Bugger the polls

141 comments on “Bugger the polls ”

• Comments are now closed

1. Blue 1

   28 November 2011 at 10:05 am

   I’ve been saying for weeks that if National got 50% or over I’d eat my hat.

   Thankfully the hat is safe now 😀

   You can’t really blame a lazy media for making a meal out of polls, which are basically a story for no effort on their part. Well, unless you think they should be making an effort to do their jobs.

   The yawning chasm between those whacked-out polls that had National at 56% the week before the election really need looking at. The 8% gap is way outside the posted margin of error.

   Personally I think they need to have larger sample sizes if they insist on continuing with landline phone polling.
   • Deadly_NZ 1.1

     28 November 2011 at 1:27 pm

     “You can’t really blame a lazy media for making a meal out of the polls” Lazy??? More like Sycophantic. My god I wouldn’t have been surprised if he’d been caught getting a sneaky blowie from some appreciative media person. And his popularity rating would have climbed. SO now here we are again with 3 years of asset stripping, higher Unemployment, (A BILLION Bucks from the public service (called savings)) Lower wages, Oh joy it’s not going to be a nice 3 years, and I hate to think what joys he has for us now X sickness and other beneficiary’s.

     You should have a poll and see how many are still feeling secure, in the cold light of day when they finally realise what they have done, and how they are going to be repaid.

     And what happened in Christchurch ? A dead heat now who would have picked that one?
2. gingercrush 2

   28 November 2011 at 10:07 am

   In fairness both NZ Herald Digipoll and Roy Morgan picked up the momentum changes. Both showed the rise of NZ First and over the threshold. Roy Morgan had NZ First at 6.5%. Roy Morgan crucially had National below 50% and Labour’s vote was very close as well. I think everyone anticipated that the Greens weren’t going to poll what they were and that was accounted for. The most spinning I’ve seen is from Horizon which is doing all sorts of justifications for why their results stated National so low.

   I would personally like to see some consolidation of polling with maybe a partnership between TVNZ and one of the Newspapers and the same with TV3. That is done in Australia. I also see a need for a tracking poll. Tracking polls are not necessarily anymore accurate but they tend to be good at showing momentum and change.

   Edit: Electorate polling on the other hand surely has been shown to be troubling. I wonder if its the same affect as 2005 and the mayoralty race in Auckland. When its essentially between two people or in 2005’s case two parties (ignoring the minor parties for a second or two). They can shown incredible difference to what the results actually were.
   • lprent 2.1

     28 November 2011 at 10:36 am

     A few days before the election they were quite incorrect.

     A week before the election what were they showing? Something that was quite different from the result – differences that far exceed the error rates. Why – because they only show people they managed to contact and not those they failed to contact.

     Basically you’d have to say that chicken entrails and a ear to the political ground would have been more accurate. Quite simply the only thing that polls are good for is to show momentum. Because the pollsters don’t publish either their methodology or their actual numbers (like the various failure to contact rates or in most cases the refused to answer rates) they are essentially useless.

     But needless to say the lazy commentators insist on using the inadequate inaccurate things as being gospel for prediction numbers.

     Polls these days are essentially useless because they don’t sample randomly. They sample listed land-lines of those who are home at the right time and are willing to answer a poll – a group that you’d have to characterize as being rather strange in NZ urban society these days. The no-contact rate is enormous. The refusal levels are high.

     I think it is time to ban them from the public discourse. Either that or require that polling companies publish their whole methodology and results so it can be analyzed for the real error rates.
     • wtl 2.1.1

       28 November 2011 at 11:43 am

       I agree re: banning. I’m pretty sure other countries do so during the election period (France maybe?). The led up to this election has shown how poll-focussed the political discourse in this country has become, which isn’t a good thing as it moves the focus away from discussing advantages/disadvantages of various policies etc.

       The only real argument for publishing poll results is to aid tactical voting, but our voting system can be tweaked to make move tactical voting unnecessary.
       • Draco T Bastard 2.1.1.1

         28 November 2011 at 5:12 pm

         I agree re: banning.

         We could make it slightly more complex. Have it so that polls that ask about candidate/party support are banned but polls that ask about issues and policies are allowed.

         The only real argument for publishing poll results is to aid tactical voting, but our voting system can be tweaked to make move tactical voting unnecessary.

         Tactical voting shouldn’t be necessary. That’s one reason why, while we keep a representative system, that I think electorates should be done away with (I mean, really, how many politicians live outside of their electorate anyway?) and go full proportional. Local conditions can be addressed at the local government level.

         EDIT:
         Here is a study on countries that ban polling.
     • Ari 2.1.2

       28 November 2011 at 11:26 pm

       Disclosing methodology and contact rate statistics, more information about undecided voters, etc… would be very useful, especially a “why are you going to vote for X?” type question.

       Polls have a place in election coverage, but it ought to be a low-key one and be less frequently published. A ban on releasing polls to the public during say the month before the election could be very helpful.
       • lprent 2.1.2.1

         29 November 2011 at 12:14 am

         Outright bans are difficult because they’d run directly into BORA. Someone (Draco?) put in a link to a 2003 paper that is worth reading. Methodology looks more effective..
3. Jimmie 3

   28 November 2011 at 10:13 am

   You could also take the message that the polls got Labour’s support just about bang on – no more harping on about cell phones vs land lines.
   • Lanthanide 3.1

     28 November 2011 at 10:16 am

     Yeah, I read something on the horizon site about a survey they did for land lines vs cell phones, that they used as the basis for why their internet methodology was so much better. Their results to me looked like those using cellphone only were spread fairly evenly across the income bands, meaning the actual cellphone-only bias should be fairly balanced and easy to ignore. Roy Morgan have said that they actually do include a small sample of cell phones in their polling, as well.

     I think now, we can all agree to ignore Horizon.
     • felix 3.1.1

       28 November 2011 at 11:27 am

       Where did Horizon have NZFirst and the Greens?
       • Lanthanide 3.1.1.1

         28 November 2011 at 12:26 pm

         Here’s their last forecast: http://www.horizonpoll.co.nz/page/188/6-3-more-could-vote-no-party-governs-alone

         Here’s a slightly older one: http://www.horizonpoll.co.nz/page/180/nz-first-10-

         Here’s their rather bizarre assertion that they were “close” to the final results: http://www.horizonpoll.co.nz/page/192/election-results-close-to-forecasts

         It seems they’re claiming the percentages are “of total eligible 18+ population”, so they claim that their forecast of National at 33.7% was therefore close to the final outcome of 29.2%. But that means their other figures for NZFirst at 11.2% and Conservatives on 5.3% were wildly off. If 29.2% of all eligible 18+ voters gave National 47.99%, it means that 11.2% of all voters would’ve given NZFirst 18.4% (11.2 / 29.2 * 47.99) and Conservatives 8.7%.

         Using the same extrapolation, they had the Greens at 20.3% and Act at 4.43%.

         In this case Horizon have tried to use a strange statistic that no other polling company uses and cherry-pick what parts of the results they show have been predictive, while completely ignoring the rest.
         • Ari 3.1.1.1.1

           28 November 2011 at 11:27 pm

           Yep, Horizon is clearly even worse than traditional polling with their current methodology, and should be having a re-think, not spinning their figures.
     • Olwyn 3.1.2

       28 November 2011 at 11:34 am

       Although if you look at the actual percentage that voted for Nation, including the non-voters who amount to something like 27% and 32%, it is closer to 30% that 48%, which is roughly where Horizon had them. Horizon, however, did overestimate the vote for other parties, so their weakness may lie in the opposite direction than the other pollsters – pressing people to say who they would vote for if they did vote does not take into account the possibility that they may not vote at all.
     • Deadly_NZ 3.1.3

       28 November 2011 at 1:35 pm

       Yeah but they give stuff away so it’s only time and very little effort. I am more inclined to hang up on phone pollsters ( the forced happiness gets me ) but offer me a chance to win a grand or an I pod for 5 mins of my time. I’m in, and I was.
   • RedLogix 3.2

     28 November 2011 at 10:18 am

     Or you might want to consider that 35% of the registered voters, that’s almost a million people, have either not voted or are specials that have not yet had their vote counted.

