My post for the Ministers Of Design Blog
How do we measure racism and homophobia across the United States? Humboldt State's Dr. Monica Stephens teamed up with Floating Sheep, the same group that mapped post-election Twitter hate speech to broaden the scope of the study and give a more panoramic view of America's bigotry. The Geography Of Hate map
was created by geo-coding 150,000 hate tweets between June 2012 and
April 2013, dividing the tweets in three categories--racist, homophobic,
and disability-hating, including the words "chink," "gook," "nigger,"
"wetback," "spick," "cripple," "dyke," "fag," "homo," or "queer,"
amongst others. You might argue, however, that context is everything
when it comes to these words so how did the research control for that
variable? They used humans (probably woefully underpaid or even unpaid
Ph.D. students, natch) to analyze and code the 150,000 tweets, eschewing
machine inability to read tone and coding the usage as negative,
neutral, or positive.
To add more rigor to the study, the researchers accounted for tweet
density by creating a scale, essentially measuring something akin to per
capita hate, accounting for population density.
So, what can we conclude from all this? On a micro level, there are
some rather surprising results--click on the n word, for example, and
you will see for yourself...the Deep South is not the hotbed of racism
it is often stereotypically cast as. On a more macro level, hate speech
is clearly alive and well-spread across America. In addition, the study
demonstrates that Twitter has become a really vibrant (and vociferous)
platform for the spreading of hateful ideas and even recruiting people
with that sort of rhetoric. Now you might argue that 150,000 tweets is
not a wide enough sample to make conclusions on, but this is a prime
example that Twitter *can* have scholarly utility (don't worry, consider
me as shocked as you are).