ReachTEL moves to Labor, Essential to Coalition

This week's polls are confusing, with Essential now having the Coalition ahead, while ReachTEL had Labor gaining, particularly on respondent allocated preferences. ReachTEL is a far more credible poll than Essential, so it is more likely that ReachTEL is correct. We will need to wait until next week for other highly rated polls. Here is this week's poll table.

It appears that ReachTEL will be doing a weekly poll during the campaign. The poll will be taken every Thursday night and released on Friday night's 6pm Channel 7 news.

ReachTEL has been using respondent allocated preferences, which were 52-48 to Labor, a 2 point gain for Labor. Previous election preferences can be calculated from the decimal primaries, and were about 51-49 to Labor. I will continue to use previous election preferences in the tables, as all polls provide them.

Turnbull's (total good) minus (total poor) rating was -7, down one point. Polarisation has increased: Turnbull's "satisfactory" rating has dropped four points in three weeks. Shorten's rise continued, his (total good) minus (total poor) rating was -8, up three points on last week; this is Shorten's second best ReachTEL rating, and his highest since December 2013.

There was also a single seat ReachTEL poll of Corangamite, which was gained by the Liberals from Labor at the last election, and is held by a 3.9% margin. The poll has the Liberals easily retaining, 54-46. Single seat polls have been poor, but Corangamite would be expected to be difficult for Labor to regain.

Morgan had Labor leading by 51-49 on respondent allocated preferences, 0.5% better for the Coalition than the previous election method. Morgan has the total primary vote for the two major parties at only 70%, when all other polls have the total major party vote in the high 70's. This gives Morgan a credibility problem.

In Essential, Turnbull's net approval was +2, up four points from a fortnight ago, and Shorten's net approval was -10, down one point. Other polls have had Shorten improving and Turnbull steady, so these ratings could be the result of a pro-Coalition sample.

Essential and ReachTEL have roughly cancelled out in the aggregates, and there is only incremental improvement for Labor. Kevin Bonham has Labor at 50.1% Two Party Preferred (2PP), and the Poll Bludger's BludgerTrack has Labor at 50.2% 2PP. Both aggregates still have the Coalition winning a narrow majority of seats.

BludgerTrack primary votes are 40.8% for the Coalition, 34.8% for Labor, 10.6% for the Greens and 3.9% for the Nick Xenophon Team (NXT). Since last week, the Coalition's vote is down 0.8%, but Labor and the Greens are also down. The lost vote has gone to the NXT and Others.

Following Monday's Newspoll release of April to May state by state aggregate data, the Poll Bludger has published BludgerTrack data for each state. This has 3-4 point two party swings to Labor in NSW, Queensland and Victoria, a 9 point swing in WA, and little change in SA and Tasmania. In SA, the 23% primary vote for Others, mostly the NXT, could be distorting the two party swing. The big WA swing is partly explained by the unpopular state conservative government.

Notes on these polls

In ReachTEL, 29% thought the economy most important in deciding their vote, followed by jobs on 18%, health 17%, climate change 12%, education 11% and border protection 9%; these figures are little changed from last week. Labor has reduced the Coalition's economic lead from 11 points to 5 since last week, and also increased its lead on health from 23 points to 25, and education from 17 points to 21. The Coalition gained one point on border protection, leading that issue by 12 points.

In Essential, 77% thought that the election would be held in July. 50% correctly thought they would be voting for the House and the full Senate, with 8% for just the House, 8% for the House and half the Senate, and 34% who didn't know. 64% thought that Scott Morrison was the current Treasurer. 50% thought laws to prosecute whistleblowers should apply only to national security matters, while 30% approved of laws as they currently stand.

Additional polling from last week

In Newspoll, Shorten has gained ground on Turnbull on attributes and issues. Shorten now leads 47-40 on health, 47-41 on education and 41-36 on climate change after having been narrowly behind on all three in February. Turnbull maintains 20+ point leads on the economy, national security and asylum seekers. Shorten now leads on "cares for people" and "in touch with the voters", but Turnbull still leads on other positive Newspoll attributes. Turnbull has a 13-point lead on being perceived as arrogant. Unfortunately, tables for these questions have not been published online.

In Essential, Turnbull's attributes were worse since early May, particularly on being perceived as "out of touch" (up 7), arrogant (up 5), hard-working (down 5) and "understands the problems facing Australia" (down 5). Shorten's attributes were relatively unchanged. Turnbull now has a 21-point lead over Shorten on the negative attribute of being out of touch, and an 11-point lead on being arrogant, though the other five attributes with a difference of more than 10 points are all positive.

Both the party and leader trusted to handle particular issues are marred by a high undecided rate of at least 34% on all issues. 38% thought Labor, if elected, would change the Coalition's policy on asylum seekers, while 28% thought Labor would keep the Coalition's policy.

Why Labor probably needs at least 51% 2PP to win

At the last election, many Labor members lost their seats, and were replaced by Coalition members. Members usually have small personal votes that are not associated with their party, but with people in the electorate getting to know and like their local member. Of course some members become very unpopular, and their personal vote turns negative, for example Liberal Sophie Mirabella in Indi in 2013.

If a Labor sitting member was defending a marginal seat in 2013, and lost, the Coalition will benefit both from the loss of that Labor member's personal vote, and the new Coalition member's vote. Corangamite was one such seat, and the Liberal margin of 3.9% is underestimated, and is probably effectively about 6%.

When a new member wins a seat, they will usually receive a personal vote boost at the next election - this has become known as the "sophomore surge" effect. After that second election, personal votes are factored into the swing required to gain the seat. However, if a sitting member retires, their party will lose the benefit of that member's personal vote, leaving the seat more vulnerable than it appears.

On Antony Green's calculator, without factoring in retirements and sophomore members, Labor needs 50.4% 2PP to win more seats than the Coalition. With retirements and sophomore members factored in, Labor needs 50.6% 2PP to win more seats.

If Labor got 50.6% 2PP and won more seats than the Coalition they would be very lucky, because two of the seats Green's calculator has as Labor gains would be won by less than 0.1%, while all Coalition seats would be retained more comfortably. Kevin Bonham thinks Labor needs at least 50.9% 2PP.

Adrian Beaumont, PhD Student, School of Mathematics and Statistics, University of Melbourne

This article is republished from The Conversation under a Creative Commons license. Read the original article.