Change you can’t believe in
From One News and a Colmar Brunton poll about Judith Collins and the Oravida affair
Which of these statements best describes how the issues will influence your vote in the upcoming election?
23% These issues will be a factor in your decision about who to vote for
75% These issues will not have much influence on your vote
1% Don’t know/won’t vote
Graeme Edgeler pointed out on Twitter that it matters what starting position people are being influenced from. That information wasn’t in the Colmar Brunton summary, because reporting it would also involve reporting the split of party affiliations in the sample, and the poll wasn’t designed for that split to be a reliable estimate.
I’m not going to report the split, either, but you can get it from the detailed poll report. I do think it’s reasonable to note that among people who identified as Labour/Green voters, about 1/3 said it would influence their vote, and among those who identified as National voters, less than 10% said it would influence their vote. The difference is more than twice the margin of error estimated from those proportions and numbers. Looked at the other way, three-quarters of respondents said the issue would not make much difference to their vote, and three-quarters of the rest were Labour or Green voters.
It’s not impossible for Labour or Green supporters to have their votes influenced by the Oravida affair. You could imagine someone with a long-term philosophical or emotional attachment to Labour, who had been thinking of voting National at this election, but who decided against it because of the scandal. But if there are enough people like that to show up in a poll, the left-wing parties are in real trouble. It’s more likely that most respondents said whatever they thought would make their side look good.
Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »
Although… Trying to make your side look good is a valid reason for voting for a particular party on Election Day. :)
11 years ago
Sure. It’s not unreasonable behaviour by the respondents, but it does make the headline percentage a bit pointless.
11 years ago
I think in any political poll, for any supplementary question referring to a political party, people will answer in a way that makes their party look good.
Unless, that is, the issue is important enough to them to influence their vote. The results suggest to me that this is not one of those instances, for a lot of people.
11 years ago
Indeed. Which is why it would be better to do a poll that allowed for the party split. The results would then be less interesting, perhaps leading to the conclusion that the question wasn’t worth asking.
It would be nice to know if National’s lost many votes because of this, but I don’t think it’s feasible to tell by polling.
11 years ago
In think it’s feasible, but would take a longer interview.
11 years ago
To me support for a party is a continuous variable but the outcome seen is a discrete one. So a Labour/Green voter could have had luke warm support for Labour/Green but this “scandal” could have made them hard line. That voter would felt the “scandal” had influenced their vote even if the outcome of their vote hasn’t changed.
Alternatively, the “scandal” could have changed a voter from Green to a Labour supporter (or vice versa) based on the way the two parties handled the issue.
There is also the issue of how the pollsters decided who was a National/Labour/Green supporter. Was the question 1) “who are you going to vote for at the next election” or 2) “who did you vote for at the previous election”. My suspicion is the former (the latter question is old news so why ask it) – which means people who have been influenced by the “recent news” to change their vote would change from National to Labour/Green i.e. they’d be recorded as a Labour/Green supporter so you would expect to see a higher proportion of Labour/Green supporters having the “recent news” influence their votes.
However, if that were true it would also be expected to see more Labour/Green supporters than Nationals supporters but at 153 for the former and 222 for the latter something looks decidedly odd.
What you would expect with 20% saying don’t know/not voting and using the latest RM poll, there should be roughly 170 National supporters and 182 Labour/Green supporters in the Colmar-Brunton poll.
11 years ago
Hi Megan
You’d also need to filter out those who are unlikely to vote. Once you factor that in, and also weight the data, and also consider that the two polls you mention have historically not tended to show similar results, then those raw numbers may begin to look a little less odd.
***
Hi Thomas
I don’t agree with you that the poll question was worthless. I actually think it was really interesting, especially when considered alongside some of the other results in the poll. Forgetting about party support for a moment, what we have is a situation where a large bunch of eligible voters think the whole ‘Judith Collins/Oravida/comments about a reporter’ thing is actually a fairly big deal – 42% think she should not remain a minister and 50% think these issues will have damaged National’s level of public support.
However, when asked if these issues would be a factor in their *own* voting decision, most say the issues won’t have much influence. Note that the question did not ask people if they would *change* their vote, it asked whether these issues would be one of the issues they would consider in their decision.(+) There are many other issues, of course, such as education, jobs, housing, child poverty, crime, and the list goes on.
What interesting things could we take from this combination of results?
I think what’s interesting is the results suggest that eligible voters can consider an issue to be a fairly serious one, but they may also be able to separate it from their overall assessment of a party. There are many things that may help this separation along – a person may quite like the Prime Minister, or may not yet see a viable alternative to National, or may simply see other issues as even *more* serious that this one.
So I don’t think the question is worthless. I actually think it helps to illustrate the complex nature of political preferences. I’m not sure your title ‘change you can’t believe in’ is really a fair reflection of the question that was asked.
(+)This is the reason why I’m not at all surprised that a higher number of Labour/Green supporters say it will be a factor in their decision. Like Megan, I see support for a political party as a continuous attitudinal variable. I think these issues may have helped to move some Labour/Green supporters further along the continuum in support of the Labour and Green parties.
11 years ago
So how did you decide who was a Labour/Green supporter and who was a National supporter?
11 years ago
I decided not to filter out the “will not vote” because no poll every does (mainly because people generally don’t tell the truth on this issue).
Polls do filter out the “don’t knows” which are running at 20% and so I did. And if you look at the two sums
170+182
[1] 352
222+153
[1] 375
My total counts for Labour/Greens is reasonable.
11 years ago
Hi Megan
Yes your counts are totally reasonable given your assumptions. My post was not criticism at all – just additional information.
We did exactly as you suggest to determine current party supporters, plus we also probed ‘don’t knows’ for a general preference. When calculating party support normally we filter out the unlikely voters, and those that won’t vote, and weight. Plus our numbers have always been quite different to RM, it’s maybe not a good idea to expect we would show a similar result to them at any given time.
Our raw counts were within the range we expected.
11 years ago
That should have been…
My total counts for Labour/Greens and National is reasonable.
11 years ago
@Andrew
So I finally get your point about the weighting – you were using raw numbers not weighted numbers of National and Labour/Green Supporters.
11 years ago
So I’ve blogged on this, if anyone is interested.
Joyce gets it wrong, and all we see are the trees –
https://grumpollie.wordpress.com/2014/05/11/joyce-gets-it-wrong-and-all-we-see-are-the-trees/
11 years ago