Well, we had fun* crunching the KORE Poll this week. It appears that protesters in Melbourne were doing the poll while killing time hanging around the city, and thought answering every question with expletive-laden comments would be a great way to say how they felt about Dan Andrews. Um, ok.
Not to worry, this is why we use multiple sources of recruitment (half the KORE Panel, half recruited via email, social media, and a panel provider). After discarding more than 400 responses as noise, the usable replies were reduced to 889, which means we have a smaller sample than we’d like and a high MOE of 3.5%, but the numbers are still usable. The remaining respondents were reasonably well balanced, representing voters from 125 of the 151 electorates, with our normal demographic skew of older men being slightly over-represented and younger women being under-represented. Responses were weighted by age, gender, and state. However, the two measures we are trialing to replace 2PP – Effective Vote and Incumbent v Challenger seem to be pointing in opposite directions.
First Preferences
First preferences continue to be about where they should be for this point in the cycle with no party being particularly inspirational. Major primary vote in the low 30s, high support for minor parties, and a whopping 8% undecided.
Effective Vote
Effective vote is calculated using the Hypothetic 6 question of asking respondents to rank 6 hypothetical candidates – ALP, Coalition, Greens, a moderate Independent, Minor Left and Minor Right party – and then effecting a preference vote for their electorate, based on the two candidate preferred count in the 2019 election. This is an attempt to find something better than the 2PP and properly account for the increasing number of seats that are not a battle between the two majors.
The effective vote gives us an indication of the likely seat breakdown, and thus is a better indicator of likely winner than the 2PP.
| EFFECTIVE VOTE | Percentage | Seats |
| Australian Labor Party | 51.4% | 78 |
| Liberal Party/Nationals/LNP | 44.5% | 67 |
| Others | 4.1% | 6 |
The Effective Vote measure indicates a clear Labor win, but by no means the landslide indicated by the 2PP numbers bouncing around at the moment.
(Note: The hypothetical 6 works exceptionally well in every seat except Mayo, where some voters consider Bek Sharkie an independent, while others view Centre Alliance as a minor party, but there is no agreement on whether it is left or right. We use their first preference vote for Mayo voters, as almost uniformly they are voting for Bek or they’re not.)
Incumbent v Challenger
The other model we are testing this cycle to see if it is superior to the 2PP is a simple Incumbent versus Challenger measure. To get this figure we code if the respondent’s effective vote is for or against the sitting member. If the incumbent figure is above 50% then the current Government is not likely to change hands.
Hang on, the Effective says Labor win, in the IvC says no change in Government?
Yeah. We triple-checked and dove deep to figure out what was going on, and the best explanation we have is this: the most committed Labor voters in the sample are in Labor-held seats, and all 6 of the crossbench seats are extremely committed to returning their sitting member. Usually, the two numbers are pretty consistent at least in the direction of what’s going on – but if you take a step back, all these numbers are within MOE of a dead heat.
We’ll see how this plays out throughout the election, but the IvC correctly projected no change in Government from as early as January in 2019.
Introducing the Momentum Tracker
The last 5 years I have been working on a system or method to detect momentum – where the vote is going to go. In the Voter Choice Project this was a stunningly good performer, but I didn’t trust it or believe it and therefore didn’t publish it. This time, I’m publishing it – but I have tinkered with it a bit, and have no idea if it is right, we’ll find out after the election I guess.
The way it works is we calculate a score of how committed each respondent is to their declared vote intention. That score puts them in one of these 5 ranges:
- Uncommitted – the uncommitted voter is either so uninterested or unimpressed by the election there is a very high likelihood they will change their vote or not vote at all
- Soft – this voter may need a reason or catalyst to change their vote, but are actively looking to switch.
- Medium – if some significant reason presented itself they’d consider switching their vote.
- Firm – a near-certain vote. The voters may switch if the party or candidate they were intending to vote for did something stupid, unlikely to be swayed by another candidate or party.
- Hard – a locked-in vote. Nothing will budge it.
The goal for the campaign strategist is obviously to lock in and firm up as many votes as possible, and reduce the number of soft votes in the process. The theory I’m working on is that the parties with more firm voters than soft voters, and an increasing hard vote poll on poll, have the momentum in the election. In the simplest explanation possible of why:
- The only thing that effectively moves votes is personal conversations and political discussion. Ads, news and other media coverage, and everything else can keep issues salient and start conversations, but they have minimal effect in changing minds.
- Firm and soft voters are the only two groups engaging in political discussion. Firm voters are more likely to reassure and retain soft votes in social discussion: Soft voters are likely to put doubts in the otherwise firm voters’ minds. As long as the firm number is bigger than the soft number, the party’s vote share should continue to firm up.
- Moderate and uncommitted voters are unlikely to engage in the debate either way – neither seeking reassurance nor giving it. They will pick up on other’s discussions, so if there are more firm voters than soft, they too will become firmer in their voting decisions.
