Yglesias and Kristof Get Illiteracy Wrong

Recent headlines touted a “shocking new study” revealing that 47% of adults in Detroit are functionally illiterate. The story got big play in both the traditional press and the blogosphere (see this and this; the reliably superficial Nick Kristof even tweeted it to his million-plus followers).

Journalists loved the story because it meshes with the agenda of the big-money foundations and corporations that want to scare us into privatizing our schools. After all, if things are that bad, surely any kind of “reform” is justified – even it if it involves breaking unions, trashing teachers, and turning urban schools into drill-and-kill testing mills. Wrote prominent “progressive” blogger Matthew Yglesias: “Under [these] circumstances, I find it difficult to be seized with worry that schools are going to be ruined by teachers ‘teaching to the test.’”

We were curious about the source and methodology of the study, so we spent a few minutes researching it.

Turns out the “new study” is actually a report by the Detroit Regional Workforce Fund (to his credit, Yglesias links to it).  The report contains no research but simply states, in a single bullet point, that “the National Institute for Literacy estimates that 47% of adults (more than 200,000 individuals) in the City of Detroit are functionally illiterate.”

That’s it. No new research; no new study; no new data. So where did the figure come from?

We tried to track down the research on the National Institute for Literacy Web site and couldn’t anything remotely relevant. (One thing we did find while Googling is that this same “new” factoid, variously sourced, has been a staple of  white supremacist Web sites for at least a decade.)

So where did this oft-cited, little analyzed statistic originate?  After a fair amount of hunting we located the source: It’s a 1998 National Institute for Literacy Report entitled “The State of Literacy in America: Estimates at the State, Local, and National Levels” where — if you look hard enough — you can indeed find an estimate that 47% of adults in Detroit are at “Level 1 Literacy.” (Level 1 Literacy is not quite the same as functional illiteracy, by the way, but it’s close enough that we’ll let the distinction slide.)

The 47% figure is ultimately derived from a 1993 federal study, the National Adult Literacy Survey (NALS). The NALS, it should be noted, did not include statistics at the municipal level because the sample size – 26,000 adults nationwide — was regarded as too small for confidence at the local level.

So the 1998 study crunched raw NALS data for US counties with 1990 census data, applying a computerized “regression model” to “predict literary proficiencies for other aggregates [i.e., cities].” The resulting estimates, the study warned, had a 95% confidence interval larger than plus or minus five points and “should be used with corresponding caution.”

There you have it: not new research, but a 1993 computerized extrapolation of information collected two decades ago. It’s not even on point, since the low literacy rates among adults in US urban areas have at least as much to do with recent immigration as with failing schools.

To be sure, the statistic is not exactly encouraging, but it’s far from an urgent mandate for doubling down on testing and other corporate reforms. Too bad Yglesias, Kristof, and their unthinking followers couldn’t be bothered to do some elementary research before granting new life to an ancient and misleading factoid.