Those “election” data points appear to be almost exactly what the polls were forecasting.
So all the polling companies must have been in on the conspiracy with John Key, Warner Bros, Merrill Lynch, the Learned Elders of Zion, the Illuminati and the Electoral Commission to rig the election with the fake “result”.
Comment by Matthew Hooton — September 29, 2014 @ 9:22 am
I dunno. National estimates were kinda all over the place. And the polls again overestimate the Greens and underestimate NZ First. They’re pretty good at picking Labour though.
Comment by danylmc — September 29, 2014 @ 9:24 am
What I would really like to know is there an underestimation of Labour’s support while at the same time a depression in Labour’s vote because of a low turnout. This could also explain Green’s polling better than they scored. The turnout this election overall (unregistered and did not vote) was about the same as last time. The polls are still all over the place although for Labour they painted a depressingly consistent result.
Comment by Greg Presland — September 29, 2014 @ 9:29 am
That would be clearer if there were separate trendlines that excluded the elections.
Comment by James Butler (@j20r) — September 29, 2014 @ 9:31 am
Here you go: http://imgh.us/nzpollsresult.svg
Comment by danylmc — September 29, 2014 @ 9:38 am
One day the polls will shit the bed like they did in the last UK general election, or in the US 2000. In fact, they might have individually done that by the looks of some of the results. Confidence intervals don’t dictate that something is always right, only that something is right with a probability of less than 0.05 that there’s an error. And I would challenge that idea that they are that right to that standard all the time, because that P<0.05 is a standard used for controlled experiments, not polling in the field.
I don't know how polling companies can claim that level of confidence. Real world results suggest they can only predict a general trend, and not specifics like they claim. The closer you get to 50% the less the error margins are supposed to matter. Some polls are looking well outside of that ~3% at the outliers (Greens, NZF) and at the centre of the curve (National).
But then again, they'd all be out of a pointless job if they had to be honest about what they were doing.
Comment by Chris (@slackjawdtownie) — September 29, 2014 @ 9:47 am
I am now utterly confused by the continued underestimation of NZF. Winston didn’t have a tea-tape moment this time. The ‘big reveal’ by KDC and co wasn’t something Winston could tap into. NZF’s voter base is overwhelmingly old, white, and land-line owning.
My only remaining guess is that NZF benefits from a ‘reverse-Bradley effect’ where a whole bunch of people lie to polling companies and say they won’t vote (or will vote for someone else) and then vote for Winston anyway.
Comment by Phil — September 29, 2014 @ 9:48 am
One day the polls will shit the bed like they did … in the US 2000.
All the pre-election polls predicted Bush vs. Gore would be neck-&-neck and, shock horror, it was. The ‘shit the bed’ moment was solely down to a bunch of cable news networks all being desperate to call the result in Florida first. The problem wasn’t even exit polling per-se, it was a bunch of TV network executives being more interested in dick measurement than data interpretation.
Comment by Phil — September 29, 2014 @ 9:57 am
Maybe, contrary to the urban myth, the poll is accurate on Labour’s numbers because they do still have landlines.
Comment by whaleoil — September 29, 2014 @ 10:00 am
@9 Rancid is as rancid does. Fester in your own backyard.
Comment by paritutu — September 29, 2014 @ 10:09 am
Try this. It talks about bouncing polls now and in the 2000 general election.
Comment by Chris (@slackjawdtownie) — September 29, 2014 @ 10:11 am
Reblogged this on Talking Auckland and commented:
How the poll of polls panned out in the end
Comment by Ben Ross - Talking Auckland — September 29, 2014 @ 10:13 am
There is probably something to this. It would be interesting to know what proportion of voters vs non-voters had landlines (I suspect that it’s quite high given the turnout by demographic).
My only remaining guess is that NZF benefits from a ‘reverse-Bradley effect’
Given that NZF directly appeals to the old, white, scared, and befuddled, rather than lie I suspect their supporters are the mysterious ‘undecided’ who either make their mind up on the day or adhere to the notion of the privacy of political opinion so don’t feel comfortable talking to pollsters.
Comment by Gregor W — September 29, 2014 @ 10:20 am
Lots of NZ First voters tend to be soft Labour voters; I think they’re people who saw that Labour couldn’t win so switched to Winston Peters in the last week so that he could ‘keep National honest’.
