Association for Public Policy Analysis & Management

This presentation expands on a set of analyses from a larger study on college access and success in New York City (Black & Coca, 2017), a portion of which described the virtues and challenges of using a neighborhood-level measure of poverty to describe differential college-going trajectories for students from NYC public schools. In this presentation we deepen that work to describe the extent to which median neighborhood income overlaps with a more commonly used measure of poverty: free and reduced price lunch (FRL) eligibility. We also examine how this overlap has changed over time and what those changes might signal for college access and success in the city, as well as for researchers working in that space.

Our neighborhood measure uses the Median Household Income variable from the American Community Survey (2006-2010) pooled 5% sample, which is census tract-level data updated every five years. The variable is linearly interpolated, producing a yearly estimate of median household income by census tract. We find that although this neighborhood income measure obscures individual family differences to a certain degree, it is also far better at capturing a wider range of variation in college access and persistence than FRL eligibility, likely because 74 percent of the students in our sample were eligible for FRL. College enrollment rates of students in the poorest neighborhoods (bottom quartile of median income), for example, were more than 10 percentage points lower than they are for the broader group of students who were eligible for FRL.

From a research perspective, the aggregate measure of poverty has several advantages, including that it is on a continuous scale and is flexible to various forms of analysis. And, importantly, the measure is not dependent on families’ abilities to complete the paperwork for the FRL eligibility—a problem that has been reported anecdotally by our CBO partners and in with respect to Pell paperwork by our partners at CUNY. But the measure also has limitations. Perhaps most importantly, New York City students tend to live in poorer neighborhoods than the national median while, at the same time, the city has a much higher cost of living. Together, these facts mean that our highest quartile neighborhoods capture a broad swath of families—those who face real financial difficulties and those who are very wealthy—while our middle neighborhood income category may be quite a bit poorer than students from median-income families elsewhere in the country.

To extent that FRL eligibility obscures the impacts of economic disparities, and that neighborhood income obscures individual family differences, a combined index of poverty is particularly promising. We describe the additional power of this combined index by examining several cohorts of college-going outcomes from this perspective. We also describe the iterative and ongoing work with our partners, and the extent to which those collaborations have shaped our understanding of the complexities of measuring poverty in the New York City school context.