Remember the collective angst about how badly the opinion polls got it wrong in the 2019 election, with all of them predicting a 51-49 ALP victory?
Back then the (now) Nine newspapers The Sydney Morning Herald and The Age led the public handwringing with a vow to do things better. Gone was the Ipsos polling company which, the SMH/Age declared, “can’t walk away from the fact its overall polling forecast the wrong result”.
What’s more, as the SMH/Age declared straight after the election, polling companies were “the main reason Saturday night’s result took voters, the media and many political operatives by surprise. The implications of our major pollsters making the same mistakes in a consistent way are serious”.
So, how has the SMH/Age moved to repair the shattered trust?
Nine’s new pollster, unveiled this week, is Jim Reed, whose polling work for the last 15 years has been tied closely to conservative politics. Reed began with the Liberal-aligned Crosby Textor market research company back in 2006. By 2015 he had risen to the rank of group director of global research and strategy, as conservative parties surged in Australia and the UK (where Crosby Textor has worked closely with the Tory party).
Reed then joined Newgate Research, another Liberal-aligned company set up by former NSW Liberal government adviser Brian Tyson.
Last year as the COVID-19 pandemic hit, Reed’s new company, Resolve, was signed up to conduct more than half-a-million dollars of work for the COVID Commission. As Crikey reported, Reed’s company was awarded the lucrative contract without an open tender. Then in June he was handed another $502,000 limited tender contract by the Department of Treasury, again for market research for the government’s economic “comeback” campaign. That contract ended just months ago, on December 31, 2020. The ALP alleged that the job was no more than political messaging research for the Morrison government, done on the taxpayer dime.
Reed was also engaged to conduct research for the Australian Republican Movement, where former Liberal campaigner Sandy Biar is ensconced as national director.
So, how does this sit with Nine’s vow to restore public faith in polling? Particularly given the company’s top-level links to the Liberal Party through Nine chairman Peter Costello and senior editorial executive James Chessell, a one-time adviser to then-Howard government employment minister Joe Hockey?
SMH/Age national editor Tory Maguire told Crikey that it was irrelevant who else Reed had worked for.
“We hired Jim because we wanted to do good quality political research,” she said.
“Our only agenda is to find out what voters are thinking and get behind why they’re voting.”
Reed referred Crikey to the SMH/Age’s explanations covering the new methodology it was applying and the ways in which this would avoid the errors of 2019, as well as reasons for now omitting the two-party preferred poll.
On the question of his independence, Reed said in an email that he was not affiliated with any party. “Nor am I doing any work for any party, government, etc,” he added.
Abandoning the two-party preferred poll
As part of that larger look at how it does polling, Nine has done away with the traditional polling on two-party preferred voting — a measurement that, critics say, has produced a fixation on politics as a two-horse race.
Yet, according to veteran pollster John Utting, the two-party preferred vote is what political polling is ultimately all about.
“If you haven’t got the confidence to make that call, why are you in it?” Utting asked. “It’s what people are ultimately interested in.”
Utting, the managing director of Utting Research, worked as the ALP’s federal pollster for 20 years. He acknowledges that there was “a massive polling failure” in 2019 with respect to the two-party preferred vote that hasn’t yet been “truly explained”.
What of the 2019 polling review?
In the wake of the 2019 failure a review of pollsters’ performance was set up under Macquarie University emeritus Professor Murray Goot.
He told Crikey that Nine had a responsibility to be transparent about Reed’s background.
“I think it’s a good idea for papers to be upfront about this stuff. There may be a perceived conflict of interest,” he said.
He said that while the new survey went some way to address the problems of online polling — by getting samples offline — it did nothing to address the main problem of 2019, which was to do with misreading the primary vote, not the two-party preferred.
“What happened in 2019 was not a matter of how the sample was chosen. It was a problem with all methods of sampling. They over-represented the Labor vote. That may be because they over-represented people interested in politics,” he said.
“There are a lot of things about this poll that are new and interesting. [But] it doesn’t go to what happened in 2019.”