“Communities of Interest” and City Council Districting in New York, 2012-2013

The Center for Urban Research prepared a paper on behalf of the NYC Districting Commission discussing how "communities of interest" might be considered, measured, and understood in the context of drawing new City Council lines in the 2013 districting process.

The paper was included in the Commission's March 2013 submission to the US Department of Justice (see Exhibit 69 at that link) regarding the final proposed Council district lines. We have made it available here with the permission of the Districting Commission.

Download the full paper

The following text is excerpted and summarized from the paper. The text below recapitulates the paper's key findings and its discussion about how communities of interest can be visualized through available mapped demographic and socio-economic data. Omitted below for brevity are the paper's insights and analysis of how the idea of "communities of interest" can be contextualized and understood in relation to social science theory and the literature -- essential reading to fully appreciate the value of the findings and the maps below.


As it does each decade, a public conversation has taken place over the past year about the Constitutionally-required redrawing of federal, state, and city legislative boundaries. In New York City, the debate has focused on the new boundaries that the New York City Districting Commission has proposed for the city’s 51-member City Council. A key term of interest in this debate is the concept of “communities of interest” and how best to apply it. The purpose of the report is to:

  • provide a brief history of how social scientists have thought about urban communities and their relevance to Council redistricting,
  • look at developments related to the term in the theory and practice of political representation (including voting rights litigation and advocacy and practices by other jurisdictions) concerning how to define and apply the term in redistricting the New York City Council, and
  • utilize existing data to illustrate some different ways to empirically define the concept in New York City – and to highlight how data limitations constrain that process.

Key findings

Our paper provides an in-depth discussion of the first two bullet points listed above. Based on that discussion, the paper draws three basic lessons:

  1. First, it is clear that multiple, conflicting, and subjective definitions of any specific community in New York City as well as the overall mosaic of communities will coexist at any given point in time. Moreover, there is no “objective” way to adjudicate among these competing conceptions. This stems partly from the fact that different kinds of communities do in reality overlap with each other; partly from the fact that we have no systematic measurements of key dimensions of community (shared sense of belonging, social interactions and networks, community organizations); and partly from the fact that communities are always undergoing change.
  2. Second, the corollary is that any official designation or recognition of communities of interest is a value judgment made among alternative possible definitions for community -- it is, in other words, a perhaps somewhat arbitrary selection of what should count and what does not. Given that such choices are inherently political, empowering a wide range of neighborhood residents to articulate what they think defines their communities and taking their views into account during the districting process is a valuable way to solve the subjectivity problem. In other words, process may count as much as substance.
  3. Finally, however, we are not completely bereft of systematic data on the question of community. It may be unfortunate that the Census Bureau does not collect small area data on social ties, civic engagement, community organizations, or subjective sense of belonging, but it does collect data on a great many relevant correlates of community: racial and ethnic group membership; family form and stage in the life cycle; work, income, and reliance on public provision; housing tenure and type; and education and occupation, among others. These data may be supplemented with small area data on party registration and voting, housing values, neighborhood facilities and conditions, and even patterns of reported crime and health problems. While we cannot measure all important aspects of place-based communities, we can measure a great many. The availability of such systematically collected data provides an important baseline against which to compare claims of community that are based on unsystematic, subjective data.

Mapping key demographic and socioeconomic data

The Center for Urban Research has mapped and analyzed hundreds of measures that might be taken to indicate some aspect of community of interest. In our paper, for the sake of concise illustration, we restrict ourselves to four clusters:

  1. Race-ethnicity-nativity-religion;
  2. Income-tenure-family form-poverty-reliance on social services;
  3. Education-occupation; and
  4. Political engagement and orientation.

These overlap, of course. Different ethno-racial groups are clustered in different parts of the occupational structure with direct consequences for the education levels required and the incomes derived. They also have distinct levels of political engagement and orientation, reflecting their interests. Nonetheless each of these dimensions is to some degree analytically distinct.

The most common way to think about community is to distinguish first by race, then by ethnicity, nativity, and religion. The most basic racial categories include whites, blacks, Hispanics, and Asians (New York City has comparatively few native Americans). While the Census does not (yet) consider Hispanics to be a race -- and currently provides them with the option of choosing any race -- in practice, many Hispanics choose to categorize themselves as an “other” race. Demographers and the person on the street tend to unite in classifying all Hispanics as a distinct group and separating them from the other racial groups. However, these four groups are clearly too broad a classification. Each of these groups is characterized by important differences relating to ethnicity or national origin, nativity (i.e. native born or immigrant), and religion. Whites can therefore be broken down into a number of major components, including white Catholic groups (primarily Italians and Irish), Jews (further distinguished by degree of religiosity), and secular whites. Similarly, blacks may be broken down into African Americans, Afro Caribbeans, and Africans; Hispanics include Puerto Ricans, Dominicans, Mexicans, Colombians, Ecuadorans, and so on; and Asians include Chinese, Koreans, Indians, Bangladeshis, and Filipinos.

