News

Find contact details for any general enquiries.

Events

Find contact details for any general enquiries.

Our research

Our research is focused on issues that are highly significant for organisations, the Australian economy, and society at large.

Student resources

Whether you're a new or continuing student, you can find everything you need here about managing your program and the opportunities available to you.

Alumni

Our alumni may be found in the world’s leading companies, policy agencies and universities.

Contact us

Find contact details for any general enquiries.

Image: Shutterstock

Image: Shutterstock

Democracy, Statistics and Sausages

24 May 2019

6 minutes read

Martin Michael CMYK

This month’s Australian Federal election has shown two things: pre-election opinion polls are more fragile than we thought; and the democracy sausage could spell the end of effective exit polling, writes Professor Michael Martin.

On Election Day, after the polls closed at 6 pm, I read several news headlines that declared that exit polls pointed to a Labor victory. However, as the evening unfolded, the true situation was not only contrary to these headlines, but also to the 50 Newspolls (since 2016) that consistently predicted that the Coalition would lose the 2019 Federal election.    

I wondered, like thousands of Australians, how (and why) did the polls get it so, so wrong? And not for the first time, having witnessed the 2016 US presidential election and the Brexit vote both defying polling and punditry alike. 

So, I thought, what has brought us to this? Are polls just another example of “fake news”? And, how can it be “fake” when the theory underlying using sampling to predict population outcomes is both established and sound?

This thinking raises three questions. Has sampling changed? Are pollsters asking the right questions? And, are we interpreting the polls we read about correctly?

The battle of the polls: landlines vs mobile phones

For several decades, telephones or landlines have been fundamental to pre-election polls across the world. Under the premise that everyone had a telephone and that comprehensive directories were available. Samplers could randomly dial someone’s landline and hope to gain insight into “typical” Australian voters.

Additional tools, like stratification, could be used to fine-tune the sampling so that it could be regarded as representative – a true reflection of what would happen on Election Day.

However, we are now in the era of mobile phones – who uses a landline anymore? – for which there are no comprehensive directories like the White Pages. One of the major polls now has 46% of its sample based on mobile calls. Given our new reality, how does one gain a representative sample? And why does randomly dialling mobile numbers not necessarily gain access to “typical” Australian voters?

Here’s one possible reason. The makeup of mobile phone users varies considerably across different age groups. The younger Australians – many of whom are not able to vote – dominate as opposed to the older Australians who are less likely to be glued to their mobile screens. We’ve heard headlines that older Australians “defied the polls” through higher support for the Coalition. But were older Australians properly represented in polls based heavily on randomly dialling mobile phones?

It’s a coin toss, folks, and we ought not be all that surprised when the polls say “Heads” and it comes up “Tails”.

Making this question even more relevant is the fact that the dominant issues in this election were the “retiree tax” and “climate change” – big issues that disproportionately impact older Australians and the younger generation, respectively.

So, our supposed “climate change” election – based on polling that may not have captured enough older Australians – might have actually been a “retiree tax” election – based on who actually voted.

'What's the question?'

Another key factor in polling is the questions that pollsters ask once you answer your mobile.  While it seems straightforward to simply ask “Who will you vote for in the upcoming Federal election?”, our preferential voting system of electing a government is considerably more complex.

Especially in tighter races, pollsters really need to get a handle on how a prospective voter will rank not just their first, but second, third and even fourth choice candidate. Three-cornered contests, for instance between Labor-Liberal-Greens are even trickier.

What pollsters end up doing is to infer a two-party preferred outcome from the response to the simple question about the next Federal election using preference flows from the previous election. Needless to say, a lot can change in three years – even if the Government doesn’t. 

Democracy sausages will be the end of exit polls

And this all assumes that Australians being polled tell the truth. A prevailing attitude to ballot box choices that “it’s none of your business” can lead to people either refusing to participate in the poll or not necessarily telling the truth about what they will do on polling day.

Both factors corrode the accuracy of what polls can tell us. Even then, when the polls say 52% to 49% with a margin of error of 2.3%, both sides cover the middle. It’s a coin toss, folks, and we ought not be all that surprised when the polls say “Heads” and it comes up “Tails”.

As to exit polls, they also got it disastrously wrong. The reason why for this one came to me a lot more easily. Exit pollsters intercept people as they leave the polling area.

But these days, after voting, most of us make a beeline for the barbecue at the other end of the school yard. Woe betide the pollster who gets between the hungry voter and their democracy sausage.

Tell them anything, and head for the snag. No wonder pollsters got it wrong. I’ll have mine with onions – and don’t spare the sauce.

Read more about CBE students

‘The ANU scholarship has given confidence and a reassurance to my capabilities.’

Find out more »

As a strategy associate for the largest e-commerce operator in South-East Asia, Huiqi Low shares her insights

Find out more »

Westpac International’s CFO Rod Jackson reflects on his 40-year career

Find out more »

PolicyStreet Co-Founder Winnie Chua shares her thoughts and experience on the developing world of insurance

Find out more »

Updated:   24 May 2019 / Responsible Officer:  CBE Communications and Outreach / Page Contact:  College Web Team