The Guardian reports the fortnightly Essential Research poll has followed Newspoll in recording the Labor lead narrowing from 52-48 to 51-49 – and also in doing so from primary votes that you would think more likely to convert to 52-48. Labor are actually up two points from an unusually weak result last time, from 35% to 37%, while the Coalition are up a single point to 39%. The explanation for Labor’s two-party decline must lie in the two-point drop for the Greens, from 11% to 9%, and the attendant weakening in their flow of preferences. One Nation are up a point to 6%; no response option has been added for the United Australia Party, and there is nothing to suggest their ascent in the combined “others” tally, which is down a point to 9%.
If preference flows from 2016 are applied to these crudely rounded numbers, Labor starts with its 37% primary vote and gets 7.4% from the Greens (82% of their total), 3.0% from One Nation (50%) and 4.4% from others (49%), plus a 0.1% boost to correct for preference leakage between the Liberals and the Nationals. Add all that together and Labor comes out on 51.9%. Since this is, to the best of my knowledge, more-or-less the formula Essential uses, the explanation must lie in rounding. Dial Labor back to 36.6% and the Greens to 8.6%, and boost the Coalition to 39.4%, and you get primary votes that round to the published totals, but which produce a Labor two-party result of 51.4%, rounding to 51-49. There can’t have been much in it though.
The poll also features Essential’s occasional measure of leadership ratings, but all we are given at this stage is preferred prime minister. Scott Morrison’s lead as preferred prime minister is 40-31, down from 44-31 when the question was last asked in early March. So here too the poll reflects Newspoll in finding leadership ratings headed the opposite way from the two-party headline.
We will have to wait until later today for the full report, but The Guardian report relates that 59% expect Labor to win compared with 41% for the Coalition (so presumably a forced response); that “voters have logged news stories about the Liberal party’s preference deal with the controversial businessman Clive Palmer’s United Australia party, and are noticing the debates about tax and healthcare”; that the top rated issues were health, national security and the economy; and that 19% reported taking no interest in the campaign, 29% a little, 33% some, and 20% a lot.
UPDATE: Full report here. The preferred prime minister is the only leadership ratings result – nothing on leaders’ approval and disapproval.
Further poll news:
• Roy Morgan, which either publishes or doesn’t publish its weekly face-to-face poll in irregular fashion, has released its results for a second successive week. Polling conducted over the weekend had Labor’s two-party preferred lead steady at 51-49, according to both respondent-allocated and previous election preference measures. Both major parties are up half a point on the primary vote, the Coalition to 39.5% and Labor to 36%, while the Greens are steady on 9.5% and One Nation (which doesn’t do well in this series at the best of times) down two to 2.5%. Also not doing well in this series is Clive Palmer’s United Australia Party, steady on 2%. The poll was conducted face-to-face on Saturday and Sunday from sample size unknown, but probably around 700.
• The Advertiser has a YouGov Galaxy poll of Sturt, the Adelaide seat being vacated by Christopher Pyne, which had the Liberals leading 53-47, compared with their post-redistribution margin of 5.4%. The primary votes were 42% for the new Liberal candidate, James Stevens (44.7% post-redistribution); 35% for Labor candidate Cressida O’Hanlon (23.1%); a striking 9% for the United Australia Party (triple what Palmer United managed in Sturt in 2013); and 6% for the Greens. The poll also gives Scott Morrison a 45-31 lead over Bill Shorten as preferred prime minister; finds 40% less likely to vote Liberal because of Malcolm Turnbull’s replacement by Scott Morrison, compared with 25% for more likely; and finds only 22% more likely to vote Labor because of its franking credits and capital gains tax policies, compared with “almost half” for less likely. The poll was conducted last Wednesday from a sample of 504.
• The weirdest poll story of the campaign so far turns out to be the revelation that a supposed ReachTEL poll of the Curtin electorate, provided by independent candidate Louise Stewart to The West Australian and run as a front page story on Saturday, was fabricated. The Liberals reacted to ReachTEL’s denial that any such poll had been conducted by calling on Stewart to withdraw from her campaign, but Stewart says she believes she is the victim of a trick by her opponents. However, a follow-up report in The West Australian relates that Stewart told the paper she had “committed two polls from ReachTEL/Ucomms before election day”, and is now refusing the provide the email she received either to the paper or to ReachTEL. ReachTEL principal James Stewart said Louise Stewart had told him the email had been “deleted somehow”, but Louise Stewart says this is “not true”. Alex Turnbull, the son of the former Prime Minister, who has loomed large in independent candidates’ efforts to unseat sitting Liberals (though not, so far, in Stewart’s), said he believed he had been impersonated as part of the ruse. Stewart tells Andrew Burrell of The Australian that Turnbull’s investigations linked the distribution of the fake poll to a source “close to a senior conservative WA Liberal MP’s office in Perth”.