Roy Morgan has released results from a face-to-face survey of 1723 respondents conducted over the past two weekends, which has Labor’s primary vote down from 51 per cent to 48 per cent from the previous such poll a fortnight ago (which had only 909 respondents) and the Coalition’s down from 38 per cent to 37 per cent. Labor’s two-party lead has increased from 58-42 to 58.5-41.5 based on how respondents said they would vote; on the basis of how preferences went in 2004, it has narrowed from 58-42 to 57.5-42.5. The face-to-face results are distinct from the two smaller-sample telephone surveys conducted by Morgan in the interim, both of which had Labor leading 55-45.
What are you on about?
A poll is an estimate of a population value, with a margin of error. Galaxy provided that, and recent polls suggest 44-42 is probably (just) within the margin of error. News Limited interpreted it as they are wont too. Nothing to do with the polling organisation.
What does Galaxy have to excuse? Why aren’t you demanding Newspoll apologise for their (probably) even more misleading 52-35 a week earlier?
Leopold, what are you on about? Are you saying Galaxy are as out of step as Newspoll were a week earlier? If so you’re right. Rogue polls my friend.
“the excuse Galaxy need to provide and to take back the misleading headlines they created and wanted.”
Could you explain that sentence? My impression is you are saying a) Galaxy is intentionally biased, b) they should apologise (‘provide…excuse’) in some way for their poll and News Ltd’s headlines associated with it.
All the opinion polls have the Coalition’s primary vote around 40% at the moment, say if the Greens were to win 7% of the vote, Family First 4% and others 4-5%. If an election here held anytime soon like August, the Coalition would be defeated, possibly as bad as they were in 1983.
The Coalition would have to score around 44% to win the election. It remains to be seen if the Coalition can gain that extra 4% of the vote between now and the election in Late October/Early November.