In an effort to differentiate the Labour Party from National and ACT, and to rule out Winston Peters, Chris Hipkins took aim on Sunday at a putative right-wing coalition of ‘cuts, chaos and confusion’. He’s clearly rattled by Peters who’s been on the hustings courting post-pandemic disaffection and, as always, nostalgia. In contrast, Hipkins asserted his ‘vision for an inclusive, progressive and prosperous society’.
Although Hipkins was calling for unity, many New Zealanders may think his ‘vision’ is getting short-sighted. During the economic shutdown, Labour’s policies denied people their normal social and economic rights. That wasn’t very ‘inclusive’ if you couldn’t fly home from abroad. The other left-wing parties may not agree that Labour is ‘progressive’ after Hipkins ruled out capital gains or wealth tax. Families on low and middle incomes won’t be feeling ‘prosperous’ in the cost-of-living crisis. They may wonder if Hipkins is really in it for them.
Though his desire to include certain minorities is laudable, Hipkins may end up shadow-boxing in the corner, rather than fighting in the centre. The majority of voters might look elsewhere.
Some may defect from the left to NZ First so that Winston can apply the handbrake to the Luxon–Seymour express.
Should we trust polls, or treat them critically?
The Research Association, which upholds standards for polling companies, is on a publicity drive to convince Kiwis that they ‘should trust polls’. Are they asking people to abandon critical thought?
More sensibly, pollster David Farrar advised people not to read too much into one poll and not to cherry-pick the ones they like. People should look at trends over time and at average results. That’s a duly cautious and critical approach. These surveys are based on samples and they’re statistical estimates. Hence they report margins of error, within which the ‘true’ result probably lies.
So don’t read anything into one-off changes that are within margins of error.
The polling companies are staffed by trustworthy people, but that doesn’t mean you should blindly trust their results. People trust other people, not numbers. In other words, people might trust surveyors to be honest and not to fabricate results, but critical thinking is called for when reading any research results.
And political opinions aren’t objects in space waiting to be identified and measured by brainy people with horn-rimmed glasses.
Now, in the fortnight before the 2020 election, three published opinion polls indicated that Labour would win, and they did. But, in the aftermath, few commented on how inaccurate those polls had been. They’d put Labour ahead of National by about 15 percentage points, but the election gave Labour a lead of 24.4 points. This discrepancy between 15 and 24.4 was well outside of margins of error.
It's been argued that this was due to the high numbers of advance voters. Many survey respondents may already have voted, so asking who they’d vote for ‘if the election were held tomorrow’ was confusing. But similar discrepancies can be seen in polls done before the advance voting booths had opened, so that story doesn’t cover it.
A better explanation may be about who actually votes. Research companies are polling those eligible to vote, but not everyone who’s eligible is enrolled, and not all those enrolled will vote.
We’re doing well if three quarters of eligible people actually vote in a general election. And the turnout rate varies between elections and across the population. Younger people are less likely to enrol, and they’re less likely to vote even if enrolled. Māori in all age-groups are less likely to vote than non-Māori.
Surveyors aim to get samples that resemble the eligible population, but some people just don’t have time for them, or don’t tell them the truth. Those who aren’t interested in helping surveyors are probably demographically similar to – if not the same as – those unlikely to vote.
Some surveys rule out those who say they don’t intend to vote. But people change their minds. When there’s a surge in election turnout – as in 2005 and 2020 – lots of people who wouldn’t normally have voted decide to go out and vote. And many people decide who to vote for on the day.
So a strong turnout may widen discrepancies between opinion polls and election results.
The 2005 election saw turnout rise to 80 percent of those enrolled. The two published polls immediately before it had National six percentage points ahead of Labour, but in the election National was behind by two.
The pollsters do their best to be on target; it’s just that the target’s blurry and won’t stand still.
All surveys have margins of error, but, in addition, the population of people who actually vote, and their voting choices, are changeable. And you never know how many survey respondents have scant regard for the truth.
