Political polls with the headlines and column inches they command. A cynic would say it probably depends on whether you're going up or you're going down. There's a story in Newsroom today from Professor Ben Gray that looks at how the different companies poll the population to get their results. And it doesn't give you a great deal of confidence, does it? He cites the New Zealand Health Survey as a good example of a high-quality survey because the selection of households is random from census data, representing the total population by age, sex, region, ethnicity, whether Māori, Pacific, Asian, other, and deprivation. So, if the population is made up say of 20% Māori, then they'll make sure that they have 20% of the respondents will be Māori. If there is 51% of the population that's female, they'll make sure that 51% of the respondents are female. Interviews are in person with multiple callbacks, resulting in a 75% response rate. Interpreters are used for the 5% of the sample not fluent in English and represented in the population.  

The methodology is approved by an independent research ethics committee and complies with National Ethics Advisory Committee research ethics guidelines. Political pollsters on the other hand, all use panels, online or telephone, who may or may not be paid, from which they claim to select a random sample. None provide any information about how their panel is recruited. It could be people who are self-selecting and they say, well, we make sure that we get 20% who self-select as Māori. I don't think that's true and they don't show you so you can't test it. 

In the past, apparently, New Zealand had a good reputation for accuracy in political polls when compared to other countries. We couldn't lie to save ourselves. Wouldn't have even occurred to us. You ask us who we're voting for and we'll tell you, without a word of a lie, it will be our absolute intention to vote for this. But declining trust in institutions and the media, as well as the impact of social media, is starting to erode that. We just don't trust the media, we don't trust institutions, we don't trust politicians, we don't trust pollsters. And political scientists point to sensationalist media reporting, bam, kapow, freefall, as being "not very helpful from a scientific point of view", from Professor Grant Duncan.  

But then when a media organisation is paid top dollar for a survey, they'll want to milk the results for all they can for headlines and by crikey, they do. 1News a couple of years ago pointed out that an analysis of the final 1News poll, the final poll before each of the last eight elections, shows the last survey of the campaign had been with an average of 1.17 percentage points of the actual party vote, according to the accuracy data. They say we're actually pretty accurate. 2020 threw everybody. That was the one that knocked most of the polling companies out of the water because it overestimated how much of the vote National would get, underestimated the love affair the nation had with Labour. 

But for 1News, they say with their pollsters Verian, the figures covering every election from 2002 to 2023 translate to the poll just being up by one to two seats per party, which they say shows it's pretty accurate, and it is. But do you care about them? They must have an impact. They absolutely must. Because if you see a poll showing the Opportunity party at 3% and you're thinking, oh, quite like the new shiny thing, oh, but at 3% my vote will be wasted, I'll go with the Greens. But then you see a poll showing the Opportunity party at 6% or 7% and you think, oh, oh, that's good. I'll vote for them. I'll vote for the shiny new Greens. Surely, it must have some sort of influence on how you are going to vote, especially if you're not 100% sure and a lot aren't. Now, the latest one that came out, that sort of I suppose it was a fortunate accident that we had the Newsroom story at the same time as we had the Freshwater poll come out, showed that Labour had dropped seven percentage points, which seems ridiculous. 

I mean, I totally get that National and Labour are unappealing to the electorate. We've seen that over a number of different surveys. People aren't enthused about voting for either of the main parties. They're flirting with the minor parties and having a look at the Opportunity party. But seven percentage points, which either means the poll was overinflated the first time or people are just having a laugh. I have never been asked in a survey. I've never been randomly selected —people want my opinion, they can tune in between 9am 12pm Monday to Friday on Newstalk ZB— but those who have say the questions are so weighted that they are trying to get a response that they want. It's not a "random question, who will you vote for?" Which I imagine a survey would be. No, nothing like it.  

Do you look, especially if you're flirting with some of the minor parties, at the results of a poll and think, hmm, my vote might be wasted, I'll go somewhere else? Do you look at your favoured party perhaps dropping in the polls and think, right, I'm going to make sure it's two ticks now, not just one? I don't know. I mean, they certainly don't influence me. I get where the political scientists are coming from saying, look, we're just producing a random sample, it's the media who ramps them up and milks every last piece of sensational headline out of them. But do they matter to you? Is there any point in publishing the bloody things other than making it an easy day at work for journalists?