The latest investigation into allegations of election fraud isn’t being spearheaded by a respected bastion of journalism or the Federal Election Commission—it’s being conducted by a group of statistics-savvy redditors.
Earlier today, a member of the social news site who goes by rhetorikill posted a series of data-intensive arguments and visuals attributed to people who claim to have uncovered “undeniable evidence of voter fraud” benefiting Mitt Romney in primary races dating back to his previous presidential run in 2008.
According to rhetorikill’s post, he wasn’t alleging fraud: He wanted to draft his fellow redditors and use their collective expertise to either prove or debunk the findings.
The analysis he provided was fairly raw and more than a little intimidating to those of us who aren’t stats wonks. Nevertheless, many redditors were up to the challenge.
Tribbletrouble, a master’s student in finance, was among those who took a crack at the data. The user concluded that the statistical function used in the original analysis was flawed, and while the approach might work for random variables like pulling different-colored marbles out of a jar, elections are not random. If Romney beats Ron Paul by 20 points, it’s not because Romney was lucky or Paul was unlucky.
Users also took issue with rhetorikill’s failure to consider demographics as a factor. For example, he claims that Romney’s success in larger precincts is suspicious, but that’s not a statistical anomaly, says SwillFish, that’s just politics:
“Urban voters are generally less conservative and are thus probably more likely to vote for Romney because he is considered to be the most ‘mainstream’ and centrist.”
But even though redditors shot down the claim of “undeniable evidence” advanced by rhetorikill’s sources, the discussion revealed two important trends: First, the ease with which statistics and fancy language can be spun to make inaccurate or incomplete assertions, and second, the intellectual curiosity of the Reddit community.
I can’t think of many online communities where robust statistical analysis is business as usual. If the sharpest users of every news site were willing to collectively devote hours of their time to debunking questionable assertions, we’d have a much healthier climate for political debate.
Then again, sometimes the most basic explanations hold the most truth. As a user named Bryndyn described Romney’s success, “More money != More votes.”
Photo by WEBN-TV