Do you think it could be used to judge the "truthiness" of media websites online? I don't think it works with rounded numbers that are more likely to be found in the news.
Do you think it could be used to judge the "truthiness" of media websites online? I don't think it works with rounded numbers that are more likely to be found in the news.
I'd love to spend some of my copious free time auto-pulling the 10-Q's from a financial site and applying it to earning statements. I'd bet a serious amount of money that many companies financial reports disagree with Benford.
Nate Silver actually did a bit of analysis on it around the time of the 2009 Iranian elections and the Al Franken-Norm Coleman recount. In short, he explained that while it does hold in certain circumstances, it's not really a 'law' that can be applied very broadly, unfortunately, even if the idea makes sense intuitively. For example, I believe he said that the Iranian ballots did not follow Benford's law, but that this was not necessarily very problematic, because election results in general have no reason to follow Benford's law.
(I don't have the link, but that should be enough to find it in the 538 archives for anyone who's interested).