The Federal election is today. Polls close at 6pm Melbourne time in the eastern states, 6:30pm in SA and the NT, and 8pm in WA.
All final polls except Ipsos have the Coalition ahead. Here is the final poll table, sorted by first fieldwork date. The final Newspoll had a massive sample of over 4000.
Respondent allocated preferences in Ipsos were 51-49 to Labor, one point better for Labor than the previous election method. Labor is also doing well on respondent allocation in Essential. However, ReachTEL’s respondent allocated preferences were the same as the previous election preferences for the second consecutive week. Previous election preferences can be calculated from the decimal primary votes.
If Labor does better than expected, it could be because they received better preference flows than at the 2013 election. Inflated win expectations for the Coalition may also benefit Labor, as some people may think they can lodge a protest vote without removing the government. In particular, the hard right would like Turnbull to get a bloody nose without losing office.
Another way the polls could be wrong is herding. All final polls are between 50 and 51% Two Party Preferred (2PP) to the Coalition, with four between 50.5 and 51% 2PP. The range of results is too narrow to occur purely by random chance, and implies that polls are herding to a consensus result. If the actual result is in that consensus, the polls will all look good, but if not, they will all look bad. It is possible that either side could perform much better than polls indicate.
Kevin Bonham’s final poll aggregate is at 50.8% 2PP to the Coalition, and he is predicting a seat outcome of 78 Coalition, 65 Labor and 7 crossbenchers.
The Poll Bludger’s final BludgerTrack is at 50.9% 2PP to the Coalition, from primary votes of 41.7% for the Coalition, 33.4% for Labor and 10.7% for the Greens. Charts of respondent allocated preferences show that the Coalition’s share of respondent preferences fell continuously under Abbott, recovered to 2013 levels under Turnbull, and has now fallen a little below 2013 levels. If respondent allocated preferences were used in BludgerTrack, the Coalition would be reduced to 50.3% 2PP.
State BludgerTrack breakdowns show the Coalition has recovered ground in WA, as voters there focused on Federal issues and not on the unpopular conservative state government. However, Labor appears to be doing better in Tasmania.
In SA, the Nick Xenophon Team (NXT) was at 20% according to Newspoll. That level would probably be enough for one or two seats. Unfortunately, there has been no single state poll of SA, so we only have a rough idea of how well the NXT is doing.
In seat polling, Port Adelaide will be easily retained by Labor, despite a mooted challenge from the NXT. In the Labor-held marginal of Adelaide, Labor is ahead by 53-47 vs the Liberals. In the Victorian Labor-held marginal of Chisholm, Labor is ahead by just 51-49, but this probably reflects a bad sample of Greens voters; Labor would be leading 54-46 by last election preferences.
In Ipsos, Turnbull’s approval rating was 49% (up 2), and his disapproval rating was 41% (down 1), for a net approval of +8. Shorten’s net approval was -8, down 4 points. 61% thought the Coalition would win, with just 17% backing Labor; this was 54-26 to the Coalition last fortnight.
In ReachTEL, Turnbull’s (total good) minus (total poor) rating was -6, down 6 points. Shorten’s equivalent rating was -4, up 3 points. It is the first time Shorten has recorded a better ReachTEL rating than Turnbull.
In Newspoll, Turnbull’s satisfied rating was 40% (up 3) and his dissatisfied rating was 47% (down 4), for a net approval of -7; this is Turnbull’s best Newspoll rating since early March. Shorten’s net approval was an unchanged -15.
In Ipsos, Turnbull led by an unchanged 14 points as better PM. In Newspoll, Turnbull’s better PM lead was 17 points, up from 15. In ReachTEL’s forced choice better PM question, Turnbull’s lead shrank from 17 points to 6, as more non-Coalition voters backed Shorten.
Notes on these polls
In Essential, 38% thought the Liberals had run the best campaign, with 29% for Labor and 6% for the Greens; last week, this was respectively 30%, 28% and 8%. 48% thought the Coalition would win, with 21% backing Labor and 15% a hung Parliament; last week, this was respectively 39%, 24% and 16%. 26% thought the Brexit vote would be bad for Australia’s economy, 14% bad and 34% thought it would make no difference.
In ReachTEL, 55% expected a Coalition win, with 22% backing Labor. Last week, this was 50-22 to the Coalition. Shorten was thought to have campaigned better than Turnbull by a 5 point margin. 6% said they were more likely to vote for the Coalition as a result of Brexit, and only 3% said they were more likely to vote Labor.
Authors: Adrian Beaumont, PhD Student, School of Mathematics and Statistics, University of Melbourne