Where Is Real America?
Oklahoma City is America. San Jose is Not.
Update: So, this isn’t related to the post, but so many of you are reading this, I felt I had to mention that I have a podcast! It’s about the history of migration in America, and my mom tells me it’s very good. Check it out!
So Jed Kolko, of whose work I’m a big fan, had a piece last week at 538 arguing that, based on a simple estimate of demographic abnormality, the most “normal” place in America was New Haven, Connecticut, specifically claiming that white, small-town America was not “normal.”
It’s a neat claim, if true. But I have serious beef with Jed’s method. See, he used 3 variables: race, education, and age, to proxy for “normalcy.” His method looked at how typical a given “race” group in a given city was on educational/age factors, and a given educational group in a given city on race/age factors, etc. In other words, he didn’t truly ask “What city is most normal?” He asked “In what city is each group of people most typical of that group of people nationally?” That’s a cool question, but it’s totally not “normalcy.” The reason is simple: as best I can tell, Jed doesn’t fully capture the role of aggregate composition. He’s trying to get specific and avoid calling a place “abnormal” just because it has one weird demographic lump; he wants cell-specific abnormality. But nobody cares if Graduate-Degree-Holding Native Americans happen to be much younger in St. Louis than elsewhere. We care if St. Louis has a weirdly large number of Graduate-Degree-Holding-Native-Americans. Composition of the population is the most important measure of normalcy, and one that Kolko’s method will tend to under-emphasize.
Plus, just three variables? I argued on Twitter he should at least have included proxies for the foreign-born population, geographic area, and the rurality/density of a region. He responded by politely encouraging me to get off my fat, lazy bum and do the work myself. Challenge accepted!
I used 20 variables from the American Community Survey’s 2014 5-year sample. They are:
- Foreign-Born Share- It seems to me a key demographic that makes a place “normal America” is how many people there aren’t born as Americans. I define places as “weirder” if they have a foreign-born population much larger or much smaller than is typical. So you can be “weird” by having too many immigrants, or too few.
- Married Share of the Population- Family and social structure is a vital part of American life. Yes, a huge part of this is driven by age factors, but that doesn’t matter too much, because regardless of age, being married is fundamentally different from being single.
- Labor Force Participation- The number of people who are totally out of the labor force is a valuable indicator of economic activity and the structure of socio-economic life.
- Percent of Employed People in the Armed Forces- The military is a vital part of American life, and a monumental social institution that radically alters the lives of those who are within it. So an estimate of the prevalence of the military is a huge part of describing an area’s “normalcy.”
- Percent of Employed People in Civilian Government Work- Areas with government hubs are weird. Whether state or national centers, these areas are culturally unlike their surrounds, and so are worth highlighting. Similarly, areas of governance-deficit, places where for whatever reason the footprint of government is very light, are worth pointing out as oddities.
- Percent Self-Employed- Self-employed people are odd. Their lives are different. Their schedules are different. So if there are very few or very many of them, it makes a place less “normal.”
- Percent White Collar- White collar work makes up a large share of employment generally, but its deviation from the norm is worth highlighting both because many white collar industries are highly concentrated, but also because they are associated with the elitism that drives this whole debate about “real” America.
- Percent Agricultural & Natural Resources- But while the “real” America stereotype is rural, I do intend to punish places with weirdly high employment shares of ag or mining work. Though, of course, I also punish places with weirdly low shares.
- Median Male Annual Earnings- Very low wages and very high wages are alike penalized.
- Average Number of Workers Per Household- Areas with very large numbers of workers per household or very small I will alike penalize, as this is a nifty indicator of the role of work in family life.
- Family Poverty Rate- How many people are actually in poverty? Aside from wages, I want to highlight places where poverty is unusually common or unusually rare.
- Percent of Households Renter-Occupied- Lots of renters or few renters, either is penalized, because housing tenure says a lot about lifestyle.
- Average Owner-Occupied Household Size- How many kids ya got? Few things drive lifestyle more than kids. Baby happy Utah? You’re weird! Childless New England? You’re weird!
- Median Number of Rooms Per House- McMansions? Weird! Studios? Weird! Ranch-style 2-bedroom? Just right.
- Percent of Households With a Car- Literally everyone has a car? Weird! Nobody drives because we all ride bikes and jetpacks to work? Weird!
- Percent of Households With No Phone
- Percent of Owner Households Where Housing Costs Are Under 30% of Income
- Percent of Metro Area Population Residing in Rural Settings
- Percent of Population With Bachelor’s or Higher Degree
- White % of the Population
Note: I didn’t do age. That’s because most age measures were highly correlated with several of the above-listed items, so including age didn’t change that much, and because how you define age matters a lot. I also didn’t do a more detailed dive on race. Both age and race have big categorization issues. Do I penalize a place for its oddness on each racial group? How specific do you get? And at some point, you get colinearity, and they start cancelling out. Do I use a diversity index for each city, like HHI? But again, what level category do I use? What about foreign-born whites? Ultimately, I decided the white share was sufficient, as the debate about “real America” has a large subtext about whether “real America” means “white America.” I penalized areas with abnormally high white shares, or abnormally low white shares.
