Richard Florida is a co-founder and editor at large of CityLab and a senior editor at The Atlantic. He is a university professor in the University of Toronto’s School of Cities and Rotman School of Management, and a distinguished fellow at New York University’s Schack Institute of Real Estate and visiting fellow at Florida International University.
The list of America's Most Literate Cities suggests there's little relationship between literacy and economic development. Our analysis suggests otherwise.
I’m an unabashed fan of lists. I’m not talking about made-up, off-the-top of someone’s head lists, but real data-driven rankings.
That’s why I’ve been so impressed by the America’s Most Literate Cities list. Created back in 2003 by John W. Miller, president of Central Connecticut State University, it parses the relative literacy of America’s 75 largest cities, based on six key indicators: newspaper circulation, magazine and journal circulation, online reading and book purchases, number of bookstores and public libraries.
The index has gotten terrific coverage in the press (including this story by my Atlantic Cities colleague John Metcalfe). But one thing that jumped out at me was the broader question of urban literacy.
The study's authors claim that literacy is not connected to a city’s income or wealth. "Using U.S. Census data for income in the relevant cities, I learned that wealthier cites are no more likely to rank highly in literacy than poorer cities," Miller writes. "While poverty has a strong impact on educational attainment, its impact on literacy is much weaker."
I was surprised by this finding. A huge body of research documents the connection between human capital (typically measured as the percent of adults with a college degree) and income. It would only stand to reason that literacy rates across cities and metros would vary in much the same way.
So I decided to take a quick look. I asked my MPI research team to run an analysis of the index, comparing it to income, education and other key economic and demographic factors. We did the analysis at the metro level, which we believe is reasonable and appropriate for several reasons.
First, the metro level is the standard for urban and regional analysis, which allows us to compare apples to apples. Second, several of the indicators used in the index clearly reflect metro-wide trends. Bookstores and large central library facilities, for example, typically service people from the surrounding suburbs as well as the center city. As a high school student in a Newark, New Jersey, suburb, my father frequently drove me to the Newark Public Library to do research.
Even indicators like newspaper and magazine circulation and online activity can spill between city borders. Some people who live in suburbs subscribe to newspapers and magazines at their office locations in the city; many more surf the net from computers at various locations where they live or work.
Ultimately, we were able to match our data to 60 of the index’s observations. As usual, I note that correlation does not in any way imply causation; we are simply measuring associations. The associations we found do suggest a fairly strong connection between literacy, income and education.
We found that a city’s rank on the literacy index was significantly associated with median metro income (.62) and slightly less so with economic output per capita (.48), a standard measure of regional productivity. There was a strong negative correlation between literacy and the metro poverty rate (-.61).
We also found correlations between literacy and the percent of adults in a metro with a college degree (.7) and the share of a metro’s workforce in professional, knowledge-based and creative class jobs (.55). We found a strong negative correlation between literacy and the percent of the workforce in blue-collar jobs (-.38). Literacy was positively related to a metro’s share of high-tech industry (.39).
One explanation for the difference in findings could be that we used metros as the unit of analysis, while the original study used cities. In an email exchange, Miller noted this and several other reasons we draw different conclusions.
The first is the use of metro area as opposed to city as the unit of analysis. Second, the connection between education level and income level is clear and never in doubt. In fact, part of the correlation between our overall assessment of literacy and income which does exist is partly attributable to the fact that educational attainment is one of the sets of variables in the determination of the literacy score.
My hunch though, given not just our findings but the existing body of research on the subject, is that literacy remains tied to both income and education. The sad fact of the matter is that access to books, newspapers, magazines and other "literacy" resources, like so many other things, remains terribly unequal in our society, tied to key factors like income, education and socio-economic class.
In pointing that out, I do not mean to in anyway minimize the incredible literary resources available in many less affluent areas.
Nor do I seek to undermine the Literacy Index. Efforts like it provide the kind of hard data and baseline indicators which are critical to helping us better understand and devise solutions for overcoming inequalities in education and literacy that continue to plague our nation.
Photo credit: Shannon Stapleton/Reuters