Statistics collected by @alittleracist, 10 April 2012
The Twitter-account @alittleracist looks at who claims ‘not racist but…’ on Twitter. This is how the data is gathered. And the following are some of the findings so far…
Stats updated as per Tuesday 10th of April 2012.
Total sample size of all tweets analysed so far: n=18370
By Gender
A twitterer’s gender is determined by examining their profile photograph and/or name. If neither gives an indication of their gender, or if the two data sources conflict (for example, a photo of a guy, but a name marked as ‘sarah’), they are marked as ‘unknown’
By apparent ethnicity
A twitterer’s ethnic appearance is a guesstimation based on their profile picture
They claimed they weren’t racist… Were they?
If a twitterer says something racist, they’re marked as such. ‘Unknown’ is used if they say something that may or may not be racist, but where it is not immediately obvious from the context which it is. They may, for example, making a joke (‘I am not racist, but I prefer brown bread’) or similar. If someone says something like ‘If you say Not Racist But, you may as well stop speaking, nothing good will follow’, then this is counted as an anti-racist message; someone is using the words ‘not racist but…’ in an effort to encourage other Twitter users to be less racist.
And was their racism highlighted on the @alittleracist twitter account?
Other interesting statistics…
Obviously, there’s tons of interesting statistical analysis we can do here; which ethnicity tends to be more racist? Which gender is more racist? Are white men more racist than black women? Etc. This will all be done once we have a sufficiently large data set. For now, enjoy the following:
Number of followers
Average number of followers for people who tweeted something racist: 457
Average number of followers for people who tweeted something anti-racist: 431
Literacy & spelling
On average, racists spell 6.21% of their words wrong
On average, anti-racists spell 3.89% of their words wrong
More literacy stats
On average, males who tweeted something racist misspelled 5.9% of their words. n=5956
On average, females who tweeted something racist misspelled 6.57% of their words. n=6497
On average, black people who tweeted something racist misspelled 7.44% of their words. n=5238
On average, white people who tweeted something racist misspelled 5.15% of their words. n=4548
On average, black males who tweeted something racist misspelled 7.24% of their words. n=2291
On average, black females who tweeted something racist misspelled 7.6% of their words. n=2892
On average, white males who tweeted something racist misspelled 4.8% of their words. n=2371
On average, white females who tweeted something racist misspelled 5.55% of their words. n=2060
On average, males misspelled 5.68% of their words. n=7685
On average, females misspelled 6.51% of their words. n=8094
Even more statistics
There is a strong inverse correlation between the number of followers people have, and the number of words they spell correctly.
In other words: The more followers you have, the better your spelling. Or: The worse your spelling, the fewer people can be bothered following you:
p value: < 0.001 / Spearman’s R statistic: 0.037 /Degrees of Freedom (df): 10791
Women tend to be more racist than men:
p value: 0.003 / Pearson’s Chi Square statistic: 11.459 / Degrees of Freedom (df): 2
… But men tend to post more things that are racially ambiguous
