Update on Australian Polling

After the surprise turnaround last month that resulted in Kevin Rudd replacing Julia Gillard as leader of the Australian Labor Party, and thus Prime Minister, I’ve been keeping a close eye on opinion polls.  It’s been widely assumed that Rudd is a far more popular politician than Gillard, perhaps because his as Prime Minister between 2007 – 2010 is being seen in a positive light or due to Gillard’s unpopular leadership, and it was hoped that he could save the party from the wipe-out it seemed to be spiraling towards.  Labor had only barely clung on to power in the 2010 election, winning the same number of seats as the Liberal-National coalition (and actually having one seat less now, keeping afloat only through cross-bench support), and their subsequent loss of support has not looked positive.

Yet, Rudd has improved the party’s prospects significantly.  While a month ago Labor was polling at least 10% behind the Coalition, 12 polls since have shown a distinct change.  Granted, Labor have only had a lead in 3 of these polls (4 if you’re counting two-party-preferred vote) and all from the same company, but every poll has shown a narrowing of voting intention.  The Coalition only has an average lead of 3% in the last couple of weeks of polls, which constitutes a major reduction.  Polling shows that, although it’s an uphill struggle, Australian Labor still has everything to play for.

Using this nice page to calculate seats, we can get an idea of what these votes will actually mean on election day.  If we take an average of every poll since Rudd became Prime Minister, we get a result something like:

Labor – 71 seats
Coalition – 76 seats
Other – 3 seats
(150 seats in total)

This would give the coalition the narrowest of majorities.  Using this data, we can predict the probability of various results:

Labor majority: 36% likely.
Coalition majority: 45% likely.
Hung parliament: 18% likely.

This election, the date of which has not yet been decided but must be held before the 30th November, is not predictable by any means.

(I often find maths boring, but when calculating election things I’m incredibly thankful for that Higher maths course!  I could think of many worse jobs than this…)