     This is potentially a significant effect that the polls don’t pick up.
     • Lanthanide 3.2.1

       28 November 2011 at 10:21 am

       I wish all those people had voted for Aotearoa Legalise Cannabis :/
     • Vicky32 3.2.2

       28 November 2011 at 3:55 pm

       or are specials that have not yet had their vote counted.

       Does anyone know how many specials there are this time? My son, daughter in law and niece all had to cast special votes as their changes of address had not been registered (although they notified them in plenty of time!) I heard this morning that the specials won’t be counted until December 10th. Shocking! (All 3 voted Labour – so therefore I have some hope.) Being young people they will have been tarred as apathetic non-voters – not so!
   • lprent 3.3

     28 November 2011 at 10:41 am

     Jimmie: you are a bullshit artist. What was being said was that the National support was over represented – you know the 56% a week before the election that was splashed all over TV. Because it was being advertised as being a foregone conclusion people didn’t vote

     What did National get? 48%. I wonder what happened to the other 8%?

     What was the turnout. Massively down, much of it from the left.

     Now have a look at the polls again you idiot, and this time use your brain….

     I guess that the polls should be banned as their inaccuracies appear to be affecting the election results.
     • Gosman 3.3.1

       28 November 2011 at 11:27 am

       Hang on here. Are you trying to state that a Poll showing a likely National majority government will impact on the support for the the left or the Right? If it is the left what is the logic behind this? Surely this should galvinise the left to get out the vote to avoid the situation ever coming to fruition.
       • lprent 3.3.1.1

         28 November 2011 at 11:38 am

         That is exactly what I am saying – inaccurate polls are dangerous for the democratic process (regardless of which party they favour).

         You’re being disingenuous if you’re trying to say that having grossly inaccurate polls published week after week and daily during the last weeks doesn’t have an effect on voting patterns.

         We live in an age of mass media. In the mass media you find that people tend to see the lead stories more than others.

         Go back on the front page of the herald for the last 6 months and count how many stories there are about poll results compared to Labour policies or Nationals policies. I’d bet that poll results are more than National’s policies and a order of magnitude more than Labours. This is in a period when Labour has been releasing quite innovative and new policy. Same for TV. Policy gets washed out by cheap polling headlines.

         So your advice to overcome the cheap publicity engendered by lazy journalism is for Labour supporters to contact people one by one? Yeah right.

         Polls aren’t part of the democratic system. They are now getting in the way of it operating properly. I don’t think that they can be fixed as a predictive mechanism.

         Nope I think that banning polls all together as being grossly inaccurate is the most effective solution..
         • McFlock 3.3.1.1.1

           28 November 2011 at 12:10 pm

           The classic example was the “worm” in 2002. 
            
           At least this time Winston’s exposure was due to the good old fashioned beat up of a story, rather than inventing something to go berserk over.
         • Gosman 3.3.1.1.2

           28 November 2011 at 12:48 pm

           I’d suggest this is your opinion only and one that doesn’t have much in the way of empirical evidence to back it up. Indeed I know of the opposite effect being mentioned in elections in the past where Political Parties well out in the lead in the Polls were worried they would fail to get the support on election day because their support base would think it was a done deal and they didn’t need to bother to vote. Now this is the polar opposite to the effect you mention, one that I have never seen discussed before.

           You seem to be using the fact that the National party had consistently high support in opinion polls as an excuse for Labour’s loss when the reality is that one term Governments in NZ are rare, especially on the right of the political spectrum. There wasn’t much Labour or the left could have done to swing this one unless they pulled something miraculously from the campaign bag of tricks. This they failed to do. It had little to do with what the Polls stated prior to the election.
           • McFlock 3.3.1.1.2.1

             28 November 2011 at 1:15 pm

             Um, no – people seem to be suggesting that there is a systemic bias in the polls which makes them useless, and can skew the vote. Nobody has stated that this was the sole, or even main, contributing factor to labour’s returns at the election booth.
              
             The fact is that the polls consistently overestimated national’s support, seemed okay-ish for labour, but were quite a bit off for the smaller parties. Which kind’ve suggests that the orgasms the media were having over each almost daily poll result, losing no opportunity to stroke their touch screens, was all a complete waste of time.
             • lprent …

               28 November 2011 at 2:32 pm

               Pretty much my point. The journo’s appear to use polls as cheap political filler. But they are of dubious value in the political debates normally because they are so inaccurate and the media keep the details opaque. They are outright harmful in ultra-short election campaigns like this one wound up as.
           • lprent 3.3.1.1.2.2

             28 November 2011 at 2:51 pm

             I suspect you’re missing the point (deliberately?). Work it the other way around.

             What value do the polls have for the political process?

             I’d suggest that because they are so inaccurate they have little to offer apart from cheap media filler.

             We just had 4 weeks worth of an election campaign where the polls probably accounted for at least a third of the media news coverage of the election (and probably closer to a half). Those polls were either essentially meaningless in political terms or harmful in that they dominated the political debates that people saw on mass media.

             But that media discussion was quite inaccurate because it was well outside the usual expected error levels. They were predicting results that weren’t even close to the eventual results and that virtually every political commentator was predicting that at least one part of which (the mid-50’s National figure) would never happen.

             It is in exactly line with what was predicted by the people like myself who’d looked at the issue of systematic inaccuracies of polling methodologies on the basis of far far larger samples done directly on the electoral roll. Specifically the ever reducing numbers of listed landlines and the demographics of who had the remaining ones.

             So remind me again – what is the value of the highly inaccurate polls?
             • Gosman …

               28 November 2011 at 3:10 pm

               “So remind me again – what is the value of the highly inaccurate polls?”

               You claim they are highly inaccurate yet I am arguing they were surprising accurate meassures of both trends and outcomes.

               As for their use well for a start any information on the mood of a population is useful in my view. It allows you to guage impacts of various actions. For example the Teapot tapes sage. Without the polling information National was getting on the topic the only driver in the discussion would have been the media’s response. As it was National had advice that people thought it was all much ado about nothing and this led to the PM call to get back to policy while ignoring calls to release the tapes.

               It also allows voters in MMP to make decisions over who they give their party vote for in an election so that they can be more confident they don’t waste them. This probably led to more people voting for NZ First as the polls showed over the last two weeks that he was likely to get over the 5% threshold. It also probably lessened ACT’s party vote as people were unsure that Bank’s would win Epsom. I think this is quite a useful service myself. I hate to see my vote going to waste, don’t you?
               • lprent …

                 28 November 2011 at 3:49 pm

                 So having a poll saying that National is going to get 55-56% a week out from the election is “accurate” when they actually got 48%.

                 That over the course of a week the same polling company dropped almost 5% is “accurate”. When they’d had the figure for National at that 55% sort of level for months.

                 I’d suspect that there was a systematic sampling error somewhere. But I can’t tell because the information to evaluate the accuracy of the technique is conveniently unavailable.

                 I guess you have a different idea of what is “accurate” than I do. And I’d say that you idea of “accurate” would make me to never give you accounts to control, programs to write, or hardware to maintain. Salesman are you?
         • Deadly_NZ 3.3.1.1.3

           28 November 2011 at 1:41 pm

           But if you ban all polls then how are you A: Going to know where you are in relation to your opponents. And B: How are you going to let your supporters know whom to vote for. Therefore inaccurate or not Polls seem to be needed. Sort of a necessary evil.
           • lprent 3.3.1.1.3.1

             28 November 2011 at 2:25 pm

             A. All parties poll already.

             B. WTF? The polls don’t tell anyone that.

             The polls that are done by the media are inaccurate and don’t give the full information about the poll figures and lack any significant information about methodology which means they are not transparent. Therefore they distort the political debate with consistently inaccurate information. Quite simply they are pretty much complete bullshit except in terms of overall trend.

             I fail to see any particular point of use to them. Tell me why you do (but please please use your frigging brain first).
             • Gosman …

               28 November 2011 at 3:17 pm

               The only supporting evidence you supply that the polls are inaccurate is that in the outcome NZ First got more of the party vote and National less than the average of polls over the preceding period of weeks and over the past three years. In every other criteria they were remarkably accurate and they also picked up on the trend where National was falling back towards 50% and below and NZ First was on the rise above 5% in the past two weeks. That is accurate in my book.

