It's time for a complete set of updated polling charts. We will start with the primary votes. What is important to note here is that both Labor and Coalition have lost primary votes to independents and other parties in past 6 months. While most pollsters have Labor as the clear favourite to win the next election, if these primary vote polls are accurate, it would be an historic win. Labor has not won government with less than 38 per cent of the primary vote in the past (well 37.99 per cent in 2010 to be precise).
When it comes to the two party preferred (2pp) votes estimates, the different pollsters use different approaches. Resolve Strategic does not calculate a 2pp estimate. Roy Morgan relies on respondent preferences. Other pollsters use preference flows from the most recent federal election. For consistency, I have calculated a 2pp estimate for each poll based on the primary vote poll results and preference flows in 2019.
The most recent published Roy Morgan polls point to a Labor landslide (although this reduces somewhat when I calculate the 2pp based on preference flows in 2019). The most recent Newspoll and Essential polls suggest a comfortable Labor win. The Resolve Strategic polls suggest a narrow Coalition win. Interestingly, while the pollsters had similar estimates to each other in the first half of 2021, the pollsters are more diverse now.
The following charts start with the pollsters estimates, and my calculations of 2pp based on the most recent election. We then look at the exponentially weighted polling average by pollster. Finally, I use a Bayesian implementation of a Gaussian state space model to aggregate my estimate of the Coalition's 2pp based on the 2019 preference flows and the assumption that the polling house effects sum to zero.
The attitudinal polling follows.
Note: the data for these plots was sourced from Wikipedia. The code for producing the plots is available on my GitHub site.