Slate, you're doing it wrong

Slate, you’re doing it wrong

There’s a map that Slate made recently that’s been getting a lot of shares.  The map is part of a post on language diversity, and it shows the top language in each state other than English or Spanish.  We dug a little deeper, and found a few problems.

First, looking at the most recent American Community Survey data,* we find more Asian languages in the top spot than Slate found.  Like Chinese in New Jersey and North Carolina.  And Hmong in Wisconsin. And Vietnamese in Mississippi.

Corrected Slate Map

Being interested in Asian American and Pacific Islander data, we were also curious to see the top Asian or Pacific Islander language in each state. Here they are.

Most Common Asian Language Spoken at Home

Still, there is a deeper problem with the Slate chart than just getting the original numbers right.  It goes to the heart of why we should care about language in the first place.  Should we care about language as a matter or idle curiosity?  Or should we care instead about language need as it relates to barriers and discrimination for those who don’t speak English too well?  In some ways, this “limited English proficient,” or LEP, population is much more important for public policy, because it affects how resources on language assistance are allocated, from public hospitals, to signs in stores, to voting booths.

And when we look specifically at language need, we find that outside of Spanish, Asian language needs are at the top of the list in many more states.  Not too many German speakers have a need for English language assistance.

Largest LEP Populations Outside of Spanish

And, for those who are curious what this list of top language need would look like if we only looked at Asian languages, we get this.

So, there you go.  First, Slate makes some errors of fact.  Then, some problems of priority, drawing attention to language but not paying any attention to language need.

*Note: In these maps, we rely on the 2012 American Community Survey (ACS) 5-year estimate tables from American Factfinder.  Some of the results vary when we rely on the Public Use Microdata Sample file from the ACS.

**Update May 16, 8:30am PT: I should add that, in states where there was a language group listed as the top category rather than a particular language, I dug into this earlier set of published tables from 2006-2008, to see if there might be a particular language that might be in the top spot.  That’s how I picked up on Dakota in SD as (very likely) bigger than Chinese, and Aleut-Eskimo languages as potentially equal to Tagalog in AK (the older table shows Yupik as larger, but not among limited-English proficient speakers).  Got a very nice email from Ben Blatt at Slate this morning, noting this older table was the one he used.  Hope to have a “live chat” session in the next week or two, to dig into the various issues raised here, including language definition, aggregation/disaggregation, and importantly–language policy as it relates to these data.


Karthick-WWICSKarthick Ramakrishnan is associate professor of political science at the University of California, Riverside. His research focuses on civic participation, immigration policy, and the politics of race, ethnicity, and immigration in the United States. Ramakrishnan directs the National Asian American Survey and AAPIdata.com. He received his Ph.D. in politics from Princeton University, and has held fellowships at the Russell Sage Foundation, the Woodrow Wilson International Center for Scholars, and the Public Policy Institute of California.  Connect with Karthick on Twitter, @karthickr