               I also note noone has commented on the Ipredict market which has proven to be remarkably good at predicting the results. I believe National was sitting at a likely 48-49% on Ipredict during the last week. Would you ban Ipredict as well during the last part of the election?
               • lprent …

                 28 November 2011 at 3:45 pm

                 iPredict is quite open about their methodology and its weaknesses – unlike the polling companies. That methodolgy has certain flaws as well (the classic insider trading ones and small market manipulation) that showed up in various ‘stocks’ at different points, and as you say it has proved to be more accurate especially further out from the election. They also don’t claim to any kind of ridiculously over-hyped “scientific” basis. Effectively it is reporting the aggregate opinion of a number of people about the aggregate opinion on the actions of other people. It is opinion….

                 You’d probably have noticed that with the exception of the Stratos hour, a few blogs, and the continuous promitional e-mails from Matthew Hooten they weren’t widely reported in the media? Probably because the figures didn’t make a good story because they were too mundane.
   • Peter 3.4

     28 November 2011 at 11:32 am

     I disagree.

     Landline factor is significant, it’s just that the landline factor decided not to vote 🙂
   • Colonial Viper 3.5

     28 November 2011 at 12:18 pm

     Landline polls predicted landline voters intentions, and with a million people not voting that’s what we got.

     If the youth (under 25) vote had turned out, or turnout had suprassed 80%, the polls would look as outrageously skewed as they are.
     • Gosman 3.5.1

       28 November 2011 at 1:10 pm

       Yes and IF my Aunt was born with testicles she would be my Uncle. The polls were broadly accurate even in predicting the number of undecided’s (that in the end didn’t vote it seems).
       • Ari 3.5.1.1

         28 November 2011 at 11:34 pm

         I’m not sure how you can call a consistent error of four times their margin accross all landline-based polls for a major party “broadly accurate”. Next you’ll be trotting out the joke about the statisticians going hunting.
4. Lanthanide 4

   28 November 2011 at 10:14 am

   The only real big differences here are National and NZFirst. The Greens and Labour were about right.

   Adding the numbers up for poll of polls: 97.6% voted for the listed parties. Actual result: 99.34%. Taking out Conservatives and Mana drops it to 95.58%.
   • felix 4.1

     28 November 2011 at 1:07 pm

     Good point.

     If a bunch of Nat voters decided to vote for Winston at the last minute (which wouldn’t be surprising given the horrible things Key is presumed to have said about old people on the teapot tapes) then that’d just about square everything up.
5. Ms X 5

   28 November 2011 at 10:14 am

   I think polls do shape the voting patterns, and while most of them rely on landline calling – if not all – what does that say? That people with cellphones only don’t vote? Are they the missing 40%?
   Personally, I’d like to see the ban mentioned on here yesterday, for a couple of weeks before the election. If people are going to vote, they should not be deciding by what others think. And yes, we do have a lazy news media – how often do we see stuff that might be going to happen, hasn’t happened but could…. that’s NOT news!
6. Brendon O'Connor 6

   28 November 2011 at 10:20 am

   What is this modern day obsession with polls who phone 1000 or less phone landlines and those that dont tell them to bugger off, ask them who they are going to vote for. Apart from the land lines flaw, they never say how many calls get hung up on and they seem to all ignore the fact that they don’t ignore the people who say they are not going to vote which seems to be about 30% of the eligable voters, which means that even the closest polls were out by 30%, which is a massive margin of error.
   Is there anyone out there who has half a brain who actually answers phone polls? I personally don’t know anyone who admits to not hanging up the phone on pollsters, especially as they all phone you at dinner time.
   • Vicky32 6.1

     28 November 2011 at 4:07 pm

     I personally don’t know anyone who admits to not hanging up the phone on pollsters, especially as they all phone you at dinner time.

     I live in a state housing area (in a State house – er dump…) and in 15 years here, have been phoned exactly once for a political poll… (It shows the power if stereotypes that I am constantly having to tell Sky TV to bleep off)… Oh, no, I remember now, I  was phoned twice – the first time, I got as far as expressing my preference – supervisor comes on and tells me “We won’t be using your result – we have already got enough in your demographic…” Pretty blatant! The pollsters don’t want to hear from people in poor areas. When they do call one, and get an old lady, they’re fine with that, until they find out that the old lady is a leftie… BUGGER THE POLLS ALL RIGHT! 😀
7. Carol 7

   28 November 2011 at 10:22 am

   I support a ban on polling throughout the election period.

   The problem with the apparent accuracy of the Labour and Green poll stats, is that it may have been a self-fulfilling prophesy, with many potential voters feeling they these parties were a lost cause – hence the low voter turn-out.

   But also the amount of time the MSM spends on looking at polls is indicative of very lazy journalism – that and the way they focus on “the game”. They treat politics as if it were a politics idol game show, complete with undemocratic phone-ins and online comments & digipolls that get flooded with right-wing astroturfers.

   Instead they should be encouraging critical debate of all policies and political activity, including a much more in-depth debate of government & opposition performance.

   The MSM are like poll junkies and it’s time, and in the interest of the country they they get over their addiction. Maybe they should go cold turkey for a while, and see if they can find a different focus for their political coverage.
   • Gosman 7.1

     28 November 2011 at 10:25 am

     As far as I am concerned the media was focused on the policy debate, at least until the Teapot tape intervened. The media generally treated Labour’s policy focused campaign very well. That is why every commentator I have read or heard, (including those on the right), have stated that Phil Goff ran a good campaign. Name me one thing you think the media should have focused more on other than perhaps not getting caught up in the distraction of the Teapot tapes?
     • rosy 7.1.1

       28 November 2011 at 10:34 am

       It’s not often I agree with you Gosman, but yeah. And that’s why the teapot tape derailed Labour’s campaign more than it derailed NActs because Labour had set the policy agenda and the MSM had picked up on it. It gave National to take the focus off policy and no-shows, and re-set the narrative to gossip and innuendo.
     • kriswgtn 7.1.2

       28 November 2011 at 11:37 am

       Nationals lack of policy
       • Gosman 7.1.2.1

         28 November 2011 at 11:56 am

         The cornerstone of Labour’s campaign though was attacking a key part of the National party’s policy – Part selling of certain SOE’s. It is obviously illogical to state that the Media should have focused on a lack of policy on National’s side when Labour was drawing attention to perceived dangers of it.

         Now I agree that National didn’t have much in the way of new policy but I suspect that was a deliberate ploy on their part to make the case that it would be pretty much business as usual. This obviously worked for them.

         It is also difficult to see how the media could criticise this when National’s obvious response is that it doesn’t want to radically change things from where they were before the election. What’s the media to do then?
     • Hanswurst 7.1.3

       28 November 2011 at 2:50 pm

       I disagree. There was a lot of reporting that certain policies had been announced. However, any analysis was largely focussed on how those policies would affect polling, rather than how they would affect the country in the event that they were enacted. Policy seems to be presented as a sort of makeup that parties occasionally put on to make themselves more sexy, rather than a tool that they use to get their important work done.
8. Gosman 8

   28 November 2011 at 10:22 am

   Come on here. Other than overestimating National and underestimating NZ First the polls were pretty much spot on. What was also clear from polling was there was a move away from National to NZ First in the last two weeks. This was reflected in the election result. You can hardly state there is something seriously wrong with the polls when they generally got trends and even results right.
   • lprent 8.1

     28 November 2011 at 10:45 am

     So that explains why the polls throughout that whole of the last three years didn’t even get close to predicting the actual results.

     A two week trend? And they were still way out.

     Don’t make me laugh… You have just defined why the polls are useless in the political arena.. They carry a systematic set of biases
     • Gosman 8.1.1

       28 November 2011 at 11:19 am

       Why would polls taken over a three year period somehow reflect the reality of what happens on one particular day in the future? All they can really do is indicate respective positions and trends. This is what they were very good at predicting.

       The polls showed National well in the lead. They were on Election night. The polls showed the left split between Labour and the Greens with Labour being the bigger party but the Greens picking up a lot of support. This was the case. The polls showed all of the minor parties other than NZ First at or below the margin of error which they were.