Polling companies are trying to ensure they get representative samples by setting up large online panels, but we’ll have to wait till after the election to see how well they’re doing. I’m picking there’ll be a relatively low turnout in this election. If so, the discrepancies between opinion polls and election results will be smaller than in 2020 – if my hypothesis is correct. Obviously, I’ll have to return to this topic after the election results are all counted.
For comic relief, here are some examples of nutty polls:
In 2021, a YouGov survey asked 1,224 adult Americans if they could beat a lion in a fight, unarmed. Eight percent said they could – men and women almost equally. But we’ll never know how many really meant it. Perhaps some respondents just got confused, or were trying to please, or wanted to impress, or were lying to get one back at the system, or were taking the piss out of a totally ridiculous question. Aside from the statistical margin of error, surveys suffer from human error: people get things wrong or invent stuff, and then they change their minds, if they even pick up the phone.
So if four percent of Americans said they ‘believe shape-shifting reptilian people control our world by taking on human form and gaining power’, then we’d want to account for those who were too tired or confused to answer correctly or were just kidding. If 23 percent of American millennials in 2016 told a surveyor they’d prefer a giant meteor to extinguish human life above either a Clinton or a Trump presidency, it’s probably because survey respondents often just express how they feel.
Seriously, though, the political polls are useful information: just read them with caution.
Do the polls shape voters’ expectations or even election results?
The research evidence on this is mixed, I’m afraid. But they do seem to have some impact.
Political columnist Bryce Edwards, for example, asserted that the recent drop in Labour’s polling below 30% ‘will risk becoming a self-fulfilling factor in the campaign’ as people will cease believing in Labour.
Was that attributing too much influence to the polls?
There’s a good ‘freedom of information’ argument that the polls are a part of a vibrant democratic culture. As such, their accuracy and the integrity of surveyors matter a great deal.
Some people make a strategic voting choice based on what the polls are suggesting may happen. For example, not wanting to ‘waste’ a vote on a party that’s under 5% in the polls. Or voting for a party to moderate others that, according to the polls, could form the next government.
There’s also a suspicion, though, that polls, whether accurate or not, are unduly influencing our expectations and choices as voters. I’ll let you be the judge!
The Woodhouse Saga
Did ‘being male’ cost MP (and former cabinet minister) Michael Woodhouse a high ranking on the National Party’s list for the 2023 election?
First, it appears he didn’t actually say it was due to ‘being male’. The ODT’s editor put that spin on it. Mr Woodhouse did say, ‘there was a contest between diversity and experience, and in my case diversity won’. And indeed the National Party’s leader and president were eager to point out the diversity on their election list when they announced it. Some of those ‘diversity candidates’ lack parliamentary experience; while Woodhouse has loads of it.
We don’t know exactly where on the list Woodhouse was placed, as he promptly resigned when he found out, and his name had been removed. But, as he was contesting a probable Labour seat (Dunedin), he would’ve counted on returning as a list MP. So he must have been put at such a low position on the list that he probably wouldn’t return to parliament – and, even if he did, he wouldn’t be in a Luxon-led cabinet.
If things go well for National, they might win, say, 48 seats. On their list, there’s a number of white males contesting safe (or probable) National seats who are well below 48 – down in the 50s or lower. When those guys win those electorates, they’ll automatically displace candidates much higher up who either lost their electorate contests or were list-only. That’ll dilute the diversity that the National Party leaders were crowing about. The party caucus after the election won’t be as diverse as the party list looks now.
Those white guys are set like submersible drones to torpedo National’s diversity on 14 October.
Woodhouse must have been ranked just above the likes of Sam Uffindell (#57, Tauranga) and Tim van de Molen (#58, Waikato). But, unlike them, he was contesting an electorate he couldn’t win.
When asked about it, Luxon showed no sign of remorse at losing Woodhouse. And Woodhouse did say, ‘I wasn’t part of the caucus club. [And] I have a habit of saying things without fear or favour’.
So Woodhouse fearlessly said, ‘Screw you all’, and resigned. OK, he isn’t on record saying those latter words. I put my own spin on it!
My assessment: Woodhouse showed personal integrity. National’s diversity push looks disingenuous, and some may call it virtue signalling.
I'm sensing a new addition to the election campaign bullshit bingo board, 'chaos'