With both Galaxy and Newspoll coming in at 51/49 to the LNP, it’s highly probable that there hasn’t been much in the way of movement in public opinion since the campaign began – at most the movement is a few points either side of that central 51/49 estimate. We can pool the sample of the latest Galaxy and Newspoll results (where we simply combine them and weight the pooled poll by sample size), it gives us a point estimate of 51/49 with a margin or error running at 2.4%.
If we go to the base simulation result which provides the implied probabilities of a Labor victory from a given TPP result (with the assumption of a uniform swing and a standard deviation on that uniformity of 4%), it suggests that a 51/49 result would give the ALP around a 66% probability of holding 45 seats or more in the new Parliament and retaining government.
However, any small movements in that two party preferred result would be expected to have a high impact on the implied probability of Labor retaining government. If we take the first difference of that above chart, where we will effectively measure the change in probability of the ALP retaining government for every 1 tenth of a percent change in the two party preferred result, it shows us where the “danger area” is for the ALP in terms of delivering (or not) the seats it needs to form government from the vote level it receives.
The way to read this is a lot simpler than it looks. To use an example, if the ALP vote changes from 48.5% to 48.6%, the implied probability of a Labor victory increases by just over 5%. 5.1% to be ridiculously exact at the change from 48.5 TPP through to 48.6 TPP to Labor.
You’ll notice that the chart peaks between 48.1% and 49.1% two party preferred (which represents the TPP from 48% through to 49%) – this is telling us that the ALP is expected to make the highest implied probability gains for victory as it moves from 48% through to 49% two party preferred.
Since there is a larger downside risk for the ALP probability of victory in the 48% to 49% TPP range than there is upside risk at the 49% to 50% range, this reduces the implied ALP probability of victory with a polling result of 49/51 compared to what it would be with an election result of 49/51, since polls have sampling error that fall randomly either side of the mean.
As the true level of political support could be a few percent either side of 49/51, yet the downside risk for Labor is greater in size than its corresponding upside risk at a TPP poll result of 49/51 – the current state of political opinion suggests that the ALP is actually on an implied probability of victory of 55.7% – much less than 66% probability an actual election night result would produce if the vote were to come in at 49/51 to the LNP. We can see that by rerunning the base simulation and including the uncertainty of the pooled poll results.
It’s worth pointing out very clearly here that these implied probabilities can’t be read literally – especially at tenths of a percent level. The point of this isn’t to whack an implied probability on some given X point Y % two party preferred result, but to highlight the sensitivity of the seats a given TPP result would be expected to deliver to the ALP.
At less than 48% TPP for Labor, the LNP would be expected to form government– even though it requires a slightly higher vote than 52% on the standard pendulum – but as the ALP vote increases from 48% to 49%, the likelihood of an ALP victory grows dramatically – very dramatically. Every tenth of a percent higher the ALP can get its vote above 48, the returns in seats should be disproportionate.
So while this makes the underlying political situation effectively closer than a 51/49 LNP lead would ordinarily suggest, considering where the seats sit on the pendulum, it also suggests that the ALP will probably start trying to hold the line, which in ordinary parlance means going negative and targeted. The ALP internals would be telling them essentially the same thing we’ve figured out here, that they need to hold their current vote level of around 48 to 50% TPP and grind out a victory by flipping 1% of the electorate. That means they’ll probably start to cut voting blocks and geographical areas loose from the campaign (some of the regions and parts of the Gold Coast come to mind). They’ll probably start targeting key demographics – suburban women especially since the LNP are urban female weak and the Premier is a women – but most importantly they’ll start to go negative, since the Borg’s biggest weakness is his high level of Undecideds.
Undecideds are fairly easily flipped by incumbents – Beattie did it in the lead up to the last election, Howard did it with Latham and did it again with Rudd. Mostly, undecideds are flipped or crystalised into support for a party in net terms by fear.
Fear of incompetence was the Beattie weapon, fear of economic mismanagement and fear of the unknown was Howard’s weapon against Latham, and fear of change itself gained Howard a few points on Rudd in the last week of the campaign (or fear of unionists turning out the lights of dress shops! 😀 )
So the question for you ALP folks – if you had to launch a fear campaign, which demographics would you target, and what issue would you target them over?
For the LNP supporters, what would you do to counteract the inevitable campaign, and reduce Springborg’s undecideds in your favour?