The UK’s vote to leave the European Union (Brexit) was known by the afternoon of the 24 June Melbourne time. The ReachTEL and Galaxy polls in the table were both taken entirely before the Brexit result was known. The one point gain for the Coalition in Newspoll, which had three of its four fieldwork days after Brexit, is probably attributable to Brexit. Essential failed to change, but nothing should be read from that as Essential’s polls are too stable.
Despite a primary vote drop for the Coalition, ReachTEL’s headline two party vote, based on respondent allocation, was unchanged at 51-49 to the Coalition. Previous election preferences can be calculated from the decimal primary votes, and were also 51-49 to the Coalition. The fall of the Nick Xenophon Team (NXT) to 1.4% is because ReachTEL has excluded them from the readout beyond SA, as they are only running candidates in 7 of 139 seats outside SA.
Brexit benefits the Coalition for two reasons: the economic uncertainty helps the Coalition because they are widely regarded as better economic managers, and the “don’t risk the minor parties” message is strengthened.
Labor’s only real chance now is the hope that Brexit has produced a transitory spike for the Coalition, which will fade by election day. However, the costings debate is also likely to damage Labor.
Kevin Bonham’s poll aggregate is now at 50.6% Two Party Preferred (2PP) to the Coalition, a 0.1 point gain for the Coalition since last week. The Poll Bludger’s BludgerTrack is now at 50.9% 2PP to the Coalition, a 0.7 point gain for the Coalition. Primary votes in BludgerTrack are 41.7% for the Coalition, 33.7% for Labor, 10.5% for the Greens, 4.1% for the Nick Xenophon Team and 10.0% for all Others. Since last week, the Coalition’s primary vote is up 1.4 points.
The last Morgan national poll that I have seen was conducted in late May. Apparently Morgan is now only doing seat polls, but these seat polls have either been taken over too long a period (since January) or the sample size is too small to be credible.
I will do a post on the final pre-election polls on Saturday morning, and a wrap of the election on Sunday.
All of these seat polls were taken prior to Brexit.
A ReachTEL poll of the Nationals-held seat of Cowper showed Independent Rob Oakeshott at a 50-50 tie with the Nationals. This is a surprising result as Oakeshott supported Labor in the 2010-13 hung Parliament, and did not decide to contest until the last minute.
ReachTEL has polled all five Tasmanian seats, and has Labor winning Lyons from the Liberals 55-45, while Liberal-held Bass and Braddon are lineball. Labor will hold Franklin, and Independent Andrew Wilkie will hold Denison. These polls were conducted 23 June with samples of about 550 per seat.
While ReachTEL’s recent seat polls have been encouraging for Labor, Galaxy’s were disappointing. In particular, Dobell, Lindsay and Gilmore were all at 51-49 to the Coalition according to Galaxy, while ReachTEL polls conducted 20 June had them all going to Labor by at least 53-47. The Victorian Galaxy seat polls had little swing, perhaps owing to the Victorian government’s dispute with the Country Fire Authority. The Galaxy seat polls were taken 20-22 June with at least 500 sampled per seat.
In ReachTEL, Turnbull’s (total good) minus (total poor) rating increased 9 points in a week to a net zero, his best since February. Shorten’s equivalent rating increased 3 points to -7.
In Newspoll, Turnbull’s satisfied rating was 37% (up 1), and his dissatisfied rating was 51% (steady), for a net approval of -14. Shorten’s net approval was up one point to -15. While the Coalition gained, Turnbull did not, and this gives Labor some hope that the Brexit effect is transitory.
In Essential, Turnbull’s net approval was zero, up two points since last fortnight. Shorten’s net approval was -2, up four points.
In ReachTEL, Turnbull led by 17 points as better PM, up from 15 last week. In Newspoll, Turnbull led by an unchanged 15 points. In Essential, Turnbull led by an unchanged 11 points.
Notes on these polls
In Newspoll, 51% thought the Coalition would win, and 27% expected Labor to win; this was 44-33 in favour of the Coalition in late May.
In ReachTEL, 50% expected the Coalition to win, with 22% backing Labor. 48% thought Turnbull would not privatise Medicare, and 36% disagreed. The Coalition’s advantage on border protection increased from 16 to 21 points.
Galaxy found that Labor would be ahead 53-47 had Abbott still been PM. 38% believed Labor’s claim that a Coalition government would privatise Medicare, and 45% did not. 30% believed the Coalition’s claim that Labor’s negative gearing reforms would drive down house prices, and 40% did not. The last Galaxy was taken following the 3 May budget.
In Essential, 30% thought the Liberals had run the best campaign, 28% Labor and 8% the Greens. 39% expect the Coalition to win the election, 24% Labor, 16% expect a hung Parliament and 21% are undecided. 50% thought it likely that the Liberals would attempt to privatise Medicare if they won, and 34% thought it unlikely. By 63-18, voters would support phasing out live exports.
Vote counting procedure
At 6pm local time on 2 July, polls close and vote counting begins. Primary votes are counted first in all booths within an electorate, then there is a two candidate count between the candidates the electoral commission (AEC) thinks are likely to be the final two in each seat. Normally this choice is clearly between both major parties, and sometimes between an Independent or minor party and a major party candidate.
Occasionally, the AEC gets the final two candidates wrong on election night, and in that case, they go through the electorate booth by booth after election night with a preference count between the correct candidates.
All votes cast by voters enrolled in an electorate at that electorate’s polling places are counted on election night; this includes pre-poll votes cast at that electorate’s pre-poll booths.
Postal votes are not counted on election night, and are slowly processed following election day. Polling day votes that are cast outside a voter’s electorate (absent votes) are also counted slowly, as are pre-poll votes cast outside the voter’s electorates. In Australia, postal votes skew towards the Coalition, but absent votes skew towards Labor.
For the Senate, only the number of “1” votes will be counted for each party on election night. Following election day, preferences for all votes will be data entered into a computer system for each state and territory. When all votes have been entered, the “button” will be pressed, and the computer will quickly determine the successful candidates.
Authors: Adrian Beaumont, PhD Student, School of Mathematics and Statistics, University of Melbourne