Also, a bit about my method. I took the absolute Z score for each city for each category, where the mean was the mean of metro areas, not the national average. This was partly for simplicity of calculation, and partly because I decided I wanted to restrict myself to assessing which metro areas where more or less normal: it’s possible that all metro areas are a bit weird compared to non-metro areas, and it’s possible that some weird metro areas are also large, so if I population-weighted, things would cancel out a bit.
However, in the debate about “real America,” location matters as much as population. No matter how many people live in New York City, it will always be New York City, which is different from Upstate, and from the nation. Locations are important in their own right for these kinds of comparisons, partly because locations are political units in many cases, and partly because locations are actually what we’re comparing. So weighting by population is, in a sense, unfair: it means weird New York gets to stack the deck in its favor, even if the vast majority of locations being compared are different.
Now of course, if the result of this was that the vast majority of “normal” cities were small while the “weird” cities were all big, I’d be worried. But that’s not the result I get. I get big weird cities, big normal cities, small weird cities, small normal cities… so it doesn’t strike me that population weighting should radically change my results.
Finally, I do weight a few results. The college degree and white collar components are highly correlated, so I under-weight them each a bit. Poverty and telephone access are also correlated, so I under-weight them. Household size and rooms per house are also somewhat correlated, so I under-weight them. Broadly speaking, even large changes to the weighting of 2 or 3 variables has fairly little effect on where a city falls in the rankings.
So, without further ado, the results.
The Weirdest Cities in America
With Sincere Apologies to Austin, Texas
The chart below provides my estimate of the 10 weirdest cities in America, with my “weirdness score” (which, again, is the weighted sum of the absolute values of the deviation from the mean of metro areas for each of my 20 criteria).
San Jose tops the list as the weirdest city in the nation. This is driven by a very high foreign-born share, high white collar and educated shares, high annual earnings, high workers-per-household, a very low white share, and a low rural population. In other words, Silicon Valley makes San Jose very weird.
New York is up next. Again, a large foreign-born share makes New York weird. But the real weirdness is actually in New York’s transit access. New York’s car-ownership share is a whopping nine standard deviations below the national average. New York’s housing costs also make it weird, as does the percent of people who are renting. In other words, New York is weird because it’s just so darn urban.
Next up, Jacksonville, North Carolina. Jacksonville is weird for basically 1 reason and 1 reason alone. It’s actually pretty normal on most variables, except it has slightly more renters, slightly lower earnings, and slightly lower civilian government workers. But Jacksonville has 10 standard deviations higher employment in the Armed Services than most metros. Because, duh, there’s a military base in Jacksonville, which makes it kinda weird.
You can see the others. And I’ll have a link to the full Z-score tables at the bottom of the post. But for now, let’s look at “real America.”
The Most Normal Place in America
Are You Surprised It’s In Oklahoma?
The above table shows the places with the lowest weirdness-scores. Two of them are in Oklahoma. I’ll talk about them together. Oklahoma City is less than 1 standard deviation from the mean on every single variable. It is exactly the mean for the poverty rate, and almost exactly the mean for educational attainment. It’s biggest oddity is housing costs compared to income, which are a bit high, and the percent of households with a car, which is also just a teentsy bit high. Other than that? If you’re looking for “Normal America” then look to Oklahoma City. Tulsa’s story is the same, except it also has a bit of a low share of civilian government workers.
Next we come to Jacksonville, Florida. Confusing, right? I thought I had a spreadsheet error at first. Jacksonville’s oddities, such as they are, are a slightly lower than usual share of self-employed people, as well as oddly low housing expenses and a smaller than usual rural population. So Jacksonville sounds like your suburban example-city.
Finally, we get to Spokane. Now in fairness, Spokane is actually whiter than you’d expect. It also has fewer workers per household, and lower labor force participation generally, than you’d expect, and slightly bigger houses. But overall, this bastion of inland Washington State is a pretty good example of “normal” America.
Two Kinds of Normal
Comparing Jed’s List and Mine
The chart above shows Jed’s score for his top 10 most-representative cities, versus my score for those cities. For reference, the average score in my same was about 13 or 14, so scoring under 10 should be seen as a “very normal city.”
All in all, there’s some substantial agreement. We both agree that Tampa, Oklahoma City, Springfield MA, Wichita, and Kansas City are good examples of “normal” America. I’m less confident in Hartford or Milwaukee, but they’re probably not bad examples.
