Blog CemeteryBreaking Down Newspoll - September 2007
September 25th 2007 09:26
So for what is likely to be the final time before the 2007 election, the good folk at Newspoll have aggregated three months of polling to provide a snapshot of geographical, demographic and seat status trends. Unsurprisingly the overall figures for the three months shows the ALP going to a crushing victory. But the extra information shows that the Coalition have lost large amounts of votes in a number of areas which would make an election, if reflected in these results, even more crushing, and leave the Coalition a decimated rump.
By providing various breakdowns of their national polling, Newspoll also provide the building blocks with which to divide the nation’s 150 seats into 36 separate categories. This is done by working out the relative strength of the swing since the 2004 election in each of the ten categories provided and then multiplying different factors by the national swing. The basic result of this process is shown below.
Boring Number stuff - skip past this if you wish
And below is two (extreme) examples of how these figures are reached:
Now these figures don't take into account the relative size of the categories created. As the larger categories (NSW, Vic, Safe Government seats and Metropolitan seats) also tend to be the categories with the strongest swings to the ALP in these figures, merely multiplying the factors overestimates the total swing to the ALP. By multiplying each of the swings by the number of seats that are within those particular categories a rough national swing appears. In the case of the figures above this rough national swing would be almost 9.9%, or over 1% above what the national swing should be. This can be fixed by dividing each swing by the percent difference between this rough national swing and the actual swing to provide the figures below:
OK, you can come back now...
So, we can add the swings in the table just above to the redistributed 2004 results in order to show what, had an election been held using Newspoll's figures from July to September, would have eventuated. And for the Coalition, the figures are catastrophically bad, with no less than 50 seats falling to the ALP, and the Coalition being reduced to a mere 38 seats.
This is more seats than would fall by a uniform 8.8% swing across the country (37 seats). The main reason for this is that the large swing in safe Government seats to the ALP. Whereas in a closer overall situation this could be dismissed as solidly Liberal voters letting off steam against a long-term government but in not enough numbers to alter the pendulum, the very strong position held by Labor means that a number of seats that in normal circumstances would never be under threat would, based on this current polling, fall to the ALP. Thus seats in areas which appear to be swinging especially strongly, such as North Sydney and Warringah in NSW, and Menzies, Aston and Flinders in Victoria, are currently polling small ALP leads.
As the swing is concentrated in the metropolitan seats, the damage for the federal coalition there would be especially strong. The federal Liberal party would be without representation in six of the eight capital cities, with the Government retaining only a handful of previously safe seats in Sydney and Perth.
A Silver Lining?
If this all appears a little too incredible to be true, solace may be found in the 2004 results. Similar polling results in July-August 2004 gave Mark Latham a 52-48 lead on TPP terms which was reversed in time for the election. Interestingly for these purposes, this swing to the Howard government was well stronger in safe Government seats than across the rest of the nation. On the other hand, safe ALP seats tended to move back to Labor in the campaign.
The upshot of these developments is that a repeat of the 2004 campaign (which I'll call Latham II) would deliver votes back to the Government in the seats in which they desperately need to shore up support. Thus the Government would win back all of their safe seats, and a large number of marginals. But they also show that even if the ALP manage to match the disastrous (in polling terms) 2004 campaign in 2007, Kevin Rudd will still win government.
Thus the latest Newspoll offers some comfort for both major parties. For the Coalition it suggests that a good number of their voters may return to the fold and safe them a number of safe seats, and possibly even the government, should the ALP struggle in the campaign. But for the ALP the omens look stronger - even a swing the measure of 2004 would not be enough to cost them government from here. More enticingly, 2007 offers the ALP the possibility of not merely taking power but seizing a massive majority.
By providing various breakdowns of their national polling, Newspoll also provide the building blocks with which to divide the nation’s 150 seats into 36 separate categories. This is done by working out the relative strength of the swing since the 2004 election in each of the ten categories provided and then multiplying different factors by the national swing. The basic result of this process is shown below.
Boring Number stuff - skip past this if you wish
And below is two (extreme) examples of how these figures are reached:
Now these figures don't take into account the relative size of the categories created. As the larger categories (NSW, Vic, Safe Government seats and Metropolitan seats) also tend to be the categories with the strongest swings to the ALP in these figures, merely multiplying the factors overestimates the total swing to the ALP. By multiplying each of the swings by the number of seats that are within those particular categories a rough national swing appears. In the case of the figures above this rough national swing would be almost 9.9%, or over 1% above what the national swing should be. This can be fixed by dividing each swing by the percent difference between this rough national swing and the actual swing to provide the figures below:
OK, you can come back now...
So, we can add the swings in the table just above to the redistributed 2004 results in order to show what, had an election been held using Newspoll's figures from July to September, would have eventuated. And for the Coalition, the figures are catastrophically bad, with no less than 50 seats falling to the ALP, and the Coalition being reduced to a mere 38 seats.
This is more seats than would fall by a uniform 8.8% swing across the country (37 seats). The main reason for this is that the large swing in safe Government seats to the ALP. Whereas in a closer overall situation this could be dismissed as solidly Liberal voters letting off steam against a long-term government but in not enough numbers to alter the pendulum, the very strong position held by Labor means that a number of seats that in normal circumstances would never be under threat would, based on this current polling, fall to the ALP. Thus seats in areas which appear to be swinging especially strongly, such as North Sydney and Warringah in NSW, and Menzies, Aston and Flinders in Victoria, are currently polling small ALP leads.
As the swing is concentrated in the metropolitan seats, the damage for the federal coalition there would be especially strong. The federal Liberal party would be without representation in six of the eight capital cities, with the Government retaining only a handful of previously safe seats in Sydney and Perth.
A Silver Lining?
If this all appears a little too incredible to be true, solace may be found in the 2004 results. Similar polling results in July-August 2004 gave Mark Latham a 52-48 lead on TPP terms which was reversed in time for the election. Interestingly for these purposes, this swing to the Howard government was well stronger in safe Government seats than across the rest of the nation. On the other hand, safe ALP seats tended to move back to Labor in the campaign.
The upshot of these developments is that a repeat of the 2004 campaign (which I'll call Latham II) would deliver votes back to the Government in the seats in which they desperately need to shore up support. Thus the Government would win back all of their safe seats, and a large number of marginals. But they also show that even if the ALP manage to match the disastrous (in polling terms) 2004 campaign in 2007, Kevin Rudd will still win government.
Thus the latest Newspoll offers some comfort for both major parties. For the Coalition it suggests that a good number of their voters may return to the fold and safe them a number of safe seats, and possibly even the government, should the ALP struggle in the campaign. But for the ALP the omens look stronger - even a swing the measure of 2004 would not be enough to cost them government from here. More enticingly, 2007 offers the ALP the possibility of not merely taking power but seizing a massive majority.
| 40 |
| Vote |
Subscribe to this blog
