I tried to avoid instant reactions to the election outcome in May. But now that lots of people are making claims I regard as dubious at best, I think I may respond. Before doing that, I thought it would be useful to make some general observations about mistaken/dubious claims that are commonly made in post election analysis, particularly following a close election.
Just about everyone, including me, is prone to these kinds of reasoning. Feel free to discuss, give examples, and so on.
Update: Much of what I planned to write has been covered, much better, by Peter Brent at Inside Story.
- The winning party/leader has demonstrated a deep understanding of the electorate, unlike the losers
At least when the margin of victory is narrow, this is magical thinking. If one or two voters in a hundred had decided differently, the opposite claim would be made with equal confidence.
- Factor X was the (implicitly unique) cause of the outcome. As in a close sporting match, every missed shot and every bad call is as crucial as every other. So, there’s no necessary inconsistency between the claims that “we lost because of media bias”, “we lost because of policy X” and “we lost because the leader was unpopular”. Plausibly,
each cause waschanging any of these factors would have been sufficient (other things equal), and, conversely, none was necessaryessential. (There’s an opposite version of the fallacy. If someone says “policy X cost us votes”, a supporter of X may deny it and say “we would have won with a better leader”. As I’ve said, both can be true).
- There were big swings in electorates with many voters of Type X. Therefore these voters changed sides.
This one even has a name, the “ecological fallacy“. To give the classic example, the fact that electorates with a low average income changed hands, does not mean that low income voters necessarily changed sides.
- The outcome proves we should do X, where X is something the speaker has been advocating all along.
Essentially, this is the same as 2, though usually with less evidence that not doing X changed the result.