Comment by danylmc — September 29, 2014 @ 10:27 am
I suspect that older voters are more likely to decline to speak to pollsters.
My take home is that the Herald Digipoll seems to be more accurate than the others (Roy Morgan in particular seems to be all over the place).
Comment by Brent — September 29, 2014 @ 10:32 am
@Brent – Quite the opposite, actually.
Comment by Andrew — September 29, 2014 @ 10:42 am
NZ First do better than in most polls, because their support base is skewed towards elderly voters, and elderly voters turnout in much higher proportions than younger voters. My 2c.
Comment by dpf — September 29, 2014 @ 10:54 am
I’m sure a significant percentage of NZF’s vote is people who dislike being on the ‘wrong side’ of majority opinion but still want to vote—Winston presents an attractive way of opting out and avoiding a loss. Another way is just not voting.
Comment by Ant — September 29, 2014 @ 11:00 am
Does Curia weight results to match the demographics of registered voters, or likely voters (i.e. the demographics of the last general election)?
Comment by Phil — September 29, 2014 @ 12:25 pm
Interesting is that as the polls became more frequent close to the election, for the major parties they became more accurate ( this time). From memory the previous election they werent so frequent. The effect of frequency seems to be the poll is taken over less days, while the normal monthly poll covers about 2 weeks.
Comment by ghostwhowalksnz — September 29, 2014 @ 1:09 pm
National estimates were kinda all over the place.
The variance of estimates of a proportion is higher the closer the proportion is to 0.5. So you’d expect more scatter for National. The bias for NZF and the Greens is harder to explain though.
Comment by Matt — September 29, 2014 @ 1:35 pm
Because poll results get such prominence, I wonder how many people vote according to the perceptions so formed. Thus because National was so high constantly, does it cause voters to give up on say Labour and look for safer harbours? A bit like being pointless to bet on the Round the World Yacht race because the positions are known and plotted so precisely and this blunts the excitement.
Comment by xianmac — September 29, 2014 @ 1:37 pm
I think the election was skewed because it didn’t include people who don’t have landlines.
Comment by Liam Hehir — September 29, 2014 @ 7:34 pm
@Liam – You’ve hit the nail on the head, right there. 🙂
Comment by Andrew — September 30, 2014 @ 5:22 pm
JOINING THE ILLUMINATI BRINGS YOU INTO THE
LIMELIGHT OF THE WORLD IN WHICH YOU LIVE IN
TODAY. YOUR FINANCIAL DIFFICULTIES ARE
BROUGHT TO AN END. WE SUPPORT YOU BOTH
FINANCIALLY AND MATERIALLY TO ENSURE YOU LIVE
A COMFORTABLE LIFE. IT DOES NOT MATTER WHICH
PART OF THE WORLD YOU LIVE IN.
FROM THE UNITED STATES DOWN TO THE MOST
REMOTE PART OF THE EARTH, WE BRING YOU ALL
BEING AN ILLITERATE OR A LITERATE IS NOT A
BARRIER TO BEING A MILLIONAIRE BETWEEN TODAY
AND THE NEXT TWO WEEKS.
YOU BEING IN THIS OUR OFFICIAL WEBSITE TODAY
SIGNIFIES THAT IT WAS ORDERED AND ARRANGED
BY THE GREAT LUCIFER THAT FROM NOW ON, YOU
ARE ABOUT TO BE THAT REAL AND INDEPENDENT
HUMAN YOU HAVE ALWAYS WISHED YOU WERE.
WE DON’T DISCRIMINATE IF YOU ARE WHITE OR
BLACK. SEND AN EMAIL TO
Comment by ILLUMINATIINITIATIONHOOD — October 10, 2014 @ 6:49 pm
RSS feed for comments on this post. TrackBack URI
Fill in your details below or click an icon to log in:
You are commenting using your WordPress.com account. ( Log Out / Change )
You are commenting using your Twitter account. ( Log Out / Change )
You are commenting using your Facebook account. ( Log Out / Change )
You are commenting using your Google+ account. ( Log Out / Change )
Connecting to %s
Notify me of new comments via email.
Notify me of new posts via email.
Follow me on Twitter (prepare to be disappointed)
Create a free website or blog at WordPress.com.