Polity by Rob Salmond

Read Post

Polity: Poll Soup

63 Responses

First ←Older Page 1 2 3 Newer→ Last

  • Nick Russell,

    Which begs the question, how does the left do this? Recently we have seen 2 strategies in action - free tertiary education and (somewhat muddled) opposition to TPP. They haven't been game-changers so far, and one possible inference one could draw is that Labour might have done better in terms of winning over current National supporters if it had come out in support of the TPP, even if reluctantly. The current position seems to be an exercise of singing to the choir which excites the base without attracting any new support.

    Wellington • Since Jul 2008 • 129 posts Report

  • bob daktari,

    I don't think the left has to win that 12% of voters - they are nationals to lose and its more likely the governing power will lose voters than the opposition pick them up

    Saying that Labour especially could really do with some serious work on their messaging - case in point Little calling the protestor yesterday a scumbag

    auckland • Since Dec 2006 • 540 posts Report

  • Rosemary McDonald,

    Mr Little said he was disgusted by the incident.

    "Totally unacceptable, this is a community occasion where we commemorate and reflect on a great tragedy," he told ONE News.
    "For some scumbag to come and do that to anybody, doesn't matter if it's one of my political opponents, is totally unacceptable."

    [Unacceptable is an understatement.

    Well, Rob, here's the guy that Little labelled scumbag.

    I hope that part of the strategy to raise Labour's rating in the polls is to get Angry Andy to keep quiet until he learns the facts.

    Waikato, or on the road • Since Apr 2014 • 1346 posts Report

  • andrew r,

    Among Nationals more obvious wealthier support group are this lower middle class block of Nat voters. Almost seems unless economy tanks, affecting that group particularly, National will be returned at next election. Nothing else will change that outcome. I don't think Little or Labour are currently doing a poor messaging job its just that folk seem uninterested politically and so long as they have a job and can meet their grocery rent/mortgage budgets - all good, nothing else in the broader realm is of concern, sigh. Suits National perfectly. That Winston Peters is and likely still will be in the mix is pretty depressing. He's at heart a conservative politician who occasionally throws a spanner in the works out of self interest. But I guess all western democracy's especially those of proportional representation have a similar type weird ideologically barren protest type party there in the end game too.

    auckland • Since May 2007 • 100 posts Report

  • BenWilson,

    Yes, that’s a hill to climb. No, that’s not a towering cliff-face. I think any prognostications of doom are pretty premature at this stage.

    Absolutely. It's close.

    Yes, 1 in 12 National voters switching would get things closer. However, that's not the only way the gap could be closed, and I'm not sure that we really know enough about the way people change their votes to say that it's the way it would be closed. Also possible is that 1 in 12 National voters stop voting and a similar number stop not-voting and vote for the "Left coalition", ie Labour or the Greens.

    It seems to me more likely that this is the way support swings in NZ. People move in and out of not-voting. I think this simply because it is a much smaller step, and we're talking about two parties that have very big solid cores of lifelong voters, highly tribal. A not-vote could be a more palatable protest than a vote for their traditional enemy.

    Auckland • Since Nov 2006 • 10657 posts Report

  • simon g,

    If polled today, I would vote Beach.

    Parliament and politics kicks off in February every year, but the public don't welcome the return (by "public" I mean people whose idea of fun is music festivals and cricket and a swim, not watching points of order in the House).

    These 'State of the Nation' speeches are ignored by the nation. Even more so now than 5-10 years ago, with the disappearing media coverage of political issues (as opposed to protests about those issues).

    What will shift polls? Events, dear boy, events. And a drop in the temperature, perhaps.

    Auckland • Since Nov 2006 • 1333 posts Report

  • James Littlewood*,

    Does the one-in-twelve equation factor in the missing million? Between them, Labour and the Greens failed to inspire their marginal constituents.

    Auckland • Since Mar 2008 • 410 posts Report

  • Danyl Mclauchlan,

    Does the one-in-twelve equation factor in the missing million? Between them, Labour and the Greens failed to inspire their marginal constituents.

    The bad news: about half the missing million are probably National-leaning voters.

    Wellington • Since Nov 2006 • 927 posts Report

  • Mr Mark, in reply to BenWilson,

    Yeah but tribal voting has been declining since the 1960s, Ben. Partisan dealignment throughout western liberal democracies. More and more non-aligned and (in particular) softly-aligned voters.

    Labour lost a whole swathe of voters to National at the 2008 General Election - and they weren't lifelong/core Lab voters who saw the Nats as traditional enemies.

    Just as I'm pretty sure they don't see Labour as anathema now.

    Wellington • Since Dec 2009 • 128 posts Report

  • Hilary Stace,

    My theory is that nothing much happens of significance until the last few days before the actual election. The people who decide the election (about 5%) decide then whether to vote at all and who for.

