Labor failed to heed warnings that election was on knife edge, says secret YouGov report

Dr White says there was “a transparent agenda” to discredit some party officials in the wake of the party’s shock loss and warns that blaming the result on any individual, such as former leader Bill Shorten, is unfair and will not help the party win the next federal election.

Former Labor minister Craig Emerson (left) and former South Australian premier Jay Weatherill at the release last week of their highly critical review of Labor’s federal election campaign. Credit:AAP

The report, cited in the official review that came out last Thursday but not itself released, criticises the campaign for not acting on the research it was provided with and raises concerns about whether campaign slogans such as “the big end of town” were tested by research.

“The campaign missed an opportunity to convey the seriousness of the challenges it was facing,” Dr White writes.

“There were no face-to-face presentations of the research and most communications were with selected individuals who were not necessarily key decision makers.”

The nightly tracking showed Labor had 51 per cent of the two-party preferred vote when Prime Minister Scott Morrison called the election on April 11 but started losing ground the next day.

Labor regained a lead of 51 per cent at only one other point – May 4 – and was neck-and-neck against the Coalition in the final three internal party surveys before election day.

“The spin from whomever is responsible for these leaks [to the media] is apparently that, based on the internal polling, victory was assured. It is important to demonstrate that this is completely false,” Dr White writes.

“Our research clearly indicated challenge for the campaign throughout the track. On multiple occasions the numbers indicated that Labor was behind, and at no point did the numbers indicate a comfortable, let alone a certain, victory.”

He says he did not respond to the “provocations” of leaks to the media “despite the inferences made being defamatory” and had at “all times respected the confidentiality of the campaign”.

The submission shows the individual seat polling and the breakouts from the tracking clearly indicated that Labor would struggle to form a majority, with losses expected in Queensland, NSW and Tasmania.


The report says Labor did better in seats with more educated voters. In the inner-city Brisbane seat of Griffith, where 43 per cent of voters had a university degree, it gained a swing of 1.4 per cent. In the southern Brisbane seat of Forde, where just 14 per cent of voters were tertiary educated, there was an 8 per cent swing towards the Coalition.

“Clearly there’s a clear relationship between how educated the electorate is and the size of the swing,” Dr White says.

Opposition Leader Anthony Albanese said on Sunday that he had “raised concerns” internally before Labor’s defeat, including in a speech in June 2018 in which he argued for aspiration and against bashing big business.

“You could tell on the ground, when you listened to what people were saying, they were raising concerns,” he told the ABC.

Dr White acknowledges in his submission that all pollsters had the same systematic errors in their final numbers.

“No one got this election right; it was an industry-wide failure,” he says.

The submission shows the average error for each of the main pollsters when comparing their estimate of the Labor primary vote with the election result, including the YouGov ALP tracking done for the party and YouGov polling for Newspoll and for News Corp Australia tabloid newspapers.

It finds the primary vote average error was 1.6 percentage points for the YouGov ALP tracking, 1.7 points for Newspoll, 1.4 points for YouGov News Corp surveys and 1.3 points for the Ipsos surveys published in The Age and The Sydney Morning Herald.

Most Viewed in Politics



Related posts

Make a comment