Of course, while black-white segregation remains quite high in New York City, and Hispanics and Asians also live at lower but still significant levels of spatial concentration, groups overlap and intermix and many neighborhoods in New York City have a fairly heterogeneous makeup. So neighborhoods and communities cannot be equated with racial-ethnic groups. Still, one or two groups tend to predominate in any given neighborhood.

Map 1 depicts the city’s racial-ethnic communities based first on which racial group is the majority or plurality, then which ancestry or language groups are the plurality or majority within that dominant group. For reference, it also shows the proposed city council boundaries (using the final lines adopted on February 6, 2013 by the NYC Districting Commission). While the Census does not record religion, we know that different ancestry and language groups tend to cluster in different religious affiliations. The first cluster, shown on Map 1, of whites with English or Scandinavian ancestries, for example, tend to be Protestant, while those who are Irish, Italian, or speak Polish, tend to be Catholic.

CUR communities of interest Map 1

click to enlarge map (PDF)

A second way of looking at New York City’s communities involves household form and household income. New York City is characterized by a high degree of income segregation as well as racial segregation. Strong correlations pertain between household income and household form. Multi-income married couples have the highest incomes while single-parent families have the lowest, with people in nonfamily households or living alone (of whom there are a great many in New York, especially in Manhattan), in between. The former tend to own their housing, while most of the latter rent, with the poorest occupying public housing.

Map 2 depicts the predominant household form across Census tracts, while Map 3 shows household income levels.

CUR Communities of interest Map 2

click to enlarge map (PDF)

CUR Communities of interest Map 3 income

click to enlarge map (PDF)

New Yorkers are also significantly stratified by educational attainment and occupation, which are in turn closely associated with earnings. While residential patterns based on these characteristics resemble the income patterns in Map 3, however, they are not identical. In other words, the correlation between education and income is not perfect: a fair number of highly educated people may not be in the upper reaches of the income distribution, while other individuals may have higher incomes than their education might suggest. The correspondence between education and occupation is stronger. Another important aspect is the sector in which the occupations are practiced, with the private sector generating more remuneration for a given level of higher education than do the nonprofit or public sectors.

Maps 4 and 5 depict several polar patterns: Map 4 depicts concentrations of highly educated people and professional occupations as well white collar and blue collar occupations. Map 5 shows health care workers, who tend to have intermediate levels of education.

CUR Communities of interest Map 4 occupational categories

click to enlarge map (PDF)

CUR Communities of interest Map 5 healthcare workers

click to enlarge map (PDF)

Finally, election results (mapped here by election districts as opposed to census tracts) and party registration figures are one of the few ways to determine political orientation. Maps 6, 7, and 8 show how the voting results broke down for the 2008 presidential election, the 2009 mayoral general election, and the difference between the mayoral and comptroller votes in 2009 (i.e. highlighting areas where a higher share of the people voted for the Democratic Comptroller candidate than for the Democratic mayoral candidate).

Map 6 shows the strongly Democratic orientation of New York City voters when it comes to state and national elections. The voter registration favors Democrats over Republicans by six to one in the city, with more voters declining to state a party than affiliating with the Republican party. Nevertheless, the McCain vote in 2008 shows that the middle class Catholic neighborhoods of Staten Island, South Brooklyn, and southern Queens are the most Republican, with some support as well on Manhattan’s Upper East Side.

CUR Communities of interest Map 6 percent voting for Obama

click to enlarge map (PDF)

Meanwhile, the other communities of the city lean strongly Democratic, particularly in black and Hispanic neighborhoods. The slight majority vote for Mayor Bloomberg in 2009, however, shows that New York City voters do not always follow their party line as much as they do in state and federal elections (see Map 7).

CUR communities of interest Map 7 voters of Thompson

click to enlarge map (PDF)

Finally, the difference between the two vote patterns is shown in the final map (Map 8), which highlights the difference between people voting for the Democratic candidate for Comptroller (John Liu), but the Republican candidate for Mayor (Michael Bloomberg - running on the Republic and Independence party lines). It suggests that the groups most likely to be counted among the ticket-splitters or “Democratic defectors” include the Jewish communities of the outer boroughs, for example Riverdale in the Bronx or Borough Park in Brooklyn, as well as the white middle income professional neighborhoods like the Upper West Side or Park Slope.

CUR communities of interest Map 8 ticket splitting

click to enlarge map (PDF)

The takeaway

None of these four ways of looking at communities of interest in New York City can be taken as dispositive (through the lenses of race-ethnicity-nativity-religion; income-tenure-family form-poverty-reliance on social services; education-occupation; and political engagement and orientation). While there are similarities across all four, they also crosscut each other to some degree. By putting all four ways on display, it is hoped that the reader will see the diversity of ways of understanding community in New York City as well as the distinctive contributions of each way of seeing community.

At the same time, we should not privilege patterns fashioned by these systematic descriptors over less systematic, more subjective ways of describing community. After all, the beauty of New York City lies in the fact that it is a beehive of separate small cells, each of which creates and sustains the social lives of their inhabitants. New York is a dense city in which many amenities and facilities can be reached on foot, offering or even requiring interaction with different kinds of people as well as birds of a feather along the way. It offers an amazing diversity of small worlds within one big one.