Using Crowd-Sourced Community Identities to Revisit Urban, Suburban, and Rural Definitions

It is not uncommon to hear people described in terms of the places they live – an urban shopper, a rural voter, a suburban dad. Implicit in these descriptions is the notion that where someone is from, and whether that area is urban or rural, has meaningful import for their personality, opinions, or behaviors. But what is a rural voter, and how is she different from an urban one? And where must a dad live in order to be considered suburban instead of urban? It might surprise some to hear that there is no broadly accepted definition of what urban, suburban, and rural mean in the United States. Different government agencies often offer different classifications of a single county because the criteria on which they rely to make designations is specific to their mandate. And these classifications often differ from how residents might think of where they live.

When trying to understand the connection between community type and public opinion or behavior, researchers might first reach for government definitions in order to capitalize on the credibility of administrative data. Often, however, it is more germane to the research to classify people according to how urban they think their community is – the connection between subjective perceptions of community type and subjective perceptions of other things is at least conceptually closer.

At Morning Consult, we interview around 5,000 Americans every single day. In addition to questions about political leanings and brand preferences, we also ask how people would characterize the area in which they live: is it urban, rural, or suburban? This massive scale of data collection has allowed us to generate a “crowd-sourced” definition of urban, suburban, and rural for nearly every county in the entire United States–a kind of people’s accounting of urbanity country-wide. In this post, we dive into how we did it, what urban/rural looks like according to Americans themselves, and how a crowd-sourced definition differs from (and conforms to) those provided by the administrative bodies on which we often rely.

Crowd-Sourcing Community Character with Rich Data

Most polling cannot use public opinion to classify geographies more granular than the state level because of insufficient sample sizes: states are too heterogeneous to meaningfully classify as “urban” or “rural” and too few people are interviewed to get a good read at a level like the county. Morning Consult is able to overcome this limitation because of the unique scale of our data collection. Using respondents’ geographic information and self-reports of living in an urban, suburban, or rural area from 2020 to 2023, we generate over 5 million observations covering 99% of U.S. counties. We only retain data for counties where we observe 50 or more respondents, in order to get a reasonable read on public perceptions (ultimately covering 90% of counties). The decision rule we use to classify whether a county is urban, suburban, or rural, is a simple plurality: whatever a plurality of respondents in a county say that county is, that is how we classify it.¹ Some counties are very obviously urban (no one in our data thinks the county covering Manhattan is anything but urban) and some are very obviously rural. But we also recorded the second-most-selected classification and the proportion of people that chose it to capture intra-county variations in urbanity and the (un)certainty around plurality designations. Table 1 is an example of three counties – how the plurality classified them, by what margin, and what the runner-up was.

King’s County covers Brooklyn, NY and has undisputable characteristics of being urban – high population, housing, and retail density. But it is also not homogeneous, and some sections could be considered more suburban (Midwood, for example, as discussed here by the New York Times). Administrative data doesn’t tend to capture this subjective subtlety – King’s County is categorically considered in these data as entirely urban. Next, Orange County in California is largely considered suburban, but also not uniformly – 35% of people say where they live in the county is urban. On the other end of the spectrum, Russell County in Kansas is overwhelmingly rural and very few people say otherwise.

Comparing Perceptions to Government Classifications

How do these subjective evaluations compare to administrative ones developed by the government? To answer this question, we compare our crowd-sourced classifications to those produced by two different agencies with oft-cited definitions: the U.S. Department of Agriculture and the National Center for Health Statistics (NCHS). The following sections explain these sources and compare their definitions with our crowd-sourced definitions.

Department of Agriculture

First, we take a look at the U.S. Department of Agriculture’s Rural-Urban Continuum Codes (RUCC) definitions. These are based on population density, urbanization, and proximity to a metro area. There are 9 categories here which we again recode into 3 for the purposes of comparison: we bin the top metro category, which includes counties close to metro areas with populations of 1 million residents or more into ‘urban,’ the counties near smaller metro areas into ‘suburban,’ and the non-metro counties into ‘rural.’² Figure 3 reports the concordance between how the RUCC classifies counties and how our respondents view them. The gap between what this agency considers urban and what people do is striking: only 7% of RUCC’s 440 urban counties are actually considered urban by our respondent residents. A slightly lower discrepancy exists when it comes to suburban counties, as 36% of respondents agree these counties are suburban. And there is nearly complete concordance when it comes to what is rural: 97% of respondents whom RUCC would have classified as rural residents indeed view their counties as being rural.

The differences that emerge at the urban and suburban level can in large part be explained by the fact that RUCC considers a county urban if it is in the same metro area as a high-population city. For instance, it classifies Howard County, MD as urban because of its geographic proximity to Baltimore even though large swaths of the county are farmland and towns with fewer than 500 residents. Indeed, administrative classifications like RUCC need to take into consideration commuting and patterns of economic exchange, which are well-captured by proximities to large cities, in order to be useful to the relevant agencies. But regular people likely focus more on population and housing density, as well as other visual markers, when deciding whether a place is urban or rural.

How does this all square with the distribution of counties and populations across these classifications? In Table 2, we report the proportion of counties our respondents classify as being urban, rural, and suburban, and the proportion of counties RUCC classifies as such. We also report what proportion of the U.S. population lives in each of category, by classification system (our crowd-sourced one vs. RUCC). Only 3% of U.S. counties would be considered urban if you polled their residents (column 1), while 78% would be considered rural. Meanwhile, RUCC designates 16% of counties as urban, and 59% rural  (column 3). A larger gap exists when it comes to considering population proportions, however. According to RUCC, fully 58% of the population is to be found in urban counties (column 2), but only 19% of Americans would consider themselves as living in an urban area (column 4).