       The only thing they polls didn’t reflect accurately was the eventual level of support for NZ First. However I would suggest this was a result of the last two weeks of the campaign rather than something that the Polls missed over the past three years.

       By the way noone has mentioned how accurate Ipredict was proven. I remember people on here claiming some bizarre right wing bias in this that meant it was a complete joke.
9. brett Dale 9

   28 November 2011 at 10:31 am

   Werent the polls saying the greens would get 13%?
   • lprent 9.1

     28 November 2011 at 10:45 am

     15% at its peak. The poll of polls is an weighted average
10. Richard Down South 10

    28 November 2011 at 10:35 am

    Asking too many people driving $100k cars imo
     
11. queenstfarmer 11

    28 November 2011 at 11:00 am

    But over time, averaging polls should (if their methodology is sound) give us an accurate picture.

    This makes no sense. Parties release policies over time, debates are won and lost, and various scandals and other unexpected events unfold. Why do you think that an average of polls taken before and after such events should magically given an “accurate picture”?
    • RedLogix 11.1

      28 November 2011 at 11:13 am

      Well exactly.

      Averaging lots of polls effectively reduces their noise and improves their long-term accuracy, but introduces a delay which degrades their short-term responsiveness to events.

      Fundamental statistics makes it very hard to have both. (Although not impossible… just that the methods involved are way more sophisticated than these companies are using at present.)
    • Lanthanide 11.2

      28 November 2011 at 11:23 am

      Because barring exceptional circumstances (tea parties, earthquakes), the core support for any specific party does not change much, whereas the polls themselves will statistically fluctuate because they use a different sample each week. Averaging lets us remove the statistical noise and focus on the real underlying figures, which are what matters.

      The problem with averaging of course is that it obviously flattens out everything: noise and genuine poll movements. But the idea is that the noise makes up a bigger proportion of the fluctuations, so genuine poll movements shouldn’t be negatively impacted too much.
    • Craig Glen Eden 11.3

      28 November 2011 at 11:34 am

      Good point QF which kind of points to”bugger the polls” then aye?
12. Jackal 12

    28 November 2011 at 11:17 am

    Polling or propaganda?

    There’s never been an election campaign in New Zealand with so much polling. Nearly every day we had hacks like Duncan Garner saying that the polls were putting National into a commanding lead while all other social indicators were telling a different story…
13. marco 13

    28 November 2011 at 11:18 am

    It looks like the missing Nat vote was picked up by NZ First in the last days of the campaign. Probably several reasons for that but mainly their increased profile over the last two weeks of the campaign. Winston was everywhere.

    I’m pleased to see him back in parliament. Whilst I don’t agree with his politics he is a capable performer who would destroy Key in a one on one debate. His resurgence has, in my opinion, strengthened the opposition benches.
    • Jackal 13.1

      28 November 2011 at 11:25 am

      So effectively we had the mainstream media choosing who get’s into parliament by how much airtime they give to different factions… and who owns the MSM?
14. Pete George 14

    28 November 2011 at 11:23 am

    Generally polls are ok as a general indication of past preferences for those who chose to take part. Variations between polls show it is only approximate.

    The biggest problem with polls in relation to elections is there is a hig number of ‘undecided’ and very late (possibly in the booth) decision making, and a growing number that decided not to go and vote at all.

    Polls are to slow to react to the late movements, and they are not a standard pattern across all parties, so projections can be meaningless.
    • Jackal 14.1

      28 November 2011 at 11:31 am

      The undecided and non voter turnout is an indication that political parties are failing to engage with the public. If people are making up their mind on the day, they are not aware of the various policies and the effect the proposed changes will have on their lives.

      You might note that the results were far outside of the margin of error Pete George. Projections are not meaningless if they unduly influence the public.
      • Gosman 14.1.1

        28 November 2011 at 12:00 pm

        Only two of the poll predictions were well out. National’s level of support, (which was still at record high’s in an MMP environment), and that for NZ First. I’d suggest the reason for the second has more to do with the impact of the last two weeks rather than some inherent flaw in polls.
        • Jackal 14.1.1.1

          28 November 2011 at 12:12 pm

          I’d suggest that the election result has more to do with over-inflated polls leading up to the election than any proper discorse about privatization.
          • Gosman 14.1.1.1.1

            28 November 2011 at 12:20 pm

            That’s an interesting PERSONAL opinion. However I seem to remember the Labour Party making a big deal about Privatisation of Assets and Phil Goff banging on about this constantly during the election campaign. I also remember the Media reporting on this issue and highlighting the fact that the public doesn’t like the idea much according to opinion polls. What more did you want the media to do? Come out and state that any Government advocating partial provatisation of SOE’s was evil and didn’t deserve to govern? Would have missed a bit of balance if they did that don’t you think?
            • Colonial Viper 14.1.1.1.1.1

              28 November 2011 at 1:33 pm

              Publishing of plling results should have been banned in the last 7 days. Repeated expectations of a massive National win depressed voter turn out significantly.

              National’s mandate to sell of strategic state assets comprises what…backing from 35% of the voting public? JOke.
              • Gosman …

                28 November 2011 at 2:09 pm

                “Repeated expectations of a massive National win depressed voter turn out significantly.”

                Where is the evidence for this? I don’t mean just you stating well it must follow that because there was polls showing a massive National party lead going in that the low turn out was the result. I mean actual studies done which point to disillusionment caused by high poll ratings for other parties. The most I have seen is a link to Bryce Edward which suggest band wagon jumping as a result of high poll results. That is different from disillusionment caused by low poll results.
                • Jackal …

                  28 November 2011 at 3:41 pm

                  Coercive Persuasion and Attitude Change

                  Encyclopedia of Sociology Volume 1, Macmillan Publishing Company, New York
                  By Richard J. Ofshe, Ph.D.

                  Coercive persuasion and thought reform are alternate names for programs of social influence capable of producing substantial behavior and attitude change through the use of coercive tactics, persuasion, and/or interpersonal and group-based influence manipulations (Schein 1961; Lifton 1961). Such programs have also been labeled “brainwashing” (Hunter 1951), a term more often used in the media than in scientific literature. However identified, these programs are distinguishable from other elaborate attempts to influence behavior and attitudes, to socialize, and to accomplish social control. Their distinguishing features are their totalistic qualities (Lifton 1961), the types of influence procedures they employ, and the organization of these procedures into three distinctive subphases of the overall process (Schein 1961; Ofshe and Singer 1986). The key factors that distinguish coercive persuasion from other training and socialization schemes are:

                  The reliance on intense interpersonal and psychological attack to destabilize an individual’s sense of self to promote compliance

                  The use of an organized peer group

                  Applying interpersonal pressure to promote conformity

                  The manipulation of the totality of the person’s social environment to stabilize behavior once modified
                  • Gosman …

                    28 November 2011 at 3:52 pm

                    So you make a rather tenuous linkage to the impat of Polls to Brainwahing techniques based on what seems like papers on brainwashing and expect that to be persuassive do you? You were that person on Kiwiblog pretending to be a Zombie weren’t you Jackel? Because it looks like you might have canabalised your own brain with that last post.
                    • Jackal …

                      28 November 2011 at 4:28 pm

                      It appears that I ate your brain at some stage there Gosman. You asked for information concerning how polls could influence an election result… some of those studies outlined in that article date back to the 50’s.

                      Coercive Persuasion and Attitude Change are very well researched topics. In the right hands they can be used for immense good but in the wrong hands they can do untold damage. There is no tenuous link between those studies and the current dynamic.
        • mik e 14.1.1.2

          28 November 2011 at 4:20 pm

          Gooseman national had 100,000 votes less than last election Voters stayed at home and National got only 30% of eligible voters .But got 48% of those who turned up to vote so apathy was the winner by far!
    • Zaphod Beeblebrox 14.2

      28 November 2011 at 11:37 am

      Saw Colin Craig on Bomber’s iPredict show- he said their polling showed the getting 2-3%.

      On the same show Winston said he would get 8-10%.