    I also hear there is a lot of National Party money (of which they have a great deal) going into very sophisticated digital analysis and targeting of voters similar to that done in the last UK election and which turned the election in the last couple of weeks. Analysis that is based on publically available information such as FB likes. Is this correct, Rob?

    Wgtn • Since Jun 2008 • 3229 posts Report

  • BenWilson, in reply to Mr Mark,

    Labour lost a whole swathe of voters to National at the 2008 General Election – and they weren’t lifelong/core Lab voters who saw the Nats as traditional enemies.

    I'm curious if you know how many, and how you know. Longitudinal studies aren't that common. It's common to say "Labour lost voters to National", but actually proving this based on opinion polls of total support is not very compelling. To really answer the question of who changed their vote, and from whom to whom, you have to actually design your study to answer that question.

    Auckland • Since Nov 2006 • 10657 posts Report

  • David Hood,

    I agree with Ben here, when you look at the number not voting, and the change in that number between elections, that is a huge factor in tipping the deciders.

    In the US at the moment, large turnout is seen as favouring Sanders and Trump (bringing in people who otherwise would not have contributed) and increased turnout more generally favouring democrats.

    Dunedin • Since May 2007 • 1445 posts Report

  • Mr Mark,

    Three things:

    (1) Although I haven't checked, I'd say the December Herald-Digi outlier (Nat 51% / Govt 55% - a massive 10 points ahead of the Oppo) probably played a crucial role in the fall of the Nat/Govt Bloc average in Jan-Feb 2016.

    (2) Bear in mind that there's a post-May 2015 divergence between two sets of pollsters:
    Colmar Brunton and Reid Research have placed the Opposition ahead in every single poll over the last 9 months, while the Government leads in all of the Herald-Digis and in three quarters of the Roy Morgans.
    So the varying ratio of Govt-friendly polls ( RM / HD ) to Oppo-friendly ones ( CB / RR ) in any given two-month period potentially exerts a decisive impact on Party vs Party / Bloc vs Bloc averages.

    (3) I wonder if Rob's right when he suggests the putative Left coalition (Lab+Greens) need to - at bare minimum - tie with National to be seriously in the game (ie seriously tempt Winnie) ? Seems a big ask.
    Just wondering what Winnie would do in a Left 42% / Nat 46% / NZF 9% scenario ?

    Wellington • Since Dec 2009 • 128 posts Report

  • James Littlewood*, in reply to Danyl Mclauchlan,

    about half the missing million are probably National-leaning voters

    I wouldn't have thought so. How do you figure that out?

    National ran a well orchestrated campaign that delighted their followers and alienated their opponents, while the opposition - umm - didn't.

    Auckland • Since Mar 2008 • 410 posts Report

  • linger,

    N.B. it is somewhat premature to claim that "National's support has too been damaged" on the basis of a mere 1.5% shift in polling, because that change is still (just) within the noise level of the aggregate of 4 polls.

    Tokyo • Since Apr 2007 • 1944 posts Report

  • Nick Russell, in reply to James Littlewood*,

    I just love speculation about the voting intentions of a group of people whose only shared characteristic is not voting. Pitching your policies to people who don't vote is like marketing hamburgers to vegans.

    The post-mortem of the UK election by the polling companies confronted the question of why their polls had failed to pick up the fact that the Conservatives had a significant lead that translated into an outright majority. Most pollsters were still picking a hung Parliament. Among other factors, the review found that pollsters did not give enough weighting to the views of voters aged over 70 and those who were difficult to reach - did not answer the phone during office hours and took 3 or more attempts to contact before responding. Both of these groups, it transpired, were more likely to vote than the average, and more likely to vote Conservative. By contrast, the views of younger people were over-represented because they are less likely to vote at all.

    Bottom line - nobody owns the missing million. And their support is worthless anyway because they don't vote. Better to spend your time and resources developing policies that are likely to appeal to people who do vote.

    Wellington • Since Jul 2008 • 129 posts Report

  • Mr Mark, in reply to BenWilson,

    The Jack Vowles et al New Zealand Election Study would be the place to find data on the percentage of voters swinging from one party to another. Unfortunately, without SPSS, I don't have access to those stats (and they don't seem to have included them in their recent (published) post-Election analyses in the way that they once did in the 1990s/early zeros).

    But - I've decided to wander on down the (admittedly sometimes fraught) road of common sense. Call me old-fashioned, but there you are.

    Between the 2005 and 2008 General Elections:

    Labour lost a little over 138,000 votes
    NZF lost almost 35,000
    Progressives lost more than 5,000
    Whereas, the Greens rose by a mere 37,000

    Nats, meanwhile, gained 164,000
    with the entire Right Bloc up 182,000

    Turnout in 2008 was only slightly down on 2005.

