Australia's honest pollster (Image via roymorgan.com)

Why did Newspoll and Galaxy all wrongly predict a comfortable LNP win in last week's Queensland election, while Morgan Research had it spot on, saying it was too close to call. Bob Ellis explains.

TEN DAYS before the election, Gary Morgan said Labor was on 49.5, two party preferred and the outcome ‘too close to call’.

On the morning of the election, Galaxy, Newspoll and ReachTEL all had Labor on 48, and losing — despite Palaszczuk being, in two of them, preferred Premier.

I wrote at the time of the flawed polling method of Galaxy and Newspoll (ringing the landlines of those at home on a summer night) and the total failure of the ReachTEL method (which predicted Swan, Rudd, Clare, Burke, Bowen and Dreyfus would lose their seats) in previous elections. To this, I added the ‘as in 2012′ preferences of Katter and Palmer voters, figures no longer applicable, since they were being directed another way.

Gary Morgan weighed in at one point and said when you ring mobiles you get a one per cent better score for Labor — more young people vote Labor, or prefer it and so it goes. And when you ask people, not guess, whom they will preference, you also get, on top of this, a 0.5 or 0.75 per cent better vote for Labor.

This was why Morgan got the Federal 2013, and the South Australian, and the Victorian, and the Queensland elections dead right, and Newspoll and Galaxy got them, varyingly, wrong.

And this is why, once Labor is in, those responsible in Galaxy and Newspoll should be investigated for possible knowing fraud and ReachTEL abolished for incompetence.

Did Newspoll and Galaxy know that the Katter Party, 2012, and the Katter, 2015, preferences went different ways? Of course they did. Did they ask the Katter voters what their preferences would be? No.

Nor did they ask if they would ‘exhaust’. They used the 2012 figures, knowing they would falsely minimise the Labor vote.

Did they know many people do not have a landline? Yes. Did they ring one mobile? No. Did they text mobile numbers, as Morgan does? No.

So they knowingly, it might be suggested, cheated the election, by a deliberate fraud of reconfigured counting.

Is this a crime in Queensland? I suggest it is.

Should they be arrested? Of course they should.

Not that it matters, but I predicted on the morning of the election, in my blog ellistabletalk.com, that Labor would win 45 or 46 seats.

And I predicted in the afternoon they would win 42.

And I was, not for the first time (Victoria 1999, South Australia 2002, South Australia 2010, South Australia 2014), the only national commentator to get it right.

I did it by studying Morgan, and Morgan’s methods, and ‘deciphering’, like Alan Turing, the Enigma of Newspoll.

And so it went.

Monthly Donation


Single Donation


Join Newsletter

Please fill the text in this image in the field below to assist us in eliminating spam

Creative Commons Licence
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Australia License


Share this article:   


Join the conversation Comments Policy

comments powered by Disqus

Irenas Bookkeeping Services, Your XERO + MYOB Cloud Accounting Specialist