Fairfax pollster panned by rivals

LEADING pollsters have lined up to condemn the overreach of Fairfax’s new polling boss, Jess Elgood, when analysing Ipsos’s poll results in Monday’s Fairfax newspapers.

Ms Elgood was quoted in The Sydney Morning Herald saying: “They have read the writing on the wall for Mr Abbott ... It possibly ­indicates that the voters have ­already moved on from Mr ­Abbott.”

The Ipsos poll found a three-percentage-point rise in the ­Coalition’s two-party vote such that it trailed the Labor Party 49 to 51 per cent.

The results did not fit the ­narrative of commentators that the Prime Minister’s poor ­performance was damaging the government’s standing.

Galaxy Research managing ­director David Briggs disputed Ms Elgood’s argument.

“The idea that the surge in ­government support is because voters are already factoring in ­Abbott’s potential departure doesn’t make intuitive sense,” he said.

Liberal Party pollster Mark Textor evoked a Monty Python theme, describing the Ipsos boss’s analysis as “desperately free from the ravages of quantitative ­evidence”.

Essential Media boss Peter Lewis agreed, saying it read “like a pre-emptive eulogy, and I am sorry but I don’t buy it”.

“I have never seen a ­pre-emptive eulogy as a valid analysis for a poll outcome. One of the problems is that if you retrofit a margin of error movement in a poll with a political narrative you can end up tying yourself in knots,” he said.

Newspoll chief M­artin O’Shannessy said “had the poll asked a prospective question about the Liberal Party under ­different leadership” Ms Elgood’s conclusion might have had more legitimacy.

“At the moment Abbott is their leader and you would have to associate the vote that way,” he said.

“Just as, in fairness, we have ­associated all the negative votes to him as well.”

Briggs agreed: “The choice that respondents in the poll had is ­between a Tony Abbott-led ­Coalition and a Bill Shorten-led Labor Party. I do not see how someone voting Liberal in the poll could have interpreted ‘I’ll vote Liberal on the basis that it would be led by someone other than Tony Abbott’.”

All of those contacted by The Australian agreed that the ­comments by Ms Elgood exceeded the remit of a professional pollster, moving instead into the sphere of political commentary. “Pollsters tend to be quite noncommittal when it comes to the interpretation of poll data”, Mr Briggs said.

“And the more experienced will leave the explanation for shifts in the data to more experienced political commentators.”

One of the pollsters contacted caustically added “from market researching frozen peas to examining polls, it never quite works”.

Fairfax recently parted ways with its long time pollsters AC Nielson, transferring its polling operations to the less well known Ipsos agency.

It may well be that the surge in support for the Coalition came about because voters have started to factor in a change in leadership. Who knows? But it is equally if not more likely that any number of other factors impacted on the shift.

A focus on national security, Shorten’s defence of David Hicks, Abbott’s attacks on the Human Rights Commission President, a backlash at the internal undermining of the Prime Minister by some of his colleagues, or just a general concern among voters with the opposition’s performance as the alternative government.

The increase in the Coalition’s primary and two-party vote may have simply been a consequence of margin of error movement. The three-point shift was within the stated margin of error for the Ipsos poll.

Dr Peter van Onselen is a professor at the University of Western Australia and examined quantitative polling techniques as part of his postgraduate studies.