      I’m wondering if its the way they ask the questions that miss the votes. 
15. muzza 15

    28 November 2011 at 11:24 am

    Are the polls trying to be right or are they trying to shape peoples vote patterns?
    • Colonial Viper 15.1

      28 November 2011 at 12:27 pm

      They are being used to shape voting patterns.
      • Gosman 15.1.1

        28 November 2011 at 12:36 pm

        Being used by whom and what actual evidence do you have for this other than just your extremely left wing biased view of the world?
        • Puddleglum 15.1.1.1

          28 November 2011 at 12:49 pm

          Well, there is this series of posts by Bryce Edwards which gives an overview of some of the effects found in the literature.

          Of course, you might think he is also ‘biased’ but, hopefully, not the student’s work he mentions.
          • Gosman 15.1.1.1.1

            28 November 2011 at 2:13 pm

            There’s nothing in there that suggests that polls are being used to deliberately shape public opinion just that outcomes can influence public opinion. To argue that they are being used to shape public opinion you have to produce evidence that the people who publish them are doing so with the aim of influencing opinion. I have yet to see this.
            • Jackal 15.1.1.1.1.1

              28 November 2011 at 3:51 pm

              You’re asking for information that is only held by those producing the polls Gosman… they’re not likely going to admit to their corruption.

              58% Think National’s response to the Rena disaster was bad
              26% of people thought Keys handling of the tea pot tapes was good
              53% of people prefer Phil Goff as Prime Minister

              Yet National still wins? Either people are widely voting against what they believe in or they are being unduly influenced.
              • Gosman …

                28 November 2011 at 3:56 pm

                That’s right Jackel. It is all an evil right wing conspiracy to control your mind through manipulating polls. Of course the only evidence you supply supporting this is some polls which somehow show Phil Goff has the support of 53% of people to be PM over presumable John Key (WTF?!?) as well as some other polls. Who commisioned those polls by the way Jackel? Was it perchance the media?
                • Jackal …

                  28 November 2011 at 4:31 pm

                  So you think that a contradiction in the polling is irrelevant because it was commissioned by the MSM? They have subsequently removed their findings… WTF indeed.
16. Puddleglum 16

    28 November 2011 at 11:30 am

    Polls have far too much influence on the discourse around elections and not just directly on voting choices.

    For example, Labour’s policy announcements were often discursively framed in terms of ‘desperate’ attempts to counteract low polling.That meant that the discourse was all about whether the policy announcement ‘succeeded’ in changing the poll ratings (i.e., ‘let’s wait to see the next polls’) rather than being a detailed analysis of the pros and cons of a policy.

    That means that even when the media were supposedly covering ‘policy’ and ‘issues’ they were often largely just continuing to report on the polls. The difference is subtle but real and, potentially, more pernicious as it undermines what little coverage there is of policy. Policy effectiveness becomes measured primarily in terms of its effects on polling.

    Polls also, of course, affect general conversations about elections amongst ordinary people. 
    I had, for example, a conversation (pre-election) with an educated and intelligent friend. He said that ‘Labour is a shambles’ and so he was voting Green.

    I asked him how Labour was a shambles. His answer?: “Look at the polls.” I pressed him further. Did he think Goff was about to be rolled? (No). Did he think their policy announcements were incoherent or irrational (No). Did he think their campaign was falling apart? (No).

    I think the claim about a ‘shambles’ was a justification for a feeling he had about Labour but was, itself, not supported by anything except their unpopularity. Being educated he probably didn’t want to acknowledge that his voting preference was primarily based on that sort of concern with popularity.
    • Colonial Viper 16.1

      28 November 2011 at 12:26 pm

      Yep.

      Now I’m thinking that a ban on the publication (or discussion of) of any poll results for the 7 days before election day would be appropriate.

      That last week must be based on policy and substance, not popularity. This would also give the minor parties more of a chance to shine.

      It would stop the Friday newspaper headlines and editorials saying “Labour to be smashed on Saturday, Key complete triumph most likely”. Which is what happened on the 25th.
      • queenstfarmer 16.1.1

        28 November 2011 at 12:51 pm

        Ah, the not-unprecedented socialist response to failing to win power: restrict democratic rights, because the foolish people just can’t be trusted to make the right decision. Any other democratic rights you’d like to ban while you’re at it? Perhaps banning negative discussion about any party’s policies would also help.
        • Gosman 16.1.1.1

          28 November 2011 at 1:04 pm

          Yeah I suspect that the Media would also be forced to report negatively on any right wing policy and positively on any that the left promotes. Can’t have balance now can we as otherwise people might vote incorrectly and that would lead to (Gasp! Horror!) a Right Wing government (Dun dun dah!).
          • Colonial Viper 16.1.1.1.1

            28 November 2011 at 1:30 pm

            watch your grandkids assets being sold out to a small proportion of the wealthy, and foreigners, Gosman. And relax in the knowledge that you are helping it all happen.
            • Gosman 16.1.1.1.1.1

              28 November 2011 at 1:52 pm

              Unlike you I don’t want my Government owning what is essentially commercial enterprises. I would prefer Governments stick to the business of governing and Business stuck to the business of business. It is what distinguishes the left from the right in politics. You have your reasons for disagreeing. I have seen them and obviously reject them just as you have obviously rejected the right wing arguments over why it is better for the State to stay out of businesses. Therefore you are wasting your time using silly emotive statements such as non-existing grandkids being adversely impacted by the results of part privatisations. Tell it to one of your leftist buddies and you can all commiserate together.
              • Puddleglum …

                28 November 2011 at 2:41 pm

                “I don’t want my Government owning what is essentially commercial enterprises”

                Neither do I. That’s why I opposed them being turned into businesses via the SOE Act.

                They are not businesses and should not be forced into that unnatural form. They are the vital infrastructure of a modern nation state. I’d support your implicit call for returning them to the status of a government department.
                • Gosman …

                  28 November 2011 at 3:00 pm

                  Really???

                  So are you stating that Solid Energy, Air NZ, and Kiwibank should be changed into Government departments then? On what basis would they compete with the private companies they are in the same market with if not using the commercial model?
                  • Colonial Viper …

                    29 November 2011 at 12:02 am

                    key infrastructure should be Govt owned and controlled with no bias for profit generation.
              • Vicky32 …

                28 November 2011 at 4:20 pm

                  such as non-existing grandkids

                Why am I not surprised? If you had kids or grandkids, you might have to be responsible for someone else! :O
              • *_* …

                29 November 2011 at 12:59 pm

                +10 on cock-smaking those douches
        • Colonial Viper 16.1.1.2

          28 November 2011 at 1:19 pm

          Restricting polling release is about increasing democratic rights, fool.

          It increases it by forcing the media to focus on policy substance, you see. Something that the NATs are weaker on, coincidentally.
          • Gosman 16.1.1.2.1

            28 November 2011 at 1:55 pm

            Written like only a hard core leftist could. Restricting what someone can publish is actually increasing people’s freedom. George Orwell would be proud of the use of Newspeak inherent in that statement.
            • Jackal 16.1.1.2.1.1

              28 November 2011 at 3:58 pm

              It’s increasing democratic process to ensure the media does not unduly influence the public from making an informed decision.

              If you were properly read Gosman you would know that what you’re advocating for has been used in the past to disastrous effects. I would prefer that society learns from its mistakes.
              • Gosman …

                28 November 2011 at 4:20 pm

                And if you were properly read you will know that what you are advocating has been, and continues to be, used by totalitarian regimes the world over to get the ‘correct’ message out to the population. Zimbabwe is a prime example where they have a media commission in place to ensure noone in the country is ‘manipulated’ too much by the private media.
                • Colonial Viper …

                  28 November 2011 at 4:36 pm

                  No publishing or discussion of polling data for 7 days before an election. Gives people more free room to think for themselves, not think what someone else is finding popular.

                  Also prevents the media from using polls to shape outcomes. Eg. Friday newspaper headines saying “Labour smashed, National expected to rule alone with El Presidente Don Key!”
                • Jackal …

                  28 November 2011 at 4:40 pm

                  Wrong! I’m not advocating for that type of system at all. I’m advocating for more responsible reporting and a few rules that determine the extent of influence the media can have on the public.

                  You might not have noticed but the New Zealand public has been manipulated by propaganda over the last three years. Our news services are completely ineffectual at delivering concise and correct information that is not slanted. Even the Australian news is by far superior to ours.

