The Australian has again given us a polling fix for the summer break with a Newspoll quarterly state voting intention poll for Queensland. The poll was taken progressively over the October-December period.
The topline primary vote figures for the poll of 1,139 Queensland voters are LNP 42%, Labor 31%, Green 8%, KAP 4%, Others 15%. The TPP estimated by Newspoll was 56-44% LNP. The full poll details and the accompanying article can be found at The Australian.
Newspoll’s Low KAP Vote
Once again with Newspoll’s quarterly poll, the KAP primary vote score is relatively low and the Others primary vote score is relatively high. Below is a table of all other polls taken during the October-December period.
In the above table we have taken an average of the four polls and compared it with the Newspoll. In terms of the average KAP vote in the four other polls taken over the same period as today’s Newspoll (3 ReachTEL & 1 Galaxy), the KAP vote was nearly 6 percentage points higher at 9.8%, while the Others vote was lower by approximately the same margin at 6.1%.
If we look at the KAP/Others vote split in another way, the KAP vote as a share of the KAP + Others vote in the other polls taken between October and December is 61.4% (9.75/15.87). In today’s Newspoll the same share is only 21.1% (4/19).
So what is behind the low KAP vote? (to the benefit of Others). As only Newspoll has returned low KAP primary votes this year from all the polling firms which have surveyed Queensland state voting intentions, the answer lies in Newspoll’s methodology. The true answer can only be provided by Newspoll. However, it is this blogger’s educated guess that the low KAP vote in the Queensland quarterly Newspoll is a result of the KAP not being explicitly included as a party option in the voting intention question by the telephone interviewer. The KAP result in the Newspoll would be reflective respondents who verbally respond Katter’s Australian Party to the question despite not being given the express option by the interviewer.
Once again this is purely speculation, only Newspoll really knows the answer to the low KAP vote. Furthermore the primary vote scores for the majors and the Greens do not seem to be anomalous.
As usual we will run a state seat projection based on today’s Newspoll figures. Our methodology for the Queensland state seat projection can be found here. The seat projection is below.
The above seat projection returns an LNP majority of 29, which is the largest LNP seat projection majority since the 14 October ReachTEL poll. The above seat projection takes the Newspoll primary votes as a given. However what if we try to adjust the KAP vote to reflect the larger share of the KAP + Others vote we see in other polling firms?
The combined KAP + Others vote in the Newspoll at 19% is the largest of all the polls in the same period which had ranged from 14% in the November 24 Galaxy to 17.8% in the December 16 ReachTEL. To assess the true KAP vote we shall apply the average KAP share of the KAP + Others vote found in all polls ex-Newspoll in the October-December period to the KAP + Others vote found in today’s Newspoll.
From this process the adjusted KAP vote is 11.7%, while the adjusted Others vote is 7.3%. If we use these figures with the remaining primary votes we should get the following seat projection.
The seat projection off the adjusted Newspoll is close to the seat projection from the ReachTEL poll of 16 December. The LNP gains 2 seats from the 16 December projection at the expense of both the ALP and the KAP who lose one seat each.
As we did yesterday with the SA seat projection we will provide the seat ranges and hypothetical election outcome probabilities for this Newspoll . We have provided the seat ranges for both the unadjusted and adjusted Newspoll. The seat ranges are based on a 95% probability threshold and are set out in the table below.
It should come as no surprise that the hypothetical election outcome probabilities for both the adjusted and unadjusted Newspoll are a 100% probability that the LNP would win majority government if the poll figures were repeated at an election. Such a statistic merits no graph.