Election FactCheck: could a vote among under 30s in Australia possibly deliver a Greens prime minister?

The Conversation is fact-checking claims made on Q&A, broadcast Mondays on the ABC at 9:35pm. Thank you to everyone who sent us quotes for checking via Twitter using hashtags #FactCheck and #QandA, on Facebook or by email.

If there was a vote amongst people who are under 30 in Australia, there'd possibly be a Greens prime minister. – Greens leader Richard Di Natale, speaking on Q&A, May 30, 2016.

Greens leader Richard Di Natale told Q&A that if there was a vote among people aged under 30 in Australia, there would possibly be a Greens prime minister.

Is he right?

Checking the poll data

Asked for a source to back up Di Natale's statement on Q&A, a spokeswoman said

Published Ipsos polling regularly shows our vote matching it with the other parties amongst young voters.

The spokeswoman pointed to two Fairfax Ipsos polls released over the past few months: one from April and another from February.

(You can view the Greens' spokeswoman's full response to The Conversation here.)

Those April and February poll results are shown below in tweets from poll-watcher Ghost Who Votes.

The April poll did show the Greens doing well among 18-24 year olds, scoring 32% of the vote in this age group. Labor, in this poll, had 33%.

However, these polls have a total sample of about 1,400, and the 18-24 subset is very small. In any case, Di Natale's claim was about those under 30, not under 25.

In the 25-39 year old range in the April poll cited by Di Natale's spokeswoman, the Greens vote is 17%.

However, subsequent poll data from both Fairfax-Ipsos and Newspoll (some of which was released before and some just after this episode of Q&A aired) indicates that Di Natale has exaggerated the level of support for the Greens among younger voters.

The May 17-19 Fairfax-Ipsos poll has a high Greens vote (14%) relative to other polls, but Labor is clearly in first place among young voters, with the Coalition second and the Greens a distant third.

Even among 18-24 year olds, the Greens have only 25% in the May 17-19 Fairfax-Ipsos poll (below), with Labor on 36% and the Coalition 32%.

April-May Newspoll breakdowns show the same thing; the Greens in Newspoll are at only 16% among 18-34 year olds, with Labor on 38% and the Coalition 33%.

A June poll by Fairfax Ipsos (released after Di Natale made the statement on Q&A) puts support for the Greens among 18-24 year olds at 27%.

For the Greens to be in an election-winning position among the under 30s, they would need to be ahead of Labor. But both Ipsos and Newspoll have Labor ahead of the Greens among young voters.

Using Newspoll data to calculate Greens support among under 30s

The Ipsos breakdowns are for only one poll, with a total sample of 1,500. The Newspoll breakdowns have a much larger total sample of over 6,800. Newspoll has the Greens at 16% for the 18–34 age group.

However, Di Natale's claim relates to those below 30 (that is, the 18–29 group). We cannot directly calculate the Greens percentage for 18–29 year olds, but we can assume a Greens percentage for 30–34 year olds, and calculate the 18–29 vote from that assumption.

There are 17 total years in the 18–34 range. I have assumed that any age is as likely to be interviewed as any other within that group. There are then 12 years in the 18–29 group, and 5 in the 30–34 group.

Let x be the Greens percent among 18–29 year olds, and y be the Greens percent among 30–34 year olds. We know that the overall figure must sum to 16%.

x is multiplied by (12/17), and y by (5/17) to get the correct weights of these percentages.

We have:

Rearranging to make x the subject gives:

In the highly unrealistic case that the Greens have zero support among those aged 30–34, their support among 18–29 year olds would still only be 23%.

A more realistic figure is that the Greens have 10% support among those aged 30–34. If that is used, they have 19% among those aged 18–29.

Verdict

Given the data above on the latest poll numbers, Richard Di Natale's claim that "if there was a vote amongst people who are under 30 in Australia, there'd possibly be a Greens prime minister" is exaggerated. – Adrian Beaumont

Editor's note to readers: The Conversation's standard FactCheck process is to ask an academic expert to test claims, and then ask a second academic to review an anonymous copy of the article. But this FactCheck involved both political and mathematical calculations. So in the interests of fairness and accuracy, we sought two blind reviews of this verdict: one from a political lecturer, the other from a mathematician.

Review

Senator Di Natale has a highly idiosyncratic reading of the polls, to say the least.

If going by the numbers supplied by the Greens spokesperson, the Greens sit on 32% behind Labor on 33% for the 18-24.

Senator Di Natale's must assume either (1) the Greens were in a 1% range of beating Labor in a first-past-the-post (or plurality) fight; or (2) that Liberals of this age group would tend to send their preferences to the Greens rather than to Labor. We don't have a first-past-the-post system (which Di Natale knows).

So presumably, he was thinking Liberal preferences would break his way. But many Liberals are very antagonistic to preferencing the Greens over Labor.

The Greens have denied existence of a preference deal with the Liberals and there's no hard evidence of a Liberal decision across the nation to preference the Greens over Labor.

ABC election analyst Antony Green has shown that Liberal preferences went 67% to Labor and only 33% to the Greens at the 2013 election.

Overall, my argument concurs with that of the fact checker. Di Natale's statement is unrealistic. – Mark Rolfe

I have reviewed the article and I find the author's conclusions to be reasonably supported by available evidence. The calculations assume equal voter population for each year of age. I have performed my own calculations using Australian Bureau of Statistics population data and this assumption seems reasonable.

Even if we do not assume equal population size for each year of age, the calculations change very little. I would also add the statement cannot be fully confirmed or refuted as there is no data solely for 18-29 year old voters, although this analysis suggests confirmation is unlikely. – Jake Olivier

Adrian Beaumont, PhD Student, School of Mathematics and Statistics, University of Melbourne

This article is republished from The Conversation under a Creative Commons license. Read the original article.