Kate SlackOpinion Polls
November 19th 2007 02:34
I thought it would be useful to look at all of the opinion polls at both national and state levels from 2007 in order to see if any pattern can be seen. What I've done is look at the state-by-state breakdown in terms of seats won based on each opinion poll. In each of the bigger states I've assumed each party has a number of 'safe' seats which I haven't moved regardless of the margin of the poll in question.
In each case Red seats are seats the ALP would have won, blue are Coalition won seats, grey seats are seats between 49-51% and therefore considered too close to call, while pink seats are the two seats expected to be retained by Independent candidates Tony Windsor and Bob Katter.
NATIONAL GRAPH
Here's the graph of every federal opinion poll taken this year, as well as the aggregation of the state-by-state polls often taken. I've also looked at marginal seat polling done across the nation, attempting to extrapolate an overall seat swing from this.
There are two central points here. Firstly, even if the Coalition won every seat considered too close to call, the ALP would have won 75 seats according to every opinion poll measured - an even 50% of the total House seats. Even with the advantage inherent in the election pendulum going into the 2007 election, which sees the ALP needing to poll over 52% to win power, this would suggest that the Coalition will need something big to go their way in order to hang on in the election.
The other major point is that the Coalition has made up ground as the year has gone on. In the first half of the year they endured three straight months in which the opinion polls showed the ALP winning over 100 seats, before recovering slightly in the winter, polling over 40 seats in each winter month. After dipping back to a 65 seat loss in a disastrous September, they have bounced back to just above their winter levels during the campaign itself, winning an average of 50 seats in November and holding the ALP to 75 seats in two polls. However even this recovery wouldn't be enough to safe the Government, with a majority of polls still showing the ALP looking like comfortably winning around 90 seats before even counting those considered too close to call.
We can also look at what seats the opinion polling suggests will fall.
NEW SOUTH WALES
Below is the projected results for each marginal seat in NSW at each poll conducted. Two things need to be noted here . Firstly, the safest 17 ALP seats, and 12 coalition seats, have been assumed to remain held by the party currently holding them. Secondly, the definition of marginal seats is skewed to include more Coalition held seats, in accordance with the data within the opinion polls.
ALL POLLS
Much of this data is based on the federal polls shown in the first graph, so it is no surprise to see that the Coalition has made up some ground after facing a humiliating defeat early in the year. North Sydney, Warringah and Macarthur look safe after looking decidedly shaky in the autumn, while the Coalition would appear to be ahead in Hughes and Gilmore. But the overall impression is that the Coalition will have to do better in their marginal seats and across New South Wales than they are doing in the polls as a whole to hold any real chance of hanging on in Bennelong, Dobell, Page, Paterson and Cowper, while Eden-Monaro, Lindsay, Wentworth and Parramatta would require very big local issues for the Coalition to have a hope. Even at the Coalition's best poll in November, they were still expected to lose three seats.
The next table looks at polling that has exclusively polled NSW or has provided distinctive results for the state, or a tract of marginals within it. While I haven't included individual seat polling the table does give a better impression of the effect of any statewide swing.
STATE ONLY POLLS
The story here is little better for the Coalition. While they have improved since the 2nd Quarter Newspoll figures, which would have seen the ALP win all 15 Coalition seats considered, the opinion polls suggest they look certain to lose three seats. There appears to be a divide between Hughes and Robertson, suggesting that while the Coalition should be fine in Hughes, Gilmore, North Sydney, Warringah and Macarthur, statewide polls suggest they are behind in ten seats they currently hold. The best hope for the Coalition is that the ten seats include ministers (Bennelong, Wentworth), are outside Sydney and therefore possibly underrepresented in statewide polling (Dobell, Cowper, Page), or both (Eden-Monaro, Robertson, Patterson). But while this may be able to safe some of the members, polling of this nature would suggest there's going to be a good swing to Labor and it will almost certainly include some of these seats.
VERDICT: Given the special nature of Wentworth, Bennelong and Eden-Monaro it is possible the Coalition can hold their losses to 2 seats. However looking at the polling it would be a surprise for the ALP to win less than 6 seats.
VICTORIA
In Victoria, the ALP are assumed to hold eleven safe seats, while the Coalition are assumed to hold seven. This leaves 19 seats to be analysed - 8 ALP and 11 Coalition held seats.
ALL POLLS
STATE ONLY POLLS
In many ways Victoria is the great hope for a Government looking to minimize their losses in all but Western Australia. With the Coalition's tightest marginal on a margin of around 5% they can lose votes without losing a seat.
Opinion polling which consistently showed the Coalition losing between 7 and 11 seats through the autumn have gently faded, with a handful of recent polls showing the Coalition within striking distance in even their tighest marginals. Through the campaign the seats of Gippsland, Higgins and Dunkley have appeared safe (or at least better than 50-50), while McEwen and La Trobe are too close to call in a number of recent polls. However federal polls would suggest Deakin, McMillan and Corangamite will be tough for the ALP to hold.
Focusing on state wide polls reveals a couple of conflicting stories. Firstly it must be noted that the final four polls were conducted by differing agencies around the same time and therefore cannot be looked at as a trend over time, but rather each agencies interpretations. Further, Newspoll's results in Victoria have been very volatile. the only statewide poll conducted in the third quarter of the year, Newspoll's cumulative polling found that the ALP would win nine seats in Victoria, taking Higgins and Kooyong while threatening Goldstein and Menzies. Newspoll's cumulative polling has continued to be volatile in Victoria, showing the Coalition a chance to lose zero seats in October before moving back to predicting large ALP gains in November.
