30 March 2005
Newspoll confirms Liberal Party polling
The latest Newspoll findings tend to confirm the Liberal Party's polling for their negotiations with the National Party. While the Newspoll is for all of Queensland, and the Liberals only polled three Gold Coast seats, the results are consistent.
What it shows is that there has been virtually no movement in the vote with the Labor party down 0.5% on a two-party preferred basis. The Liberal Party polling showed a 2% swing to Labor on the Coast. Both these results are within the margin of error of the others' polling sample, and the Liberal's sample would have been more vulnerable to local issues.
The Liberals had the National Party on an 8% primary vote, while Newspoll has them on 11%. It has the Liberal Party on 27%, which is the same as the Liberal Party had themselves polling on the Gold Coast.
Most telling in the Newspoll survey is that Lawrence Springborg has experienced a strong deterioration in his personal approval rating, down from 50% to 37%. Peter Beattie's approval rating has actually increased one percentage point. Springborg's fall from grace is not as bad as it might seem. His dissatisfaction rating has only risen 4 percentage points, with the balance being taken up by those who are "uncommitted".
Springborg's drop in popularity is consistent with the polling (ppt 83kb)that he commissioned for his One Conservative Party campaign. The conclusion of this polling was that if he tried to force the concept on the Liberal Party and carried the fight on in public, the result would be worse than if he had never raised the issue.
One other matter should be commented on. The National Party says that these polls are never right, because in an election their vote rises. You can see this effect in the table of first preference votes where the National Party wins 17% of the vote at the state election, but this drops to 11% by March this year, while the Liberals, going in the other direction rise from 18.5% to 27%. There is a simple mathematical explanation for this which has nothing to do with the accuracy of the polls.
The polling questions are asked across the whole of the electorate and record what people would like to do if they could choose between all the parties. Under the various coalition agreements electors have been denied this choice. As the National and Liberal Parties contest half the seats each, in some areas Liberal voters vote National and vice-versa. Without this constraint, the Liberal Party is around two-and-a-half times more popular than the National Party. This result would most likely be duplicated if both parties contested all seats in the state at the next election.
It is also interesting to note that the Liberal Party vote rises by 8.5% while the National Party vote falls 6%, making the combined first preference vote higher when voters can notionally vote for either Liberal or National over the whole of the state. This backs up results over the years which show that there are more voters that will vote Liberal, but not National, than the reverse. By denying these voters the ability to exercise this preference in an election, the coalition agreements have acted to artifically inflate the Labor result. Again, this supports the Liberal findings that there are significant numbers of Liberal "Beattie" voters.
Posted by Graham at 12:34 PM
24 March 2005
Liberals and Nationals
Yesterday's on-air discussion centred on the worth of the Liberal Party polling suggesting that the National Party shouldn't run on the Gold Coast.
There were some pros and cons. John pointed out that at 600 people over 3 electorates, the sample size could have quite a large error in it in each electorate - in the order of 9%. I pointed out that this was quite possible, but as the samples appeared to be fairly consistent - that is they showed similar results in each of the three electorates - you would suspect that these were good samples.
It is also the case that with the polling showing the National Party on 8% of the vote across the three electorates and the Liberal Party on 27%, with each electorate showing similar proportions of support, even a large margin of error would make no difference to the basic conclusion.
National Party President Terry Bolger suggests that the Liberal polling is useless because it does not carry candidate names. John agrees with this to a point. It is certainly correct that candidates can make a significant difference, but it is not correct, as suggested by Mr Bolger, that a good candidate can win any seat.
Generally candidates are worth no more than a few percent (and sometimes that is a negative figure). I have yet to see polling that is affected by more than 10% by the candidate, but even here, good campaigning and their own party brand can bring them back to the pack. For instance, in the 1996 federal election, Labor members Lavarch and Johns were both 10% ahead of the generic Labor vote until the Liberal Party started running advertising reminding voters that while they might be popular members, when they went to Canberra they still voted with Paul Keating. The result - good night Michael and Gary.
Another finding of the Liberal polling is apparently that there is actually a swing to Premier Peter Beattie in all of the Gold Coast electorates polled (which were Mudgeeraba, Burleigh and Broadwater). Not a good sign for either party - they appear to be fighting over the silver medal!
The research was conducted by the Sexton Marketing Group, which is very close to the National Party. Apparently it was originally to be jointly commissioned, and the Liberals chose Sexton to minimise the chance of their erstwhile Coalition partners disputing the accuracy of the research. As it turned out, the National Party refused to be part of the research, so the Liberal Party continued on their own.
This is not the first time that the Liberal Party has tried to win an argument with the National Party based on joint research - the proposition was also put to them in 1994 before the 1995 state election (which the Coalition won). It is not surprising that the National Party refused to take part - the results are what one could have predicted on the basis of other publicly available polling.
Something else that comes out of the research is that three-cornered contests in these seats would make it even less likely that the non-Labor parties could win them because a percentage of voters do not intend to allocate preferences, including around a third of National Party voters. What this means is that while the generic Liberal and National Party votes add up to 35% of the vote, this will only happen if all the National Party voters allocated their preferences. As roughly a third don't intend to, that effectively reduces the combined vote to around 32%.
There's a complete air of unreality about this argument. Unless both parties start to represent electors better they'll have trouble finding any candidates to run in any of these seats, let alone good ones.
Posted by Graham at 06:38 AM | Comments (1)
15 March 2005
Troops to Iraq and the latest polls
Gary Morgan finds on the 12th of March, 2005, that 63% of voters disapproved of John Howard's decision to send another 450 troops to Iraq. This consisted of 51% of Liberal-NP voters, 84% of ALP supporters, 76% Greens, 72% Other Parties and Independent Candidates, and 70% Australian Democrats. These figures are close to ours although our sample is slightly more opposed. Morgan's figures of course are more reliable than ours because we are not trying to do a quantitatively accurate poll.
Morgan's poll also found that 51% of voters thought Australia should not have any troops in Iraq.
Meanwhile, Newspoll's latest offering, released in The Australian, shows that Australians trust John Howard over Kim Beazley on National Security by a margin of 56 percentage points to 28. They also preferred Howard on economic management by 62 percentage points to 24.
However, not everything is rosy. Over at the Sydney Morning Herald, their latest Herald Poll which is done for them by AC Nielsen, has the ALP leaping to a 4 percentage point lead over the government. Their results have 55 percent of respondents opposing the Iraq deployment, versus 37 percent who support.
The story attributes the decrease in the Liberals' vote partly to the troop deployment, but careful reading suggests this is sloppy journalism. Comments by Director of Research at AC Nielsen, John Stirton are reported in this way:
ACNielsen's research director, John Stirton, said: "Mr Beazley's approval rating is more or less steady, indicating that the improvement in Labor's standing is probably less a direct result of his performance than it is a reaction against the Government."He said the sharp turnaround in the Government's fortunes might have resulted from several factors. "There may have been a reaction to the mixed economic signals over the past few weeks.
"In particular, the interest rate rise may partly explain the decline in the Coalition vote among voters aged over 25." Support for the Coalition fell by 10 points among those aged 25-39 and seven points among the over 40s.
From this quote he obviously didn't ask voters why they had changed their mind, so any comments are pure speculation. And if he thought the troop deployment was the issue, then he would have said so.
Posted by Graham at 03:56 PM | Comments (3)