                  One new practice I’ve recently noticed in New Zealand is news that is very old being rebroadcast to fit the current discourse of information when it is totally irrelevant.

                  Not as irrelevant as your comparison though Gosman.
          • queenstfarmer 16.1.1.2.2

            28 November 2011 at 2:57 pm

            So now you want to “force” the media to report certain things (I wonder who determines what “policy substance” is?). Well., state controlled media is a feature of all socialist regimes.
            • Colonial Viper 16.1.1.2.2.1

              28 November 2011 at 4:30 pm

              Corporate controlled media is a feature of all capitalist and corporatist regimes. So?
        • Puddleglum 16.1.1.3

          28 November 2011 at 2:31 pm

          I would support some period of restriction of polling/publication of polls or rules around how they are reported.

          The election period is a regulated period of time and the election is a highly regulated event in all sorts of ways – spending limits, authorisation statements, misleading claims in election advertisements, no electioneering on election day, no reporting in the media on election day of anything that encourages someone to vote for one party or candidate or another, etc..

          This degree of regulation is accepted by all parties (so far as I’m aware) as in keeping with the importance of the election and election period and the seriousness of voter decisions.

          In that spirit, there is a legitimate debate about the reporting of polls for some period or for regulating the methodology of reported polls or in some other way ensuring that information provided to electors via polling is not misleading.

          An election is an exceptional circumstance and so exceptional steps to regulate the event seem reasonable to me. There is also enough evidence in the literature on the effects of polling to suggest that a look at some form of regulation is justified (see the links in another comment by me on this thread – 15.1.1.1), especially in light of the growing use of polls by media.

          Both ‘freedom’ and ‘truth/honesty’ are virtues but that does not mean that they are always closely aligned. That’s why we have defamation laws, restrictions on the reporting of court cases and the like.
          • queenstfarmer 16.1.1.3.1

            28 November 2011 at 2:53 pm

            There is a massive difference between requirements such as authorisation statements, and CV’s proposal to ban publication (or even discussion!) of certain things.

            What CV wants is to ban anyone from saying “I asked 500 people what they thought about Party X’s policies. Here are the answers they gave me…”. He would even ban you and me from having a discussion about the same thing.
            • Jackal 16.1.1.3.1.1

              28 November 2011 at 4:03 pm

              I don’t think that is what CV is saying exactly… the thing you seem to be failing to comprehend is that people can be influenced.

              If the polling is inaccurate as has been highlighted and you tell somebody about that inaccuracy then their decision could be based on disinformation instead of facts. That’s not how a democracy should be run.
              • queenstfarmer …

                28 November 2011 at 4:19 pm

                So you want to prevent the spread of “disinformation” (which would presumably include such vicious disinformation as someone accurately reporting that they asked 500 people who they planned to vote for, and what their response were.) So your plan would clearly need some central authority to determine what qualifies as disinformation versus fact. Perhaps you could call it the Ministry of Truth?
                • Jackal …

                  28 November 2011 at 4:43 pm

                  That might be a bit much for you right winger’s… as I don’t think you can handle the truth.
                  • Gosman …

                    28 November 2011 at 5:57 pm

                    No you just need a Media commission that determines what balance and fairness is. Of course it would need to be staffed by the ‘right’ kind of people.
                    • Jackal …

                      28 November 2011 at 6:26 pm

                      Personally I think that our Press Council and BSA are completely ineffectual in their current format. They are most definitely staffed by the “right” kind of people ie cronies.
                    • Colonial Viper …

                      29 November 2011 at 12:10 am

                      Nah fuck off what I am saying is no publishing of any political party poll results or discussion of those results for up to a week before election day. (I personally think 24 hours to 48 hours would be enough).

                      Its not rocket science, it is designed to allow people to think about their options already knowing how the parties are positioned, but without stupid polling based media narrative driving the daily discussion right up to E-Day.
          • Gosman 16.1.1.3.2

            28 November 2011 at 2:55 pm

            You actually make quite a persuassive case. However what you overlook is if in fact there is an impact caused by polls and it favours the right there is no way you can get this changed without a sh*tfight. Why would the right of the political spectrum allow the left to change the rules to make the political environment more favourable to the left? It would be like the Electoral finance act all over again but much worse.
            • Jackal 16.1.1.3.2.1

              28 November 2011 at 4:13 pm

              The simple answer to that is because the right should want a stable and sustainable political process. If the public continues to be manipulated and basically lied to we will have even more people becoming less engaged in politics.

              This might work in favour of those doing the manipulating at first, but the end result is never good. Effectively the less public engagement with politics the less influence politicians have on the public. This is partly to blame for the spiraling violence overseas.

              It might seem a bit strange to someone like yourself Gosman, so obviously stuck in rightwing ideology, but having a process that does not unduly influence the public is the only way to increase political influence. The media seem unable to self regulate in this area… so there’s really only one choice.
              • Gosman …

                28 November 2011 at 5:56 pm

                What???

                Please explain that again? How does changing a system that inherently favours the right, ( according to many views expressed on here), lead to a situation which the right will benefit from that change as well? They will only benefit if at some time in the future the system is used to benefit the left over the right. You have yet to show how this could happen.
                • lprent …

                  28 November 2011 at 6:04 pm

                  Personally I don’t know if it favours the right or the left. I really don’t care.

                  What I dislike is that we get a pile of debate about something that is grossly inaccurate whilst pretending to be statistically accurate in representing the whole voting population. It wastes everyones time and clogs up the political debate with something that is largely irrelevant to it.

                  I figure the best approach would be to require the polling companies and the media to publish the whole of the sampling methodology and all of the figures from the survey including failed contacts for each poll. Then we could get some transparency on how accurate the bloody things are based on how they sample.

                  At present the polls look to me to be less accurate than simply having a large sample of people guessing the probable outcomes (ie iPredict).
            • Puddleglum 16.1.1.3.2.2

              28 November 2011 at 8:58 pm

              I’m not convinced that the effects of polling have any inherent tendency to favour the right. Polls do, however, have some effect. It would be strange if they didn’t – as if people would have them simply go in one ear and then straight out the other.

              The kinds of effect they do appear to have, though, are ones that we wouldn’t usually argue would be beneficial to an effective political debate in a democratic society (e.g., encouraging the pursuit of momentary popularity over deep discussion and analysis).

              Polls also affect the behaviour of political parties and operatives. It could be reasonably argued, also, that polls reward behaviour that enhances perceptions independently of substantive issues. In that regard, I’ve noticed – and am particularly concerned about – an increase in ‘specially designed polls’ with curiously specific questions.

              The obvious examples of this are on media websites but there are also commissioned polls that ask about fictional scenarios (e.g., ‘If you had to choose between tax cuts and increases in welfare benefits, which would you prefer?’). We’ve seen a lot of that kind of polling this time around. I wonder both about the motives of such ‘push polling’ and the soundness of the methodology.

              An example of this latter kind of polling would be the poll that I understand was carried out by Farrar’s company (?) in Epsom that suggested to those polled various scenarios around National’s overall polling in the country and, in that event, how they would vote in Epsom.

              That seems to me to do two things: (1) Alert those polled (and those later reading about the poll’s ‘findings’) to the ‘danger’ of voting particular ways if National’s party vote were at particular levels, and (2) frame a poll in Epsom in relation to other polling, suggesting that the pollster in this case clearly believed that such other polling is something that voters do (and should?) take into account.

              In fact, now that I think about it, this kind of polling does reveal one way in which widespread use of polls does have an ‘inherent’ bias to the right. On the assumption that right wing political parties tend to be better funded, they would also be able to fund more polls and, in particular, more of these specific, scenario-based and self-referencing ‘push polls’ that then become news stories and further ‘advertise’ a particular framing of elections and election issues. 
      • Gosman 16.1.2

        28 November 2011 at 12:54 pm

        I presume you agree you would need cross party support for this restriction of free speach or do you think the left could just ram this through by itself like it did with the original Electoral Finance Act?
        • Colonial Viper 16.1.2.1

          28 November 2011 at 1:22 pm

          The media needs guidelines to prevent the use of polls from shaping election results and shaping election turnout.

