Analyzing The Connection Between School Quality and The Neighborhoods They Serve

Research

School Quality and Demographics: Charter vs. District

In this Leading Indicator, we examine the relationship between schools and the neighborhoods where they’re located. For this analysis, we matched Philadelphia School District school-level data to ACS zip-code level data. Each school can then be viewed in the context of its neighborhood characteristics. The school-level data includes both student demographics and school metrics of success, while the ACS estimates detail zipcode-level demographics and median home values. 

What You Need To Know

Charter schools enroll 28% of Philadelphia’s students and most accept students from across the city; most district schools draw from a delineated catchment zone, though there are a handful of citywide special-admit schools as well as mechanisms available for families to apply for a seat in an out-of-catchment school.
Charter schools have a median Black student population of 89%, while district schools have a median Black student population of 53%.
Compared to the zipcodes where they are located, both district and charter schools tend to have student bodies that are overrepresentative of the Black population and underrepresentative of the white and Asian population.
District schools in higher-income neighborhoods have higher student attendance and higher teacher retention rates than district schools in lower-income neighborhoods.

Students of Charter vs. District Schools

As of 2021, publicly-supported but privately-operated charter schools enrolled 28% of Philadelphia school students. Charter schools differ substantially from district schools in their admissions policies in that only 26% of charter schools use neighborhood-based admission preference, while 82% of district schools use neighborhood-based admissions. Nearly three-quarters of charter schools accept students from across the city, and some use additional criteria. Perhaps because charter schools are more likely to attract students from across the city, their student body demographics differ from district schools.

Figure 1 illustrates the distribution of demographic representation in charter and district schools. Each box plot indicates how the distribution of data points would be split into quarters – the minimum, first quartile, median, third quartile, and maximum. The rectangle indicates where 50% of the distribution falls, the interior horizontal line represents the median (middle of the distribution), while the extending vertical lines show the remaining quarters on either end. On some boxplots, there are points trailing beyond the minimum or maximum lines. These are suspected outliers, points that are so far from the first and third quartiles that they don’t seem to fit with the rest of the distribution.

For example, the green plot on the left represents charter schools’ Black student populations. Amongst all Philadelphia charter schools, student populations range from a maximum of 98% Black to a minimum of 2% Black. A quarter of schools have student populations that are between 2%-34% Black, and a quarter of schools have a student population that is 94%-98% Black. The median of 89% indicates that half of charter schools have a student population at least 89% Black.

Overlaid on top of each boxplot is a black horizontal line representing the citywide population. Philadelphia’s total population is 40% Black, 34% white, 15% Hispanic, and 7% Asian.  illustrating how both categories of school tend to be overrepresentative of the city’s Black population and underrepresentative of all other groups. The city has a 23% poverty rate, compared to the 86% median rate in charter schools and the 83% median rate in district schools. In schools, the economically disadvantaged rate captures the percentage of students whose families participate in state and/or federal public assistance programs.

Looking at the boxplots for charter and district schools, several trends appear. First, charter school populations are disproportionately Black compared to district schools. Charter schools have a median Black population of 89%, compared to district schools’ median Black population of 53%. The distribution of Hispanic, Asian, and white populations is correspondingly low at charter schools, compared to the district schools. While critics claim that charter schools are exacerbating racial segregation in schools [1], proponents argue that they simply give families a choice, and many Black families take advantage of the opportunity. 

Figure 1:

 

There are few Philadelphia schools that are representative of the city’s population as a whole. Out of the 299 total schools in our analysis, 29 schools have a student body that is at least 20% Black and 20% white, while just 7 schools have a student body that is at least 30% Black and at least 30% white. Only 22 schools have a student body that is at least 20% Black, 20% white, and 10% Hispanic.

Figure 2 illustrates the difference between a school’s student population and the demographics of the surrounding zipcode. Here, we show the median difference, aggregated for all district schools and all charter schools. Since this visual only indicates the median measure, individual schools may be more or less representative of their neighborhoods than the numbers shown. Based on this measure, both categories of schools tend to be overrepresentative of the Black population and underrepresentative of the white and Asian populations of their local zipcodes.

The difference between charter and district schools is most pronounced for the Black population. The median charter school has a Black student population that is 15.9 percentage points greater than the zipcode population, while the median district school has Black student representation 10.9 percentage points above the local demographics.

While both district and charter schools have fewer white students and fewer Asian students than would be representative of their locality, they differ in their representation of Hispanic students. District schools tend to have a higher Hispanic representation than their local zipcodes, while charter schools tend to have a lower Hispanic population than their neighborhoods.

Figure 2:

Across school types, the student population tends to look quite different than the ZIP code where the school is located. Although there’s no publicly available data illustrating where students of each neighborhood opt to go to school, this suggests that students do move out of their neighborhoods of residence to go to school. In particular, the striking trends in the visual suggest that the white and Asian populations are less likely to attend Philadelphia public schools at all – either district or charter – making schools overrepresentative of other groups.

Median Home Values and School Metrics

Next, we look at several metrics of school quality from the 2021-22 school year in district schools only. Although charter schools are included in the school quality reporting, not all metrics are available. The student and teacher attendance metrics refer to the percentage of students or teachers who are present for at least 90% of school days. The teacher retention metric is the percentage of teachers who stayed at the school from the prior school year. Each of these three measures provide some proxy for the quality of instruction that students are receiving. While standardized tests can speak to a school’s overall performance, these chosen metrics speak to the student experience, as higher attendance and higher teacher retention are associated with more effective instruction.

In Figure 3, we separated schools based on the median home value in their neighborhoods. Schools located in zipcodes with a median home value at or above the city median are shown on top, while schools in zipcodes with median home values below the city median are below. There is a notable difference between the chosen metrics between the two groups, indicating that there is some correlation between the neighborhood where a school is located and its ability to deliver instruction.

The metric with the most significant difference between the groups is student attendance. In neighborhoods with lower-valued homes, three quarters of schools have 65% or fewer of students meeting the attendance metric. For schools in zipcodes with higher-valued homes, three quarters of these schools have 80% or fewer of students meeting the attendance bar.

The distribution of schools meeting the teacher attendance metric has some variation between the two groups, but only by several percentage points, a median of 83% or 86% across schools. However, the difference in teacher retention rates is greater, with a median of 84% amongst schools in higher-income neighborhoods and a median of 78% in schools in lower-income neighborhoods.

Figure 3:

Takeaways

These data illustrate a well-known phenomenon in which public schools in wealthier neighborhoods tend to be higher performing. In Philadelphia, as in many American cities, they also tend to have a higher white population than schools in lower-income neighborhoods. Yet, Black students usually make up the largest percentage of the student body in most Philadelphia schools, regardless of neighborhood. The lower performance of schools in lower-income neighborhoods may be one factor contributing to Black families opting into charter schools at a higher rate. 

Sources

[1] https://crpe.org/we-need-a-more-productive-conversation-about-charter-schools-and-segregation/