    Can't prove anything - possible , I guess that there were huge Lab-to-non-voting and non-voting-to-Nat swings - but a significant direct Lab-to-Nat swing would seem a little more likely.

    Wellington • Since Dec 2009 • 128 posts Report

  • Felix Marwick,

    Actually there was some research, released during the last election campaign, that looked at undecided/inactive voters and their possible intentions.

    Going from memory (I can't locate the stories I wrote at the time), the findings showed the split, while it favoured the left, was only of a small margin.

    Unfortunately I can't quite recall who published the study and Google isn't being particularly helpful. But the issue has been looked at

    Wellington • Since Nov 2006 • 200 posts Report

  • David Hood,

    @Mr Mark. IF you are used to R, you can use the haven library to work with the SPSS format NZES. Last time it was discussed on Public Address, this was what I was using to convert the SPSS file to a csv via R.


    fileLocation <- file.choose()
    ds <- read_spss(fileLocation)
    ##this function is needed as there are missing value labels so can’t as_factor
    applyValLabels <- function(column){
    knownvalues <- attributes(column)$labels
    valuelabels <- names(knownvalues)
    lookup <- data.frame(column = as.character(knownvalues), valuelabels, stringsAsFactors = FALSE)
    target <- data.frame(column = as.character(column), index=1:length(column), stringsAsFactors = FALSE)
    combined <- merge(target, lookup, all.x=TRUE)
    combined <- combined[order(combined$index),]
    output <- as.character(column)
    output[!(is.na(combined$valuelabels))] <- combined$valuelabels[!(is.na(combined$valuelabels))]
    newfile <- paste(fileLocation,”.csv”, sep="”)
    for (eachName in names(ds)){
    dseachName <- applyValLabels(dseachName)

    write.csv(ds, file=newfile, row.names = FALSE)

    Dunedin • Since May 2007 • 1445 posts Report

  • BenWilson, in reply to Mr Mark,

    Yes, the NZES is a start, but you can't get the data for 2014, for starters. It's not yet released. I spent a lot of time analyzing that study last year so I can have a look at how much longitudinal use it is. Definitely way better than nothing.

    But either way, it's not helping us to make judgments about things that have happened since 2014, even if you get the data. Without the 2014 data, you're back to 2011. Would still be interesting to see what patterns there are with individual voter movement.

    Auckland • Since Nov 2006 • 10657 posts Report

  • David Hood,


    And here are the patterns from 2008 to 2011 from the NZES

    The movement between National and Labour is pretty much a wash, but Labour lost a lot more of their support to the Did Not Vote as a percentage than National lost to that group (it was about the same in raw numbers, but national had more total support).

    Dunedin • Since May 2007 • 1445 posts Report

  • Mr Mark,

    Missing Million

    A few posts from some years back ...

    ... Grumpollie (Andrew R at Colmar Brunton ?) analyses the New Zealand Election Study data on those of the missing million with a party preference in 2008 and 2011

    He was responding to Danyl's Dim-Post discussion

    and to various media claims about non-voters' political proclivities

    Grumpollie's piece, in turn, inspired Young Master Farrar to speculate

    prompting this response from Rob Salmond ..
    ... http://polity.co.nz/content/impact-labours-gotv-efforts#comments

    in turn, prompting this response from Grumpollie

    See also this detailed analysis from Puddleglum at The Political Scientist

    Wellington • Since Dec 2009 • 128 posts Report

  • Russell Brown, in reply to Mr Mark,

    Thanks for those, Mr Mark!

    It's always useful to be reminded that a question has been chewed over before.

    Auckland • Since Nov 2006 • 22850 posts Report

  • Mr Mark, in reply to David Hood,

    Cheers, David.

    The 2008-11 stats are very interesting.

    But, I'd especially like to see 2005-08. That's the bone-of-contention in this wee debate - the Election where the decisive Left-to-Right swing occurred.

    Lab down 7 points
    Left Bloc down 6 points
    Clark Govt Bloc down 7 points

    Nats up 6
    Right up 7

    Little change to the Oppo and Right Bloc percentages in elections since then (despite on-going Lab decline)

    Also, Lab swing to non-voting in 2011 isn't too surprising given that 2011 was THE low turnout election (down 5 points on 2008). In contrast, turnout only fell 1 point in 2008 (relative to 2005). So, less likely to apply to 2008 - which, as I say, is the bone of contention here.

    Wellington • Since Dec 2009 • 128 posts Report

  • Sacha, in reply to Mr Mark,

    Andrew R at Colmar Brunton ?

    that's him, yes.

    Ak • Since May 2008 • 19745 posts Report

First ←Older Page 1 2 3 Newer→ Last

Post your response…

This topic is closed.