Every pollster, including us, was saying that Labor was in trouble in the Queensland election. The predicted swing was in the realm of 7 percent. From what you can tell, internal ALP research also seems to have supported this. In our last poll it appeared that the “undecideds” were breaking towards the LNP as well, and the LNP was well-placed on the issue of health, which was much more significant to voters than jobs, which was Labor’s issue.
So my prediction was that Labor would lose 12 to 15 seats. This was based on a uniform swing of around 6.6% to 8%.
In the event the Labor government may have lost 5 seats or less and has experienced a swing against it on primary votes of 4.36%. (I’m not counting seats here that were held by another party although notionally theirs). http://www.abc.net.au/elections/qld/2009/guide/changingseats.htm
Whatever movement occurred most probably happened in the last few days of the campaign. As there were no major policy announcements, or any blunders, in that period the change is most probably the result of a shift of mind by voters.
Early on the election was framed by voters as a vote of confidence in the government. Expectations were high that Labor would win. These expectations had decreased dramatically as the election progressed as you can see from the table below, to the point where a majority of our respondents expected Labor to lose.
This would have concentrated voters’ minds on the alternative government. As our polling showed, voters were unimpressed with both parties. In the end result it appears that judged one-on-one they were less impressed with the Liberal Nationals than the Labor Party.
Another factor in the swing was undoubtedly the marginal seat campaigns of both parties, local issues and the quality of local candidates. There was a huge variation in swings from a swing to the LNP of 13.6% in Hinchinbrook to a swing of 3.2% to the ALP in Whitsunday.
To adequately map these factors you would need an opinion poll sample of somewhere around 24,000, which is obviously not economically feasible. The accuracy of opinion polls therefore relies on incumbency, candidates, local issues etc. cancelling each other out. But this is not necessarily the case.
It looks from the swings as though the LNP did a very poor job of local campaigning. For example, the seat of Chatsworth would have been expected to go close to the average swing because the sitting member Chris Bombolas retired after one term in parliament, meaning that the incumbency factor was taken out of the equation. Yet the LNP failed to get any swing at all. This was the most marginal ALP seat in the state and should have been well in focus. Perhaps they were so cocky that they took it for granted.
I have produced two Leximancer maps below. The one on the left shows the issues as they were at the beginning of the campaign, and the one on the right as they were by the middle of the last week. (I know they are too small, but click on them and you will get something you can decipher.)
The image on the left is much more fragmented than the one on the right, suggesting that both the parties had successfully reduced the election to a smaller set of propositions by the end of the campaign.
Interestingly the (Global financial) Crisis has completely disappeared, even though it was Anna Bligh’s stated reason for calling an early election. A new concept appears in the second map – Government. This is larger than all the other themes and closely related to Health and Infrastructure. It suggests that the LNP has been successful in turning the election into a referendum on the government. It should also be noted that Health encompasses more than just health, with concepts inside the Health circle indicating that it is actually a catch-all for service provision.
Climate change also played a part in some electors’ minds.
The fact that judged on these maps the LNP appears to have gotten the better of the debate, underlines an important aspect of political campaigns – they are rarely about issues. In the case of this election Labor appears to have won based on a decision by voters not to risk Springborg and the LNP even though they agree with their analysis of what is wrong with the state.
I’m happy with our performance on this poll. While we didn’t capture the final result, no-one else did either, indicating that it was determined at the last moment at a time when we couldn’t have been measuring responses.
Our polling, along with that of others, would have actually helped to cause the final result by accurately measuring the level of dislike for the Labor government. By accurately describing voting intentions during the campaign we would have alerted voters to how close it could be, and inadvertently affected their decisions. This is something that professional campaigners take into account, and is part of the price you pay for knowing what is happening.
