Harper’s bucket defence of illegal robocalls

The Harper government has mounted a classic bucket defence* against charges it illegally steered opposition voters to faraway, fake polling stations in a deliberate attempt to discourage them from voting. Their defenders say: 1. Nothing serious happened. 2. It happened to us too. 3. There’s no proof we did it. 4. In fact, it was the Liberals who did it. 5. The calls didn’t work anyway. 6. Voters don’t care about it. 7. It’ll blow over in a day or two.

Some of this commentary is just the predictable party-line pandering from pro-Harper media, but a Globe and Mail story purporting to show that the robocalls failed to reduce voter turnout rests on a surprising abuse of logic and statistical analysis.

Globe reporter Éric Grenier compared ridings with robocall allegations with ridings that had none. He found slightly higher voter turnout in robocall ridings. He also found that turnout in those ridings had increased more over the 2008 federal election than it had in non-robocall ridings

From this Grenier concluded that the robocalls “failed miserably” (the words of the headline) and that “Turnout was not much affected” (his own words). His data justifies neither conclusion. The turnout-suppressing impact of the robocalls may simply have been overwhelmed by one or more turnout-enhancing factors, such as what he acknowledges were tighter contests in the robocall ridings, or keener interest in what voters saw as a landmark election. The pertinent question is not whether voter turnout was higher or lower than other ridings or previous elections, but whether it would have been higher still without the robocall campaign.

That question—what would turnout have been without the robocall campaign?—is counterfactual and therefore not susceptible to proof. The flaw in Grenier’s reasoning is so basic and so obvious, it’s a wonder no editor caught it.** Given the “failed miserably” headline, however, it’s perhaps more likely his editors were egging him on.

* The bucket defence, a Paul Wells or possibly George Will coinage, is akin to Freud’s Kettle Defence. That’s when a neighbor returns a borrowed kettle with a hole in it. Confronted, the neighbor responds, “The hole was there when you lent it to me. I never borrowed the kettle. Besides, I returned it in perfect condition.”

** Grenier backs off his claim that “turnout was not much affected” ever so slightly in this subsequent blog post.