While the leadership change was far more about the personalities involved than the polling, if Rudd was leading 57/43 — we would probably be talking about the soccer today.
Yet while Rudd’s polling results turned sharply in October last year on every polling metric and have been on a downward trend until very recently, he was, as of this morning, still leading a party in an election winning position. If we track every pollster two party preferred result over the last 12 months, we can see that Labor had only really ‘lost’ 3 polls.
The 2007 election result is shown as the dotted line, so while the most recent polls had Labor trailing their election outcome, they were still in an election-winning position and their polls had been recovering for over a month now. This is further highlighted by today’s Newspoll quarterly breakdown published in The Australian, which suggested that were an election held over the last three months, on the basis of the state level polling data, Labor would have won with around 81 seats to the Coalition’s 66.
The public opinion kicker — insofar as there was one — appears to have come with what internal polling was showing.
ALP party polling had been delivering results for the ALP poorer than the public polling for a number of months now — though not with the headline national figures, polling is polling is polling in that regard. The major problem popping up with the swings involved came from the nature of geographical distribution — where they were occurring and how a large number of Labor Party marginal seats densely populated those particular areas, and were filled with larger numbers of newly hostile demographics than the national average.
We’ve seen a little of this from the public pollsters — such as this week’s Newspoll results in marginal QLD seats — but the broad level public polling, even with breakdowns to the state level that we see by aggregating consecutive Nielsen polls or today’s Newspoll quarterly breakdown, didn’t accurately reflect the electoral danger for Labor, particularly on the primary vote.
Whether that has been conveniently overcooked a little by the ALP — as these things inevitably are in any leadership contest — is a matter for debate. Seats generally fall at elections on the basis of a mean swing of X and a standard deviation of between 2 and a bit and 3 and a bit, so it would be quite an extraordinary result if Labor was getting the very rough end of the pineapple on that variation, to the point where they were leading the two party preferred nationally, but getting creamed on the electoral pendulum– but ultimately, we’ll now never know. Take it with a grain or two of salt.
The other concern with the polling, both internal and public, has been Rudd’s personal standing with the electorate operating as a brake on the Labor party vote. If we track the polling performance of Rudd on Better PM and Rudd’s satisfaction rating since he obtained the top job, we see it falling off a cliff from October last year:
With the recent growth in the Greens vote, there is certainly an argument to be had there — especially when perceptions of Gillard in the very areas where Labor was apparently hemorrhaging votes is holding up extremely strongly.
On the public opinion front, the ALP is gambling that the concerns voters have with the government are over issues that Kevin Rudd is seen to, literally, “own”. By removing a leader, they hope that they remove their problems.
Public opinion usually doesn’t work that way for very long — insert NSW and QLD state governments for the latest examples of a very long list — but we’ll find out soon enough.