Can anyone spell “Too close to call”?
The Wentworth By-election Quick Poll was conducted over four days from October 6 to 9, with the aim of finding better ways to poll at the seat level. We did not get the larger sample that we wanted to be able to test whether a larger sample size was needed for more accurate polling at the seat level, but at 736 (after we cleaned out everyone that didn’t actually live in the electorate and just wanted to play along too) it is a statistically representative sample with a margin of error of about 3.5%.
The two candidate preferred/two party preferred figures are not the result of an equation, guess, our usual inferred method, nor a direct preference question. We asked people to rank the top candidates in order, and simulated a preference count exactly the same way they do it on election night.
Note that there is a strong flow of preferences to Heath then Murray, while both Phelps and Sharma get very little from the preference flow. (And, Wentworth is not following how to vote cards. In fact many respondents said exactly that in the comments.) These results – remembering it’s just a poll, which is indicative, not predictive, with a margin of error to consider – put Phelps and Murray neck at neck at the final exclusion round. Either one of them will win, but which one is completely unknown.
Although, with that very healthy flow of preferences, if Murray gets the 25% of primary votes some other polls have had him at then you’d have to think he’d be the (surprise?) victor.
Poll watchers may note some unusual weightings. I hate all the polling coming out of Wentworth right now, there’s something wrong. Wentworth is a very difficult seat to poll, given the many sub-cultures (Jews, Surfies, Bondi bubble, Bayside mums, LGBTIQ… ) and the village life that promotes a very parochial sense of identity. There is not a Wentworth community, there are the many communities of Wentworth. Some of them are geographically centered, some aren’t.
So I started with the usual weighting by age, gender and 2016 vote. Numbers still looked very wrong. Added weighting by suburb – remarkable difference… but still not quite there. After a little insomnia and far too much comfort food, I tried something new. I have been working for about a year on a method to track vote intent volatility – or how committed people are to their vote. In this poll I used a different form of the scale than I have been using the in the main study, but the point is to try and capture the underlying current, the rip beneath the water if you will, to detect where in the electorate those inevitable late shifts will occur. I’ve never considered using the volatility score as a weight before, but I tried it – and those numbers I liked as it was more reflective of the sentiment in the comments and the qualitative evidence from our detailed research – so that is what I’m presenting here. Essentially, by adding volatility as a weight, it discounts those who may change their vote to, say, half of someone that has completely committed to who they are going to vote for. We’ll find out in about a week whether it was a good idea or not.
The reasons given for people’s votes were not substantially different from our detailed study. Phelps voters either like her positions or are ‘sending a message’. Heath and Murray voters like their candidates personally, or that they’re local, and some ‘time for a change and Phelps has stuffed up’ protest voters. Sharma voters want to protect the stability of the Government. There are partisan loyal voters for all parties too.
We do like surprises here at the Voter Choice Project. Today’s surprise is that climate change and refugees are not the most important issues on the minds of Wentworth voters. We asked participants to rate each of the top 10 most frequently mentioned issues from our detailed Wentworth By-election study, from very important to not at all important to them. Climate change had the largest number of people answering ‘not at all important’ at 9.8%, and a combined not important total of 27.5% of voters. Asylum seekers and refugees came close, with 27.3% of voters saying it wasn’t important to them, and welfare was the only issue that performed worse, with 30.5% of Wentworth voters saying the issue was not important to them.
The highest scoring issue was the economy, with over 95% saying it was important to them. It’s the economy stupid. It was very closely followed by health, at 94% important.
Thanks to everyone who participated.