Indeed, most Americans say they live in a suburban area, even though RUCC only considers roughly a third of the population to be suburban. Taken together, Figure 1 and Table 2 suggest that while there is some agreement between administrative and subjective definitions of urban and rural, that agreement is mostly over what is rural.

Note: U.S. population percentages derived from a collapsed version of the USDA Rural-Urban Continuum Classification (RUCC). The collapsing follows the logic described in the RUCC section below.

National Center for Health Statistics

But let us not rely only on this single administrative source. The National Center for Health Statistics (NCHS) also maintains a widely-used Urban-Rural Classification Scheme for Counties. The distinguishing feature of this data is that it classifies counties into six bins defined around their relationship to major cities; these are labeled: large central metropolitan area, large fringe metro, medium metro, small metro, micropolitan and non-core. We can think of large central metros as counties that are mostly, if not entirely, a major city, fringe metro as suburbs, and non-core as wholly rural. For the sake of comparison to our classification, we consider large central metros as urban, fringe, medium, and small metros as suburban, and micropolitan and non-core as rural. For each of these three (collapsed) NCHS designations, Figure 2 plots the percentage of our respondents that see the counties in those classifications as urban, suburban, and rural.

Overall, we observe better correspondence between the NCHS’ objective classification and our crowd-sourced one than we do with the RUCC classification, but some interesting divergences remain. Here again, it seems people’s definitions of what constitutes an urban area are narrower than the government’s: of the 68 counties the NCHS classifies as urban, only 44% are considered so by respondents in our survey – the remaining 56% are more likely to be perceived as suburban. On the other end of the spectrum, our respondents agree with the NCHS in the case of 97% of the counties the agency designates as rural. In the hazy middle, 44% of our respondents agree that the NCHS’ suburbs are indeed suburban, but a larger share would actually consider those counties rural. This distribution suggests that people indeed not only have a higher standard for what should be called ‘urban’ than the government, but that they also have a higher standard for urbanity in general.

Mapping Public Opinion

In order to get a sense of what our crowd-sourced urban/suburban/rural definitions look like, we map the plurality definition for each county in three East coast states–Pennsylvania, New Jersey, and New York– as an example in Figure 4. These charts offer a geographic visual of how a crowd-sourced view of urbanity aligns and differs from government definitions: in terms of land, most of this territory is considered rural by its residents, while only cities with prototypical metropolitan features (New York City, Jersey City, Philadelphia) are considered urban. As we move out from these metropolitan centers, the closest counties, along with certain other counties further out, are considered majoritarily suburban. But many of these suburban areas translate to urban areas under the NCHS definition, and an even greater number do under the RUCC definition. These maps clearly illustrate how the bar for urbanity moves as we move from resident-sourced views to the definitions of different government agencies. And the differing considerations that underpin the three definitions play out in the areas between the extremes: of course RUCC considers the areas the other two sources think of as suburban as urban given commuting to and from major cities is at the core of its classification strategy. At the same time, all definitions do converge on basic sensibility: major metropolitan centers are urban and sparsely populated farmland is rural.

There are three major takeaways from our crowd-sourced compendium of urbanity and the comparisons to administrative measurements:

1. People have a high bar for what they think constitutes an urban area. Although government agencies might consider any place close to a major city ‘urban,’ people don’t necessarily see it that way. Indeed, the 93 counties considered urban by Americans are those at the most extreme ends of urbanization: the ones containing Chicago, Dallas, New York, and L.A., and other major American cities.
2. But it seems everyone–administrators and regular people alike–can agree on what is rural. In America, rurality might truly be a case of “you know it when you see it.”
3. Overall, public opinion tracks directionally with government data. The differences seem to be in degree, not kind, and explicable by factors that government agencies focus on that people are less likely to find germane, such as commuting patterns.
4. These findings are in line with similar conclusions drawn by the Pew Research Center in the past, contributing to the body of evidence that demonstrates the unique contribution public opinion brings to understanding community types.

Some projects are benefitted by relying on administrative characterizations of communities because they are concerned with the same factors that underpin government definitions, like labor exchanges between centers and peripheries. Yet others, especially those concerned with public opinion and relating residence types to attitudes, would be well-served to take heed of people’s own perceptions of where they live. Morning Consult is uniquely positioned to provide insight into public views of urbanity and rurality by virtue of the unprecedented scale on which we collect data. Interviewing thousands of people daily across the country means we are able to capture the contours of community perceptions at the county level in a way that smaller scale efforts cannot.

Notes

¹ Our survey question is phrased: “Would you consider where you live to be: 1) Urban, 2) Suburban, 3) Rural” and people are thus not necessarily thinking about the character of the county they live in but rather their zip code, or even their block. Thus, aggregating up to the county level aggregates over potential perceptual disparities within a smaller geography (e.g., a ZIP code) and over the objectively disparate geographies that sometimes coexist within a county (e.g., a county that encompasses an urban center and a large suburban sector). While we weight our survey data to match Census population values on age, education, race, and gender and their interactions, we do not weight individuals’ within counties according to the proportion of the county for which their ZIP code accounts in this analysis.

² We use the decision rule the Pew Research Center applied in their own comparison of public opinion data and administrative classifications of urbanity.