Oz Politics Blog

With the latest Morgan poll my analysis of recent polling trends is returning to one of random noise. I do not think there has been any real underlying movement at all.

Nonetheless, the spin from the pollster was imaginative.

Earlier this week we released a Morgan Poll on Federal Voting Intention that reflected the electorate’s knee-jerk reaction to the interest rate rise. The ALP was ahead on both Primary support (ALP — 43%, L-NP — 42%, others 15%) and two-party preferred vote (ALP — 52% cf L-NP — 48%).

This latest Morgan Poll shows the L-NP regain its lead (L-NP — 52% cf ALP — 48%) and suggests last week’s ALP lead could not be sustained.

But frankly, if you believe that, I have a Harbour bridge you may be interested in. Just put all your cash in a brown paper bag, mail it to me, and this bridge is yours.

The Morgan poll was so out of step with Newspoll for the weekend of 5-6 March, that at least one of the polls had to be wrong. While I have not applied the appropriate statistical tests, it looks like the same story with the Morgan and ACNIelsen polls for the weekend of 12-13 March. Although this time I will run with Morgan, whereas on 5-6 March my vote was with Newspoll.

In that context, I think the only plausible conclusion is that the status quo has prevailed. Support for the Liberal and Labor parties is essentially unchanged since the last election. So far Beazley has achieved a dead cat bounce and little more. Of course, things could well be vastly different come the next Federal election in 2007. And as a week is a long time in politics, they could be different this time next week.

My graphs are here.

A birdie sent me some more how-to-vote cards. The HTV cards I published previously are here.

These additional how-to-vote cards increasingly persuade me that Labor should get over the line. (Albeit, these cards and the ones I published earlier could change when it comes to polling day).

The Poll Bludger has a bit of a wrap on the Werriwa by-election. Like me, he thinks a Labor victory is almost a certainty (if there can be such a thing in politics).

If today’s Australian is anything to go by, Labor is on message about how difficult it will be to win Werriwa (which as we all know is a dog whistle for Labor voters not to toy with minor parties and independents).

The SMH and the Australian are running a painkiller-overdose-hospitalisation story about Werriwa candidate Samir Bargshoon:

A candidate for the federal seat of Werriwa has admitted he overdosed on a cocktail of painkillers and anti-depressants, spending Thursday night in hospital.

Sam Bargshoon, a self-confessed ALP branchstacker who ran independently against former Labor leader Mark Latham five months ago, said he was depressed over the Orange Grove affair and the stress it had created with his family.

The south-west Sydney businessman and lobbyist for the doomed Orange Grove factory outlet, will run in Saturday’s by-election.

Mr Bargshoon said he took six Panadeine Forte tablets, four valiums and two anti-depressants on Thursday night to relieve a headache.

“Last night I was feeling a bit down and very depressed … there’s just a lot of stress,” he said.

“I didn’t actually know what I was going to do. I remember just waking up in hospital.”

His two young daughters found Mr Bargshoon unconscious on the couch and called for help.

He was taken to Liverpool Hospital and was told by staff he would not be discharged until Monday, however he negotiated to be released on Friday.

Mr Bargshoon said his depression had also been caused by the Labor Party’s treatment of him over the Orange Grove affair.

The results for the Werriwa by-election should be available from the Australian Electoral Commission beginning at 6pm Saturday.

Well it looks like I might be wrong. That which I thought was a blip yesterday is starting to look like a trend. However, I am curious as to why Morgan has decided to publish a statistic now — but more about that in a minute.

Morgan has the same headline result as ACNielsen (albeit for a different weekend, and on the same weekend as Newspoll did not find a significant change from previous polls). The predicted two party preferred result is Labor on 52 per cent and the Coalition on 48 percent.

However, Morgan’s primary vote differs a littler from that of ACNielsen. Morgan has the Coalition on 42 percent and Labor on 43 per cent. Of note, the error bars for Labor’s primary vote for
Morgan and Newspoll do not overlap, so at least one of these statistics is wrong. (It is quite possible the same is true of the two-party-preferred graph as the data points of one does not lie within the error bars of the other, but I could not be bothered doing the math to apply the appropriate statistical test).

