One of the most challenging aspects of anti-racism is the fact that we can only usually measure racism as an absence of a better explanation. We see an inequality and then we try to rule out the other plausible explanations, and then say “it’s got to be explained by racism”. Because there is no objective test – no screen or marker or physical indicator – that positively identifies racist intent (or even racism that happens unintentionally), it is usually left to anti-racist educators to make a case through narrative explanation rather than through empirical observation.
Their (our) task is made even more difficult by the fact that, partially because people are defensive and partially because people are assholes, any claim that racism plays a role in any event is met with a howling chorus of denials and demands for the kind of rock-solid proof that is so rarely available when discussing these kinds of social/psychological issues. When these demands cannot be readily met (‘my racism detector is on the fritz’), these voices devolve into smug pronouncements of ‘race cards’ being played, or perhaps a ‘playing the victim’ gambit being used.
Which is why it’s always interesting and gratifying to see exercises like this one:
During the day after the 2012 presidential election we took note of a spike in hate speech on Twitter referring to President Obama’s re-election, as chronicled by Jezebel (thanks to Chris Van Dyke for bringing this our attention). It is a useful reminder that technology reflects the society in which it is based, both the good and the bad. Information space is not divorced from everyday life and racism extends into the geoweb and helps shapes its contours; and in turn, data from the geoweb can be used to reflect the geographies of racist practice back onto the places from which they emerged.
Using DOLLY we collected all the geocoded tweets from the last week (beginning November 1) with racist terms that also reference the election in order to understand how these everyday acts of explicit racism are spatially distributed. Given the nature of these search terms, we’ve buried the details at the bottom of this post in a footnote .
Given our interest in the geography of information we wanted to see how this type of hate speech overlaid on physical space. To do this we aggregated the 395 hate tweets to the state level and then normalized them by comparing them to the total number of geocoded tweets coming out of that state in the same time period . We used a location quotient inspired measure (LQ) that indicates each state’s share of election hate speech tweet relative to its total number of tweets. A score of 1.0 indicates that a state has relatively the same number of hate speech tweets as its total number of tweets. Scores above 1.0 indicate that hate speech is more prevalent than all tweets, suggesting that the state’s “twitterspace” contains more racists post-election tweets than the norm.
So before we get into the results of the exercise, I want to make the point that this is not the same thing as measuring racism at a state level. Racism manifests itself in a large variety of ways, only some of which are as blatant as the search terms used by the authors (they use “nigger” “monkey” “Obama” “elected” and “won”). There are types of racism that cannot be detected through angry tweets – housing or job discrimination won’t show up in social media, nor will wage gaps, microaggressive behaviours, or underfunded schools. It might be more accurate to say that this is a proxy measure for how comfortable people in different states are with making public and overt racist statements.
That being said, because comfort with overt racism is usually correlated with prevailing racist attitudes, it’s not a stretch to conclude that someone who’s going to tweet about how angry they are that the monkey nigger got re-elected is probably not too bothered by other, non-obvious forms of racism. And a community that spawns such a person is likely a community that has attitudes on race that are, shall we say less than enlightened?
With all that in mind, let’s look at the results:
State LQ of Racist Tweets Alabama 8.1 Mississippi 7.4 Georgia 3.6 North Dakota 3.5 Utah 3.5 … … Nevada 0.5 Iowa 0.4 Indiana 0.3 New York 0.3 Arizona 0.2
So a couple of perhaps surprises here, as well as a couple of things that probably don’t surprise you at all. It’s interesting that North Dakota and Utah round out the top of the list, whereas states like Arizona and Oklahoma with recent and major issues with race and racism come in among the states with the lowest quotient. I would expect that states with a long history of racial tension, low average income, and where white and black residents live in relatively close geographic proximity to each other (although not necessarily places where they live/work together) would have more frequent racist tweets.
In the study’s FAQ, the authors specifically deal with some of the major objections to their methodology, including the fact that this only includes tweets that are geocoded (meaning that they are more likely to come from smart phones than from computers), and is thus not necessarily a representative sample of all tweets sent with this content. It is also worth noting the similarities between this exercise and a similar one performed with Google searches, which I’ve blogged about before.
Measuring racism outside of the context of the humanities is a problem that will continue to plague the conversation on race. While psychologists and sociologists are able to make great strides and have been laying the groundwork for a rigorous field of research, the lack of easily-digested numerical references continues to be a stumbling block in getting anti-racist thought accepted into mainstream discussion. I don’t doubt that most people agree that racism exists, but until we can produce findings like these for less dramatic examples than “Obama is a monkey nigger”, we’re going to continue to face an uphill battle from those who would rather deny reality than face it.
Like this article? Follow me on Twitter!