But for New Haven, Chicago, and Philadelphia, I find that some of Jed’s “most normal” cities are, by my measure weirder than average. So why the difference?
Well, I mark Philadelphia as weird because it’s got a fairly high white collar share and more rooms per house than is typical, but the biggest factor is car ownership: Philadelphia is nearly 3 standard deviations below normal in terms of car ownership. Jed has nothing telling him whether people in Philadelphia have a lifestyle that looks like typical Americans. I do. And typical Americans are much more likely to own cars than typical Philadelphians.
Next we’ve got Chicago. Chicago has a higher foreign-born share than you’d expect, and fewer self-employed people, and higher earnings. But again, the car-ownership thing hits Chicago, as they’re 2 standard deviations below average. High housing costs also hit Chicagoans.
Finally, New Haven, Jed’s most normal American city. I disagree. Once again car ownership dings New Haven, but the married share also hurts them, as do higher than average earnings, alongside high housing costs, and also a smaller rural population.
By and large, Jed’s method makes a certain type of unusual place look more normal: high-cost, transit-dense urban areas. But those places aren’t normal. Car ownership is super-dominant among American households. Foreign-born people are less common nationally than in urban centers. Wages and costs are lower as well, yielding nominally smaller disposable income. While Jed was trying to make a very precise baseline estimate, his “normal” isn’t really what “normal” means to most people, which is, “this place looks about average on key lifestyle and demographic metrics.” He picked a few neat demographic metrics, but forgot that the politics of “normal” are the politics of lifestyle: questions of commuting, family size, norms about work, and geography are all inseparable from “normalcy.”
UPDATE: I got some feedback. Some readers suggested I should drop the car variable, as it seemed to extreme. When I drop it, results don’t change very much at all, except for NYC, which does get a bit less weird, but remains top-30. Other readers suggested that I should simply weight all variables by their standard deviation, weighting variables with large variability more than those with small variability (like car ownership). I did this, and it does not change results much at all. San Jose is #1 weirdest in all specifications, OKC is #1 most normal. Under the variability-weighted method, NYC is #4 weirdest city. It was also suggested I use religious demographics too. I have not done so yet, but if I get time may try and see what inclusion of a religious variable does to weirdness/normalcy.
Cities like San Jose, New York, or Philadelphia are not “normal.” And when people talk about “real America,” if what they have in mind is something like Oklahoma City, well then, they’re right. But weirdness isn’t confined to big urban hubs: mining towns, military bases, farming communities, self-employed meccas, immigrant cities, and poor areas are all “weird” in different ways. Which gets to the basic question here: what does it even mean to be a normal American? Now if I could wax poetic here for a moment, I can’t help thinking that New York is no less “real America” than Oklahoma City, even though New Yorkers are (objectively) a bunch of bitter, angry, impolite, weird, good-for-nothing Blue-stater Yankees. But that’s all American as apple pie. And there are ways Oklahoma City is weird: if I’d added a variable for Native American populations, it’d drop significantly. In migration circles we get debates between people who favor melting-pot integration and salad-bowl integration. Personally, I’m more of a beef-stew integration kind of guy: I want some chunky bits that stand out as different from the rest, but I also want a fairly uniform tasty brownish goo for the identifiable bits to swim in.
We need some weird cities. And those of us not like the weird cities are going to hate the weird cities. Always have, always will. But at the end of the day, weird cities also serve as innovation hubs, as immigrant gateways, as deployment centers for the defense of the nation, as the communities that feed the world through advanced agriculture, etc. Not every city needs to be the same. So as far as “real America” goes: yeah, New Yorkers and Silicon Valley types are uniquely detached from and ignorant of how most Americans live in most places. True. But I’m not sure we really want that to change. Sometimes, ya just gotta keep Austin weird.
For the record Austin, you’re not that weird: 12.05 score, below-average weirdness. Go ahead. Send me your hate-mail.
Check out my new Podcast about the history of American migration.
If you like this post and want to see more research like it, I’d love for you to share it on Twitter or Facebook. Or, just as valuable for me, you can click the recommend button at the bottom of the page. Thanks!
Follow me on Twitter to keep up with what I’m writing and reading. Follow my Medium Collection at In a State of Migration if you want updates when I write new posts. And if you’re writing about migration too, feel free to submit a post to the collection!
I’m a graduate of the George Washington University’s Elliott School with an MA in International Trade and Investment Policy, and an economist at USDA’s Foreign Agricultural Service. I like to learn about migration, the cotton industry, airplanes, trade policy, space, Africa, and faith. I’m married to a kickass Kentucky woman named Ruth.
My posts are not endorsed by and do not in any way represent the opinions of the United States government or any branch, department, agency, or division of it. My writing represents exclusively my own opinions. I did not receive any financial support or remuneration from any party for this research. More’s the pity.