Making sense of the polls

Making sense of the polls

NSW polls trend to Coalition; Netanyahu likely to win Israeli election

The New South Wales election will be held in ten days on the 28 March. A new NSW Morgan SMS poll has the Coalition ahead of Labor on primary votes by a crushing 13 points. The full primary votes are 46.5% for the Coalition (up 1.5 since mid-February), 33.5% for Labor (down 0.5) and 11.5% for the Greens (up 2). The stated Two Party Preferred (2PP) estimate is an unchanged 55.5-44.5 to the Coalition, but I think these primaries would be 56.5-43.5 by 2011 preferences. Even if we take 2% off that to allow for changes in preference flows, the Coalition still wins easily.

In Morgan’s poll release, he is comparing this poll to a phone poll taken in late February. As phone polling is different from SMS polling, I am comparing this SMS poll to the previous one. This SMS poll was taken from Friday to Sunday with a sample of 1290.

A NSW Galaxy poll also has the Coalition increasing its lead, though the shift is less dramatic than in Morgan. In Galaxy the Coalition has a 54-46 2PP lead, a gain of 1% for the Coalition since the last Galaxy in mid-February. Primary votes are 44% for the Coalition (up 1), 36% for Labor (steady) and 10% for the Greens (steady). 32% agreed that the upper house should pass electricity asset sales if the government is re-elected, with 48% disagreeing. This poll was conducted last Wednesday and Thursday from a sample of 820. The Galaxy poll used 2011 preference flows.

The Morgan SMS poll was volatile at the Victorian election, though the very high Greens votes recorded have not been a feature of his Queensland or NSW polling. This SMS poll should be taken with a grain of salt, but Galaxy has moved in the same direction, and it is clear that the Coalition has gained ground since the pre-campaign polling in late February.

I think the reason for the trend to the Coalition is that voters are focusing more on the state election, rather than on how they feel about the Federal government. This helps the state Coalition because Mike Baird is far more popular than Abbott. The recent Victorian election showed a similar trend, with a big Labor mid-campaign lead reduced to a narrow win on election day. As in NSW, Denis Napthine was more popular than Abbott, though to a lesser extent than Baird. In Queensland, both Abbott and Newman were very unpopular.

Kevin Bonham’s NSW poll aggregate gives the Coalition a 54.5-45.5 lead by 2011 preferences, reduced to 53.2-46.8 using respondent allocated preferences. His seat model gives the Coalition 52 of the 93 seats to Labor’s 37, assuming that the four cross-benchers hold their seats. The Poll Bludger’s BludgerTrack has the Coalition winning by 54.2-45.8 on 2011 preferences, and by 51.9-48.1 using a respondent-allocated preference estimate. On the latter measure, the Coalition leads the seat count by 51 to 39 for Labor.

Contrary to pre-election polls, Netanyahu’s Likud wins Israeli election easily

The Israeli election was held yesterday. With over 96% of the votes counted, the right wing Likud party of incumbent Prime Minister Benjamin Netanyahu has won 29 seats in the Knesset (Israeli Parliament). The left wing Zionist Union won only 24 seats. Other right wing parties to win seats were the nationalist Jewish Home (8 seats) and Israel Beitenu (6 seats) and the orthodox Shas (7 seats) and UTJ (6 seats). On the left, the Joint Arab List won 14 seats and Meretz won 5 seats. The two centrist parties are Yesh Atid with 11 seats and Kulana with 10. These figures may change slightly as the final votes from soldiers, prisoners and people in hospitals come in.

In total, the current results give the right wing parties 56 of the 120 seats to 43 for the left, with 21 for centrist parties; in 2013, right wing parties won 61 seats and left parties won 38. 61 seats are required to form government. Although it is theoretically possible for both centrist parties to side with the left, it is much more likely that Netanyahu will be re-elected as PM. Netanyahu’s position would be even stronger had a small right wing party not missed the 3.25% threshold.

Polling published four days before election day had the Zionist Union leading Likud by an average of a little more than three seats. The exit polls taken on election day showed a virtual tie between the two major parties. I think the most likely explanation for this polling error is a “shy Tory” effect; many people who voted for the hardline Likud party may have been reluctant to tell pollsters their true voting intentions.

Nationals hold Victorian Gippsland South seat in by-election

Following the resignation of Nationals leader Peter Ryan, a by-election was held last Saturday in his seat of Gippsland South. At the 2014 state election, the Nationals had won this seat by 15.7% vs Labor. Labor did not run in the by-election, which was contested by the Nationals, the Liberals and the Greens.

The Nationals won 45.2% of the primary vote, to 26.6% for the Liberals and 15.7% for the Greens. Even though Labor won 21.9% at the general election, the Greens vote only increased by 6.1%, so over 70% of Labor voters did not vote Greens at the by-election. Turnout was a high 86.9%, so it is clear that many Labor voters chose to vote for the Liberals or Nationals rather than the Greens. The electoral commission has not yet conducted a preference count between the Nationals and Liberals, but it is obvious that the Nationals will win.