          “Free Speech” requires that the media pay attention to people and issues, not push/pressure polls. And in the last 7 days before election day, that is what they must do.

          Further let’s not get too precious about ‘free speech’ given that Key threatened the media with police action, and there are already a lot of rules around ensuring a fair democratic process via media limitations.
          • Gosman 16.1.2.1.1

            28 November 2011 at 2:00 pm

            So would you think this policy needs cross party support or is it something the left should just ram through without reference to those across the political divide?
            • lprent 16.1.2.1.1.1

              28 November 2011 at 2:19 pm

              Do you mean like the cross-party support across the divide on asset sales?
              • Gosman …

                28 November 2011 at 2:44 pm

                Ahhhh. So are you stating that you would like this to become the policy of one of the main parties of the left and they would campaign on it before an election so they could claim a mandate if they win?

                If so then I too would love to see this. I would love to see that mainstream left wing party being attacked for being anti-freedom of speech. Nothing I think would be as enjoyable politically as seeing a left wing party commit political suicide by advocating something like this during an election campaign.
                • Jackal …

                  28 November 2011 at 4:20 pm

                  Your inference there is that the media has so much control of the political process during and election campaign that no political party should move to curb that influence… thus showing that you agree the dynamic exists and that the MSM does in fact unduly influence the election result through their manipulations.
                  • Gosman …

                    28 November 2011 at 5:51 pm

                    No. I’m stating that any decent right leaning party would make mince meat of a left leaning party campaigning on the stated policy of denying the media the right to publish polls during an election campaign. The media would then run with this because the policy is a restriction on their freedom of speech. This last point is irrefutable. You might like to argue that it is a necessary restriction but it is a restriction nonetheless. I can just imagine the headlines now. It is simply too delicious to describe.
                    • Colonial Viper …

                      28 November 2011 at 6:17 pm

                      No. I’m stating that any decent right leaning party would make mince meat of a left leaning party campaigning on the stated policy of denying the media the right to publish polls during an election campaign.

                      Sorry mate using push polling to shape the outcome of an election is not a “democratic right”.
17. J Mex 17

    28 November 2011 at 12:30 pm

    The polls were FAR more accurate than ‘The Standard’ authors and posters.

    Should we ban predictions by media and bloggers, because they are generally incorrect and may shape public opinion?
    • Colonial Viper 17.1

      28 November 2011 at 1:28 pm

      Yeah which Standard posters got their opinion on to the front page of the NZ Herald, or on the 6pm news bulletin?

      Oh that’s right none, moron.
      • J Mex 17.1.1

        28 November 2011 at 3:22 pm

        I see, it’s about readership size is it CV?

        So should a small newspaper be allowed to express likely poll outcomes, but big one not? Where is the cutoff?

        According to the NZ Herald, they have an annual readership of 582,000.

        I don’t know the readership of The Standard, but last year it was stated as 212,000 unique visitors (up 45% from the year prior).

        The issue seems to be that election outcome scenarios, seen by voters, affect elections. So to those who want to ban the publishing poll outcomes (and to you CV), my question is this… “At what point (readership size) should “The Standard” or another publisher be “banned from expressing likely election outcome scenarios? ?
        • lprent 17.1.1.1

          28 November 2011 at 3:35 pm

          But we don’t run inaccurate psuedo-scientific polls with fundamental methodological flaws and try to pass them off as being accurate – to the point of stating the expected outcomes from them. What is written here is stated to be opinion.

          Can you see the difference, or are you blinded by something? A blighted future perhaps?
          • Gosman 17.1.1.1.1

            28 November 2011 at 4:24 pm

            Hang on a minute here. You stated in a post above above that many of the polls you are unable to identify if they are scientific or not because they don’t publish their methodology. Now you seem to be claiming that they have fundamental methodological flaws. How do you know this if you don’t know the methodology used?
          • J Mex 17.1.1.1.2

            28 November 2011 at 4:40 pm

            Ohhh, everyone seems a bit touchy today.

            Don’t come over all high and mighty on the polls now fellas. The Standard authors spent vast column inches every time a poll came out that was good for Labour/bad for National.

            A bit hard to take a “bugger the polls” and “ban the polls” attitude a couple of weeks after getting all wet over polls that were good for your “team”

            http://thestandard.org.nz/polls-good-for-left/

            Let’s be clear. You have people here asking to ban polls. For the primary reason that they are not accurate, and they affect the way people vote. They are more accurate than a lot of things that affect the way people vote. They are up front about what they are and their methodology. They show their margin of error. They are treated with scepticism.

            Any analysis here about whether the polls were correct within their margin of error. Nope. Just a bunch of angy left wing party supporters feeling hard done by because their team didn’t win.

            How about throwing the same spotlight on yourselves.. “National’s beginning to leak too. I’ve just been hearing about their latest internal polling, and it ain’t pretty – Eddie”. Opinion posing as fact?

            Here’s CV giving tacit approval to political parties publishing bullshit information ito supress turnout

            “Colonial Viper 15.1.1.2
            2 November 2011 at 10:04 am

            wtf? A political party suggesting that their own candidate is ahead of the game in order to, amongst other things, try and suppress opposition turnout is somehow “dishonest”?

            What are you, 16 and participating in your first election campaign?”

            That is o.k? but independent polls are not o.k? Pretty strange world you guys live in.

            Polls are never going to be 100% accurate. But let’s be honest, they are impartial*

            *Except the Horizon poll – which incidental seemed to be the most enthusiastically received around here.
            • r0b 17.1.1.1.2.1

              28 November 2011 at 4:47 pm

              The Standard authors spent vast column inches every time a poll came out that was good for Labour/bad for National.

              I mostly wrote about the polls – but only because someone has to, it would be odd for a political blog not to provide a forum for discussing them. I have made my own opinion on polls clear many times, they are not very accurate. Why do you think the image that accompanies each post on polls is a dart board?

              Except the Horizon poll – which incidental seemed to be the most enthusiastically received around here.

              Oh what rubbish. I don’t think you’ll find a single post here pushing or even discussing Horizon polls, except to be critical of the methodology. We never even reported them during the election campaign.
            • lprent 17.1.1.1.2.2

              28 November 2011 at 5:10 pm

              The Standard authors spent vast column inches every time a poll came out that was good for Labour/bad for National.
              

              I think that you’re confusing us with kiwiblog and trying to rewrite history. We don’t have any particular vested interest in polls. As you can see when you look at the top posts on polling where our authors have been criticizing their methodology for a long time….

              http://thestandard.org.nz/?s=poll&isopen=block&search_posts=true&search_sortby=relevance

              Just taking the first ones by relevance rather than your rather disturbing habit of cheery picking one post on polls by its headline (because you can’t have read the post and considered that it was good about the polls. It merely linked some msm stories, and it was critical of the media use of the polls)

              http://thestandard.org.nz/pretend-polling/
              http://thestandard.org.nz/on-the-perils-of-polls/
              http://thestandard.org.nz/reviewing-the-weekend-polls/
              http://thestandard.org.nz/why-the-polls-suck/
              http://thestandard.org.nz/polls-everywhere/
              http://thestandard.org.nz/a-week-of-polls/

              Now if you actually read the posts rather than reading your navel to try to rewrite history, you’d note that damn near every post we write on polls criticizes them and their usage. Why? Because they are as stupid as well… you are…

              Throughout the years since the last election the authors tended to write about every second Morgan poll as they were the only ones that arrived with much consistency (about every 2 weeks) and one every other month of so we’d look at another poll. We did a few more closing to the election – but mainly those would be a single post on several polls.

              The usual complaint is that we don’t write about every poll. You have to go to the obsessional blogs like kiwiblog, nominister, or whaleoil if that is your taste.

              Most of the posts written about the Horizion poll here have been either cautious or critical of its methodology

              As for the other posts. In each case where and why an opinion was formed including if it was hearsay was stated. I guess you don’t actually read the posts – do you have illiteracy issues?

              What I’m saying is that the polls are useless for political debate because they do not even come close to being inside their “margins of error”. That because of the methodology of the polls they should be regarded as being psuedo science. That the lack of transparency on the polls means that the obvious errors cannot be corrected. An d that I believe that the news media know all this and choose to treat them as being factually correct because they are a cheap way of filling headlines. All of your bullshit doesn’t change any of that.