- Hard voters are likely to spruik their side more than engage in discussion. Declining hard votes indicate the base is waning, the party is in trouble with its own, and the vote is likely to collapse quickly as the most loyal party soldiers stop talking up their side.
So, In this grid of numbers, the two you really want to look at are soft and firm. ALP has more firm than soft; Coalition has more soft than firm, so the ALP has the momentum. Greens are in neutral with their soft and firm voters at exactly the same level.
The PHON and UAP numbers are very interesting and why we love the insights provided by the Momentum Tracker. If you look at the first preferences, it looks like PHON’s a good point ahead of UAP, and should do better in the election, right? But with so few hard voters, there’s no one spruiking for the team and thus the PHON vote is likely to collapse. Meanwhile, the UAP vote is almost entirely Hard and Moderate voters, with very few soft or uncommitted. This means that (if the theory is correct) the UAP vote may hold – those two groups aren’t engaging in discussion and are unlikely to change – but it won’t grow. Both because there’s so few soft voters to firm up, but so few firm voters engaging with them to convert them. So no matter how much advertising the Clive and Craig show does, that number doesn’t move. This is exactly what we saw last election just by the way.
The momentum tracker numbers get a bit more exciting next month, when we can show movement. And hopefully have a much bigger sample without so much noise.
Issues we cared about last election
This month we asked about issues that were big in the 2019 election: corruption and a federal ICAC; Indigenous reconciliation and the Uluru Statement from the Heart; Climate Change and Adani. These questions were about testing how much the baseline has shifted thanks to the pandemic.
The level of concern about all three issues has softened since 2019, however, this should be expected given COVID-19 is sucking a lot of attention away from literally everything else, and the heated debate about these issues isn’t there to keep them top of mind.
Corruption was still somewhat top of mind given the discussion of Christian Porter’s anonymous donation, but there was a definite partisan factor not present in the 2019 figures: almost all who weren’t concerned are Coalition voters.
- 59% are extremely concerned about the influence of political donors, down from 76% in 2019
- 59% are extremely or very concerned about corruption in politics, down from 80%
- 78% support a federal ICAC body, down from 88% in 2019.
The figures on Indigenous reconciliation are, I’m sure, saddening for many, but this is what happens when you stop talking about an issue. When was the last time you heard or saw any discussion of the Uluru Statement from the Heart in the news?
- 51% generally support Indigenous reconciliation, down from 90% in 2019
- 52% support the Uluru Statement from the Heart, down from 86% in 2019
On climate change and Adani, we actually did some A and B testing to measure to what extent people had kept pace with the issue. Half of the respondents were asked about the Adani mine construction; the other half were told that Adani had changed their name to Bravos and the mine construction had begun. Interestingly, only 3 of the 389 respondents to the first version corrected us that the company had changed it’s name and the mine was near operational. The rest responded exactly the same way they did in 2019, emotively, opposing the concept. In the second group there was much less emotion or carefactor about the mine.
However, both groups responded quite equally when asked about climate change, and of the three issues, these numbers are the most stable.
- 54% oppose the Carmichael mine, down from 68% in 2019
- 42% believe the mine will create jobs, up from 31% in 2019
- 55% say that taking action on Climate Change is extremely or very important. A slightly different question was asked in 2019, finding 58% cared about Climate Change as an issue.
The issue hierarchy
Finally, respondents were asked to indicate whether a range of issues were more or less important than COVID-19 management. There were lots of complaints about this idea that issues could be compared, but the insights are valuable. And surprising.
The issues have been ranked putting issues the are more or as important as COVID-19 at the top. The first surprise is the issue that was deemed more important than COVID-19 by the largest number of people – 54% – is agriculture and food security. No, the sample isn’t stacked with farmers (yes we checked). This is the clearest indicator of just how traumatic the pandemic has been for people… the rushes on supermarkets each time a lockdown was announced have really affected world views. People are worried about whether they can get food.
Aged Care and Domestic and Family Violence are the only two other issues that had a majority ranking them above COVID-19. That none of the issues posed, not even internet access, had a majority ranking it as less important than COVID-19 is also telling that people are *over* COVID-19, and want other issues dealt with.
Respondents were also asked if there are any particular issues they think are significantly more important than the management of COVID-19 which they would like to see discussed in the election campaign. The most common response was that people were struggling to separate COVID-19 from all the other issues, and that’s a great answer, as COVID-19 has exposed many cracks in all our systems. The clearest answer was those who wanted action on Climate Change, but not as many as you’d expect, only about a dozen respondents went to the trouble of writing it in.
There was also a significant number of people wanting a vision, wanting to know what the plan was post COVID-19. Me too.
KORE Poll number 3 will be out on October 14. It won’t be asking about COVID-19, and we’re going to do some recruiting in the meantime to see if we can avoid the sample issues we had this month. Please join the KORE Panel to make sure you get your survey.
~RKC