His conclusion? Uh, Noonan needs some more data if she’s going to corroborate her contention:
[E]ven with no political targeting at all, hundreds of thousands of conservative voters would have been chosen for audits in the I.R.S.’s normal course of business. Among these hundreds of thousands of voters, thousands would undoubtedly have gone beyond merely voting to become political activists.The fact that Ms. Noonan has identified four conservatives from that group of thousands provides no evidence at all toward her hypothesis. Nor would it tell us very much if dozens or even hundreds of conservative activists disclosed that they had been audited. This is exactly what you would expect in a country where there are 1.5 million audits every year.
At the end of his post, Silver does a bit of knife-twisting, writing that some folks last year engaged in similar quantitative atrocities in concluding that Mitt Romney was likely to win the presidential election, “while dismissing polls that collectively surveyed hundreds of thousands of voters in swing states and largely showed Mr. Obama ahead.”