A note on zoom levels: it is best to view high density (urban) areas at a closer zoom level to avoid the order in which dots are plotted having an influence in the color blend.

Data Sources and Caveats

Residential race data is from the US Census. They are 5-year tract-level estimates, 2011-2015.

The geographic shapefile for schools is from the State of NJ.

School teacher and student race data were downloaded from the NJDoE's website. They were then merged (along with the shapefile). Note that there was not a perfect merge (some schools were unfortunately dropped due to non-matching ids) and the schools that were dropped had higher student-teacher mismatch than the ones that were matched. This "not missing at random" error suggests a kind of bias in the composite dataset, but this doesn't pose a problem if looking at any single school.

Charters

The State of New Jersey seems confused about how to count the number of charter schools and this makes some analyses impossible. In the student enrollment data files provided by the State, there is, for example, a single school called "North Star Academy Charter School." However, in the geo-coded data files also provided by the state there are 12 entries associated with North Star. For example, there is "North Star Academy Cs Of Newark-Downtown Middle School" located at 10 Washington Place and "North Star Academy Cs Of Newark-Clinton Hill Middle School" located at 600 Clinton Ave. In an unsatisfactory attempt to solve this basic shortcoming with the State's data, I chose to display the aggregate student/teacher race data for the charter network at each individual school on the map. From the standpoint of analysis and transparency, the State could mandate individual reporting for charter schools, the same as is required for regular public schools.

Private Schools

Private schools, according to a national report, had higher levels of black-white segregation than public schools. However, the State of NJ does not provide student or teacher race information for private schools in the files used in this analysis. Excluding private school teacher-student race data from these kinds of analyses thus understates the segregation in New Jersey.

Credits

Inspiration for the layout and more was taken from several sources, including Culture of Insight and the NYTimes racial dot map, one version of which can be found here.