A TV3 NFO poll was announced on TV tonight with the Greens best yet result in the crucial Party Vote. The poll was conducted over the past weekend, as Bill English boxed for life.
Party Vote Standings:
National 24% down 4
Greens Up 4.5 to 9%
NZ First 3.1%
Progressive Coalition –0.5%
In issue questions contained in the poll, 56% agree with the Greens that the moratorium on GE should be extended beyond October next year, and 38% approve of GE bottom-line position of the Greens.
In preferred PM stakes PM is still supreme, picked by 48% up 3 ponts.
Boxing has been good to Bill English’s performance ratings - In “Performed Well” category he is up 11 points to 46%. On “Lot of Personality” he is up 12%.
TV3’s news link for poll results. http://www.tv3.co.nz/news/index.cfm?news_category_id=11
The Poll Company NFO also has a web page for poll results. http://www.nfo.co.nz/index.html?/ins_poll.html
Finally, a Scoop reader has made a few observations about the weaknesses and dangers of inherent in polls. His comments provide useful context.
SCOOP READER FEEDBACK ON POLL WATCH
I am responding to your announcement: "Introducing Poll Watch
- Scoop will be running 'Poll Watch' as a new feature up till the election, whenever that is. As national poll results are announced we will report on the basic results and provide links to further resources. See... Poll Watch: Greens Rising Rapidly In NBR/NZ Herald  "
It would do everyone a favour if you were to background this by explaining to the public just how fickle these polls are, and how they have seldom come within a bulls roar of the real poll (the general election).
You might start by providing an explanation of the "margin of error" which is put in front of the public as a measure by which they should put faith in any poll.
Whether this is done in ignorance by news readers or in a spirit of self justification by pollsters, it is grossly misleading. It is many years since I last familiarised myself with the technical definitions of "margin of error", but as I recollect it, it is merely a measure of how many mistakes might have been made in the polling/recording/analytical process.
It most certainly does not mean that you are within 3% of the outcome of a general election. Those that sell these services would like you to think so. Indeed the results may be several hundred percentage points out and still be within the defined margin of error.
There is much anecdotal evidence of pollsters misunderstanding and/or mis-recording people's answers. Many of the questions are loaded, and above all: How can polling samples of 800 to 3500 people possibly reflect the diverse and constantly shifting opinions of millions of people in sixty-something electorates?
The danger I see in such polls is that they might unduly influence the voting of an overwhelmingly ovine voting population.
The incumbent Prime Minister almost always polls highest regardless of his/her performance or popularity. The pollster asks: "Who should be the Prime Minister", but the polled answer a different question and that is: "Who is the Prime minister?"
I hope that you can enlighten us all and provide a suitable context to your Poll Watch series.