VERDICT: The ALP look favourites in three seats and at least 50-50 in two others. However, any gains beyond McEwen would be a surprise. The polls would suggest at least three seats out of the most marginal five (Deakin, McMillan, Corangamite, La Trobe and McEwen) falling to Labor.
To be continued...
In each case Red seats are seats the ALP would have won, blue are Coalition won seats, grey seats are seats between 49-51% and therefore considered too close to call, while pink seats are the two seats expected to be retained by Independent candidates Tony Windsor and Bob Katter.
NATIONAL GRAPH
Here's the graph of every federal opinion poll taken this year, as well as the aggregation of the state-by-state polls often taken. I've also looked at marginal seat polling done across the nation, attempting to extrapolate an overall seat swing from this.
There are two central points here. Firstly, even if the Coalition won every seat considered too close to call, the ALP would have won 75 seats according to every opinion poll measured - an even 50% of the total House seats. Even with the advantage inherent in the election pendulum going into the 2007 election, which sees the ALP needing to poll over 52% to win power, this would suggest that the Coalition will need something big to go their way in order to hang on in the election.
The other major point is that the Coalition has made up ground as the year has gone on. In the first half of the year they endured three straight months in which the opinion polls showed the ALP winning over 100 seats, before recovering slightly in the winter, polling over 40 seats in each winter month. After dipping back to a 65 seat loss in a disastrous September, they have bounced back to just above their winter levels during the campaign itself, winning an average of 50 seats in November and holding the ALP to 75 seats in two polls. However even this recovery wouldn't be enough to safe the Government, with a majority of polls still showing the ALP looking like comfortably winning around 90 seats before even counting those considered too close to call.
We can also look at what seats the opinion polling suggests will fall.
NEW SOUTH WALES
Below is the projected results for each marginal seat in NSW at each poll conducted. Two things need to be noted here . Firstly, the safest 17 ALP seats, and 12 coalition seats, have been assumed to remain held by the party currently holding them. Secondly, the definition of marginal seats is skewed to include more Coalition held seats, in accordance with the data within the opinion polls.
ALL POLLS
Much of this data is based on the federal polls shown in the first graph, so it is no surprise to see that the Coalition has made up some ground after facing a humiliating defeat early in the year. North Sydney, Warringah and Macarthur look safe after looking decidedly shaky in the autumn, while the Coalition would appear to be ahead in Hughes and Gilmore. But the overall impression is that the Coalition will have to do better in their marginal seats and across New South Wales than they are doing in the polls as a whole to hold any real chance of hanging on in Bennelong, Dobell, Page, Paterson and Cowper, while Eden-Monaro, Lindsay, Wentworth and Parramatta would require very big local issues for the Coalition to have a hope. Even at the Coalition's best poll in November, they were still expected to lose three seats.
The next table looks at polling that has exclusively polled NSW or has provided distinctive results for the state, or a tract of marginals within it. While I haven't included individual seat polling the table does give a better impression of the effect of any statewide swing.
STATE ONLY POLLS
The story here is little better for the Coalition. While they have improved since the 2nd Quarter Newspoll figures, which would have seen the ALP win all 15 Coalition seats considered, the opinion polls suggest they look certain to lose three seats. There appears to be a divide between Hughes and Robertson, suggesting that while the Coalition should be fine in Hughes, Gilmore, North Sydney, Warringah and Macarthur, statewide polls suggest they are behind in ten seats they currently hold. The best hope for the Coalition is that the ten seats include ministers (Bennelong, Wentworth), are outside Sydney and therefore possibly underrepresented in statewide polling (Dobell, Cowper, Page), or both (Eden-Monaro, Robertson, Patterson). But while this may be able to safe some of the members, polling of this nature would suggest there's going to be a good swing to Labor and it will almost certainly include some of these seats.
VERDICT: Given the special nature of Wentworth, Bennelong and Eden-Monaro it is possible the Coalition can hold their losses to 2 seats. However looking at the polling it would be a surprise for the ALP to win less than 6 seats.
VICTORIA
In Victoria, the ALP are assumed to hold eleven safe seats, while the Coalition are assumed to hold seven. This leaves 19 seats to be analysed - 8 ALP and 11 Coalition held seats.
ALL POLLS
STATE ONLY POLLS
In many ways Victoria is the great hope for a Government looking to minimize their losses in all but Western Australia. With the Coalition's tightest marginal on a margin of around 5% they can lose votes without losing a seat.
Opinion polling which consistently showed the Coalition losing between 7 and 11 seats through the autumn have gently faded, with a handful of recent polls showing the Coalition within striking distance in even their tighest marginals. Through the campaign the seats of Gippsland, Higgins and Dunkley have appeared safe (or at least better than 50-50), while McEwen and La Trobe are too close to call in a number of recent polls. However federal polls would suggest Deakin, McMillan and Corangamite will be tough for the ALP to hold.
Focusing on state wide polls reveals a couple of conflicting stories. Firstly it must be noted that the final four polls were conducted by differing agencies around the same time and therefore cannot be looked at as a trend over time, but rather each agencies interpretations. Further, Newspoll's results in Victoria have been very volatile. the only statewide poll conducted in the third quarter of the year, Newspoll's cumulative polling found that the ALP would win nine seats in Victoria, taking Higgins and Kooyong while threatening Goldstein and Menzies. Newspoll's cumulative polling has continued to be volatile in Victoria, showing the Coalition a chance to lose zero seats in October before moving back to predicting large ALP gains in November.
VERDICT: The ALP look favourites in three seats and at least 50-50 in two others. However, any gains beyond McEwen would be a surprise. The polls would suggest at least three seats out of the most marginal five (Deakin, McMillan, Corangamite, La Trobe and McEwen) falling to Labor.
To be continued...
| 81 |
| Vote |
subscribe to this blog
