Making sense of the polls

Making sense of the polls

NSW ReachTEL has clear Coalition advantage

A New South Wales ReachTEL poll taken last Thursday night, from a sample of 1730, has the Coalition leading by 53-47 on respondent allocated preferences. Primary votes, with changes from last week, are 44.0% for the Coalition (down 0.6), 34.8% for Labor (down 0.2) and 10.2% for the Greens (up 0.2). The NSW election is on the 28 March.

While last week’s result was reported as 53-47 by 2011 preferences, I think it was actually 55-45, and that this week’s 9.2% primary vote gap would also produce 55-45 to the Coalition by 2011 preferences. So the effect of using respondent allocation is to improve Labor’s two party percentage by about 2%. A ReachTEL table of how minor party voters chose to allocate their preferences shows that Labor would gain 15 net votes per 100 minor party votes, up from just 3.5 at the 2011 election - this would be a greater gain rate for Labor than at either the successful 2003 or ‘07 elections.

The bottom line in this poll, however, is that a primary vote gap of over 9% in favour of the Coalition is too wide to be bridged by preference flows. In Queensland, this primary vote gap was only 3.8%.

Australian Polling on the Death Penalty for the Bali Duo

Polling on whether convicted drug traffickers, Andrew Chan and Myuran Sukumaran, should be executed has been somewhat contradictory. A Newspoll, conducted for the Lowy Institute on 13-15 February with a sample of 1210, found a strong majority against the executions (62-31). However, a Morgan SMS poll, taken from the 27 February to 1 March with a sample of 1280, found only a narrow 53-47 majority against the executions.

The difference between Morgan and Newspoll is probably because, whereas Newspoll first specifically asked whether Chan and Sukumaran should be executed, Morgan asked this question after a prior one that asked whether, in general, Australians convicted of drug trafficking overseas should have a death sentence carried out if sentenced to death. The order in which questions are asked can be important in a poll.

Other State Polling Notes

It is very rare to get a Northern Territory poll, but ReachTEL conducted one on the 1 March from a sample of 1040. Owing to the difficulty in polling remote communities, this poll was restricted to Darwin and Alice Springs, taking in 18 of the NT’s 25 electorates. The poll finds that support for the Country Liberal Party (CLP) has crashed since the 2012 election: the CLP’s primary is down 17% to 34%, with Labor up 6% to 42% and the Greens up 5% to 9%. ReachTEL’s respondent allocated preferences give Labor a 62-38 lead, an 18% swing to Labor across these 18 electorates. The Poll Bludger estimates a 12% swing using previous election preferences. Either way, the CLP first-term government is in trouble. The next NT election is scheduled for August 2016.

An EMRS Tasmanian poll has primary votes of Liberals 42% (steady since November), Labor 34% (up 3) and the Greens 15% (down 4); this is a 9% swing against the Liberals since the March 2014 election. If we take this poll at face value, then under Tasmania’s Hare Clark system, Kevin Bonham thinks the most likely seat outcome for the 25-member lower house would be Liberals 12, Labor 9, Greens 4. However, EMRS is known to lean to the Greens and against the majors, particularly Labor. With these leans factored in, Bonham thinks the most likely result is Liberals 13, Labor 10, Greens 2. This EMRS poll was taken on the 20-23 February with a sample of 1000.

Antony Green has revised his final two party estimate of the Queensland election result to a Labor win by 51.1-48.9, a 14.0% swing to Labor; the previous estimate was a 51.0-49.0 win to Labor.