              We’re not responsible for commentators. Talk to CV.

              You really are a bit of a useless pillock aren’t you…
              • J Mex …

                28 November 2011 at 7:40 pm

                Here’s my summary:

                – The polls were accurate, there or thereabouts.
                – They were certainly more accurate than every author on Red Alert
                – I think that many people changed their vote based on the final polls. They realised Labour were certainly goneburger and were almost certain to govern alone. They threw votes to NZF, Conservative as a brake on National.
                – Many Standard commentators and Authors did the same thing and voted tactically for NZF because of poll results. You can’t tell me that The Standard authors are big NZF fans, or they ‘naturally’ sensed the election outcome. Bollocks – They voted NZF based on what the polls were telling them, (which they knew was the truth)

                Questions for you (or others)

                – Do you honestly believe that polling is not useful for gathering information on likely outcomes? I think that is easily the best (but not perfect) method for finding out the most likely outcome. Asking your friends certainly isn’t. Gauging doorknocking reception certainly isn’t. That is why politicians and the media poll.

                – Do you think that publishing poll results leading up to an election should be banned? That’s what a few of your commenters are saying. And you aren’t calling them idiots. Although they almost certainly are.

                For every voter put off voting by polls, there would be two that vote because of them. Polls are one of the the only things that give elections colour and excitement. Without polls, New Zealanders (other than the political junkies) would be bored witless. Close polls inspire and motivate supporters, they galvanize candidates, they allow parties to focus resources.

                But you have commenters who want to ban polling, and you don’t call them for the idiots that they are. You are either so far into the red hole, that you have lost all perspective, or you are supremely disingenuous. I’m leaning toward you being a disingenuous pillock, but it’s most likely a little from column A and a little from column B
                • lprent …

                  28 November 2011 at 7:54 pm

                  I have already answered most of those questions. Just read down the post or use the search in comments for @author lprent by date and do some reading. If you find something that you don’t understand (quite probable based on these ‘questions’ framed as invalid statements) , or find something not covered, then ask something specific.

                  You really are lazy about reading aren’t you? Are you as lazy in real life?
                  • J Mex …

                    28 November 2011 at 8:32 pm

                    I’m going to ignore your lazy barbs. I am sure that if you and I were in the same room together, people wouldn’t be picking me as the lazy one of us two.

                    I see that you are a “ban the polls” or publish full methodology

                    I don’t think we need legislation around polls, there is no need. The ones that are more accurate get more attention and more credibility. The ones that aren’t don’t. Horizon will have a hard job maintaining any credibility after this election.

                    I stand by what I said about polls making the election more interesting. I don’t think that anyone can credibly argue the corollary.

                    I also stand by the fact that polls allowed your authors and your commenters the chance to vote tactically. Nobody would have known how far away Labour was without polls. Nobody would have known how close NZF was from 5%. People stopped National getting an outright majority because of the information that the polls told them.

                    You can’t credibly argue that the number of your authors and commenters would have voted tactically for NZF without the information that the polls provided.

                    But fuck it, ban em’. eh?

                    And drop the “you are stupid” and “you are lazy”. It is beneath you and it makes you look like a complete cock
                    • Puddleglum …

                      28 November 2011 at 9:28 pm

                      “I stand by what I said about polls making the election more interesting. I don’t think that anyone can credibly argue the corollary. [I think you mean ‘converse’]”

                      Contrary to what you claim, prior to opinion polling elections were still very interesting to many participants. (Think the first half of the 20th century).

                      The interest and excitement came from attendance at public meetings, arguments (and even fights) between opponents and a generally intense debate over deeply important issues.

                      I’d go further; elections appear to have been far more interesting and exciting prior to ubiquitous polling. That’s because elections were understood to be about argument and conflict (which, if nothing else, is pretty exciting stuff) and no-one could be sure how the debate would go, overall. That is, lack of knowledge of (or presumed insight into) the outcome added intensity and emotional investment to the event.

                      In addition, to support your claim you’d have to argue that, without constant polling, turnout today would be even lower.

                      I think we can assume that much of the non-voting adult population were not overly excited by the political spectacle (including polls every other day towards the end of the campaign).

                      I guess from your perspective, far from banning polls perhaps we should have compulsory polling every hour on the hour in order to apply the democratic defibrillators to these apparently terminally unexcitable people?
                    • lprent …

                      28 November 2011 at 10:02 pm

                      See, I knew you could find out what I’d said if you only tried a little. The accuracy of the polls is pitiful unless of course you think people magically make up their minds at the last minute and the polls just happen to see it. There is no ability to determine accuracy of a poll a month or even a week out from an election which makes your statements ridiculous. The only way to determine if they are likely to be accurate is to examine the methodology and figures. For one unstated reason or another, the polling companies aren’t willing to do that.

                      But I see that your attitude is that the polls are basically entertainment. In that I suspect you are probably correct..

                      Anyone could have known how close NZF was to 5%. All they had to do was to think about constituencies. Read my comments from three years ago – I was predicting that he would be back and why. It was obvious then that it was likely.

                      I call people as I see them.. If you act like fuckwit, then I will treat you as one. – including giving you a mirror in words. Don’t get too proud of it, I’d do it to anyone playing testosterone games here. You should know because you have seen it before – have you forgotten?.

                      And I really don’t care what people think – around here I react in a appropriate way to what others do. It prevents people from trying to set the rules
18. Faversham 18

    28 November 2011 at 12:35 pm

    IMHO the polls were out mostly in their over-estimation of the turn out.
19. Brendon O'Connor 19

    28 November 2011 at 1:48 pm

    The irony is in my heart I wanted some form of semi socialist government in NZ, but I knew it wasn’t going to happen, so I moved to Australia with my family in September. I could see no other option. It turned out I didn’t able to need polls to tell me the outcome then, and I turned out to be sadly correct.
    NZ is going down the toilet as the majority, or at least slightly more than 35% of them (which is what it would take to beat National) still haven’t worked out that what makes businesses successful is when the average Joe has money to buy the goods and services, not just blow their meagre wages on the essentials which is what the majority of kiwis are now only able to do.
    Plus there will be reduced incomes streams for the Government so this and future governments will have to get by on spending less on the essentials or tax individuals more. We know this government opposes taxing the greedy and supports taxing the needy, so the only way to redistribute income for them is to keep lowering taxes for the wealthy and increase taxes on the poor. If wages were higher and taxes were lower for the poorer then the would have more money for buying extras and then make businesses richer that way.
    • mik e 19.1

      28 November 2011 at 4:23 pm

      Not only are the right wing taking over NZ but it looks like they are trying to take over the left as well.Never happy until they have ground NZ into the ground hook line and sinker just like Muldoom did!
20. Campbell Larsen 20

    28 November 2011 at 10:17 pm

    Haven’t had time to sift all of this thread, but have skimmed enough to ask a possibly pertinent question:
    Gosman – pollster????
    Can’t for the life of me think of another reason why he would submit so many comments defending the statistical equivalent of televised crystal ball gazing.
    Good money in polling…..what was the name of that penguin-like fellow who writes PR for the Nats and has a polling company….? Friend, relative or just kindred soul/ clone?
    • Colonial Viper 20.1

      28 November 2011 at 10:35 pm

      you must be thinking of Farrar and his firm Curia. I recently met some young Tory grad who used to do polling for him. Small place, NZ.
      • felix 20.1.1

        28 November 2011 at 11:06 pm

        Gee it would be so weird if the Nats were paying Farrar to do polling and he was hiring staff who spend their work time pushing the Nats’ agenda on political blogs.
21. Georgy 21

    29 November 2011 at 10:15 am

    All polling results should be banned during the election campaign period. The unreliability, the effect on voters thinking, and the precedence poll results take over policy announcements and discussion are all very good reasons for polls to be banned during the campaign period.

    BAN THE POLLS ! ! !

Links to post

1. Bugger the polls « The Standard | Quite Interesting Stats and Facts | Scoop.it: posted on 6 December 2011 at 5:53 pm

• Comments are now closed

• Log in
• Donate
• Contribute post

• Articles RSS