Use Google Searches to Figure Out How Racist Your Neighborhood Is
Want to know whether a given area has a higher number of racists than average? It turns out that Google searches can provide you with a pretty good answer, at least according to two recent studies.
A few years ago, data analyst Seth Stephens-Davidowitz set out to study whether “racial animus,” or racism, could affect the outcome of a presidential election where one candidate was black. Luckily, he had a great dataset: Obama had just been elected, and there was a ton of data on how various areas had voted. But what he needed was a good proxy that measured the racism in these places. He decided to use Google searches on the word “nigger,” which he calls (quoting Randall Kennedy) “the paradigmatic slur.”
I define an area’s racially charged search rate as the percent of Google searches, from 2004- 2007, that included the word “nigger” or “niggers.” … The epithet is searched for with some frequency on Google. From 2004-2007, the word “nigger(s)“ was included in roughly the same number of Google searches as words and phrases such as “migraine(s),” “economist,” “sweater,” “Daily Show,” and “Lakers.” The most common searches that include the epithet, such as “nigger jokes” and “I hate niggers,” return websites with derogatory material about African-Americans. From 2004-2007, the searches were most popular in West Virginia; upstate New York; rural Illinois; eastern Ohio; southern Mississippi; western Pennsylvania; and southern Oklahoma.
He’s careful to point out that he tried to control for non-racist uses of the word:
In particular, I control for search rates for “African American,” “nigga,” (the alternate spelling used in nearly all rap songs that include the word), and profane language.
While this certainly doesn’t control for all non-racist uses, it does go a long way toward disambiguating the many connotations of the word. This is underscored when you consider that Stephens-Davidowitz found that, after controlling for “nigga,” the most popular results included clearly racist phrases like “nigger jokes” (a phrase that’s popular on Reddit’s racist “Chimpire” subreddits as well).
What Stephens-Davidowitz ultimately discovered was that racism “appears to have cost Obama roughly four percentage points of the national popular vote in both 2008 and 2012.” He determined this by showing that the higher the number of searches there were on “nigger” in a given area, the more likely it was that Obama lost votes there — even controlling for things like income, already-existing political affiliations, and more. In other words, even in an area where people typically voted for Democrats, you’d see a less-than-typical number of votes for Obama ifthe rate of Google searches on “nigger” was higher than average. As Stephens-Davidowitz put it, “An area’s racially charged search rate is a robust negative predictor of Obama’s vote share.”
Now Stephens-Davidowitz’s method of identifying racist regions has made its way into another fascinating data analysis, published a month ago in PLoS One. This paper, authored by a team led by University of Maryland, College Park, biostatistician David Chae, deals with the correlation between regional racism and higher-than-average rates of black mortality.
The team focused on 196 “designated market areas,” or regions defined by Nielsen Media Research (yes, the group that does Nielsen Ratings), and looked at Google searches for “nigger” in these areas. They show the results below, in a map that reveals which areas are a half-standard deviation away on either side from the average number of searches for the word.
“Notably, this shows that greater proportions of Google search queries containing the “N-word” were concentrated in the rural Northeast and South of the US,” the authors write.
They also found that, looking at mortality rates for blacks over the age of 25 from stress-related conditions, there was a correlation between higher rates of death and higher rates of racism. Specifically, they found a 5.7% higher rate of death among blacks in the areas with the greatest levels of racism.
DMAs characterized by a one standard deviation greater level of area racism were associated with an 8.2% increase in the all-cause Black mortality rate, equivalent to over 30,000 deaths annually. The magnitude of this effect was attenuated to 5.7% after adjustment for DMA-level demographic and Black socioeconomic covariates.
It has already been demonstrated in many other studies that racism — and indeed, social stigmatization of many kinds — can lead to greater stress, and therefore higher mortality rates. But this is the first study to quantify those rates using this new measure of racism.
These studies suggest many areas of further research to be done. You could look for correlations between racist Google searches and bank loan rates to black families, for instance; or just search your city for the most racist neighborhoods and correlate that with housing prices. I look forward to seeing the results.