Yesterday The Australian's analysis of the latest Victorian Newspoll suggested that the increase in the Victorian government's vote was due to the fires which have devastated the state.

The theory rests on the presumption that natural disasters unite the community, and that as the real and symbolic head of state, the premier or prime minister gains an advantage from that.

As the Victorian government's popularity had recovered by the end of last year (up from 51% 2PP to 57% 2PP between samples) and this year's results (60% 2PP) are just within sampling error that hypothesis looks a little shaky, however the popularity of Premier Brumby was up from 45% to 52% over the same period, so it is possible.

Which raises the issue of Cyclone Hamish, a once in a 100 years cyclone threatening the coast of Queensland right in the middle of an election campaign. If it devastates any part of Queensland, will it enhance Anna Bligh's chances?

Or will she be blamed, given that she has already been criticised for calling this early election while much of Queensland is still in flood?

Much may depend on whether Queensland's disaster services are seen as being up-to-scratch.

When I went to find the most effective TV ads from Labor and the LNP to road-test on our surveys they were nowhere to be seen. It's common political wisdom, and I think true, that positive ads don't change governments, negative ads do.

Positive ads are good at reinforcing the beliefs of the already converted, but don't motivate those who are swinging or swaying.

So the best two ads of this campaign are "Not him, not them, not now" from Labor, and the "Anna file" ad from the LibNats. You won't find "Not him, not them, not now" on Anna's site, but you will find generic "Labor file" versions of the "Anna file" on the LNP site.

BTW, someone needs to tell the webmaster at www.springborg.com that they've called the election. Seems like all the effort is going into the LNP site while the Springborg site is just another Mary Celeste on the interwebs.

Anyway, to help both sides out I thought I would put both the ads up here. I've been able to find "Not him etc." thanks to Malcolm Farnsworth, but only the generic "file" ad. The one that targets Bligh personally is very powerful and runs through some negative headlines on her career in education and treasury.

Tomorrow's Galaxy poll (with the teaser released today on the CM website) confirms our polling. It shows a six percent swing against the government. My polling suggests 7% or thereabouts. Pretty damn close, particularly as we don't really set-out to measure quants.

So, emboldened by this result I'm going to make a few more quant predictions. On my sample it looks as though the Gold Coast and Central Queensland will swing most against the government. Brisbane is looking at the average swing, Far North Queensland and the Sunshine Coast a bit less than average and areas west of Brisbane a little more.

I am not going to put figures to this because the sample sizes are just too small. But it will be interesting on election night, if these figures hold true in our polling, to see whether they translate into results.

In the first week of the election there has been a strong swing in sentiment against the government, somewhere in the vicinity of 7.5%. If this were reflected in the election result Queensland would most likely have a new government and a hung parliament.

Voters see this election as being about the government, not the opposition, and they appear to be punishing it for its governing of the state.

Our surveys are of people who are most likely to lead debate. They are people who write on blogs, ring talkback radio, write letters to the editor and talk politics to friends, relatives and neighbours.

While they may not be a randomised sample of the total population they provide a good lead as to where public opinion is heading.

In June last year they told us that the government was in trouble. On a two-party preferred basis they favoured the then Coalition by 52% to 48%. Since then there has been a 7% improvement in the coalition vote on virtually the same group of people.

(Note: most pollsters allocate preferences as they were allocated at the last election, we ask voters how they would allocate. These figures catch all those proposing to just vote one and not allocate a preference.)

This improvement in the two-party preferred vote is mirrored in the first preference vote, with the Liberal National Party (I have added the Liberal and National Party votes together before the merged entity came into being) first preference vote up 7% since June last year.

While Labor was down in January the lost voters were initially warehoused mostly with Greens and Independents. It was only after the election that they moved across to the LNP.

What has caused this movement?

It doesn't appear to be anything to do with Lawrence Springborg. His approval rating with this group has actually been declining at the same time his own party's vote is increasing.

Although not as quickly as Anna Bligh's approval rating.

Despite what they will personally do with their votes, our respondents are telling us that they still expect Labor to win.

Although they'd like the Coalition to win, but not by as much as their voting intentions suggest. A significant minority would like to see a hung parliament.

If Springborg hasn't caused this shift in sentiment, what has?

It's possible that calling an election is a large part of the problem as the table below showing the numbers more ore less likely to vote for the government as a result of calling the election early suggests.

But note how the net figure hasn't changed much, even though those on both sides of the proposition have.

Which suggests that the major reason is general dissatisfaction with the way the state is heading.

All of which means that the elction is currently framed for our voters as a vote of confidence or no-confidence in the government, rather than a vote for the government or the opposition. If it stays that way, and on the assumption that our respondents are ahead of the curve, there is a very good chance that the opposition will win.

I'll be posting results of the qual in the next couple of days.