Marking predictions to market

A couple of weeks ago, I wrote about a Sydney Morning Herald story on new electorate-specific robopolls and their surprising predictions. Tim Colebatch at the SMH wrote

Uh-huh. Lonergan’s own national poll reports only a 2 per cent swing against Labor. Yet in the three seats it polled individually, it found an average swing of 10 per cent. That’s huge, far bigger than we have seen in any Federal election since 1943.

and I was similarly dubious.

We now have the facts: the national (two-party preferred) swing against Labour is just over 3%, and the swing in Kevin Rudd’s seat of Griffith, one of the three specifically polled and predicted to have a swing of 10.5%, was 5.42%. The other two seats mentioned in the story as polled by Lonergan were Forde (robopoll swing 8.5%, actual swing 2.5%) and Lindsay (robopoll swing 11%, voters 3.9%)

It looks as though the robopoll skeptics were right. Even though the national swing was larger than the poll predicted, the swings in the target electorates were much smaller.