I am intrigued about why Morgan decided to publish a statistic for the weekend of 5-6 March. Morgan normally runs its poll over two consecutive weekends with a sample of some 2000 voters. This is a single weekend poll with a sample of 1015 voters. Morgan decided to publish this statistic presumably after it polled the weekend of 12-13 March. In the normal course of events I had expected Morgan to publish its two weekend statistic this coming Saturday. I am curious.

Check out my updated poll graphs. (You may need to hit the refresh button on your browser to get the absolute latest graphs).

A kind correspondent sent me the how-to-vote cards for a few in that gang of 16. They are as follows:

If we assume that James Young will pick up the lion’s share of the Liberal vote from 2004, then the fact that he placed Labor’s Chris Hayes ahead of the Green’s Ben Raue on his how-to-vote card should be enough to ensure a Labor victory.

Update: The Socialist Equality Party is not providing how-to-vote cards. It is recommending people vote 1 for Mike Head, and it then advises voters to number every square from 2 to 16, but the party indicates no preference for how that is done.

Update #2: One of the candidates told me the how-to-vote cards from some candidates will differ for Saturday from the pre-poll cards currently being used. It has also been suggested to me that most of the independent candidates have decided to put Labor last on their how-to-vote cards for Saturday.

Today’s Fairfax press (here, here, here, here, and here) reported the ACNielsen poll from last weekend.

The headline result is that Labor would have won an election called last weekend with a two party preferred result of 52 per cent for Labor and 48 per cent for the Coalition.

Critically, the survey had the Coalition’s primary vote falling from six points from 47 per cent to 41 per cent. Labor’s primary vote rose 3 points to 39 per cent. The sample size was 1415 voters.

Howard’s approval rating was 54 per cent compared with Beazley’s 49 per cent. Howard is preferred Prime Minister by 50 per cent of voters compared with Beazley’s 40 per cent.

The Fairfax commentators are attributing the result to the Iraq deployment of 450 soldiers and the 25 basis point increase in interest rates. In respect of Iraq the ACNielsen poll asked some questions.

The poll, conducted at the weekend by ACNielsen, shows that Mr Howard’s shock announcement on February 22 that he had agreed to more than double the number of Australian troops on the ground in Iraq is a polarising issue.

His decision to break an oft-repeated promise not to significantly increase Australia’s contribution to the war effort was supported by 59 per cent of Coalition voters. But it was overwhelmingly opposed by everyone else. Just 20 per cent of Labor supporters agreed with the move, as did 14 per cent of Democrats and 15 per cent of Greens supporters.

Even among those who voted for One Nation, support was just 44 per cent.

The poll also highlighted a sharp division of opinion between men and women. Forty-four per cent of men supported the additional deployment; among women it was just 30 per cent.

Even among the Government’s own constituency, however, support for Mr Howard’s move is best described as soft. Just 10 per cent of all respondents - and 18 per cent of Coalition voters - identified themselves as “strongly” in support of Australia becoming further involved.

Thirty-four per cent - 12 per cent among Coalition voters - were “strongly” opposed.

More than one in three of those who supported this extra commitment of troops said they were opposed to any further increase. This apparent concern that Australia could become trapped in a cycle of escalation was far stronger among people old enough to recall the Vietnam War.

Only 10 per cent of respondents would unequivocally support any further commitment of forces, with another 19 per cent indicating they might, depending on circumstances.

Is this a blip or the beginning of a trend? The short answer is that we do not know for certain. No single poll result should be accepted uncritically. It must be interpreted in the context of the events of the day and other polls. Statistical theory predicts that around one in twenty polls will get the result wrong. At the moment, I suspect this poll will turn out to be blip. It is out of step with the latest Newspoll and Morgan polls. But we will need to wait for a while to see.