The recent Australian Federal Election was a watershed moment for pollsters. In this exclusive EKAS research report, we reveal that the wholesale inability to accurately gauge the sentiment of voters across the country appears to have led to a loss in confidence amongst the public.

Introduction

With the majority of polling companies calling the Australian federal election wrongly in favour of Labor, we wanted to understand what impact this has had on people’s faith in pollsters to get it right, as well as their own habits when it comes to talking about politics.

To do this we ran an online panel, attracting a sample of 1000 Australians from across the political spectrum. Indeed, the self-reported voting habits of the respondents correlates almost perfectly with the declared results of the elections – suggesting the panel was highly representative of the Australian public.

This representative sample of voters suggests our political polling process is in a bad way.

When we asked the panel ‘Do you believe the results of political polls?’, more than half of respondents (54%) said they do not. Just 28% said they still believed them, whilst 17% said it depends on the circumstances.

The quiet Australians

Prime Minister Scott Morrison made some interesting comments on election night which were seemingly aimed at the pollsters, crediting the “quiet Australians who have won a great victory”.

So we wanted to test the theory of the ‘Shy Tory Factor’ – whether people who are intending to vote for conservative parties are less likely to admit it.

We thus asked the question ‘Would you feel uncomfortable admitting your voting preferences to a polling company?’.

The vast majority of respondents (68.2%) said they were happy to tell pollsters which way they are voting, whilst 16.2% said it would depend on the circumstances.

That left just 15.5% of people who were uncomfortable admitting who they were going to vote for.

Nearly half (45.7%) of those who admitted this identified as having voted for parties on the right of the political spectrum (Liberal, One Nation and Christian Democrat), with 38.1% more to the left of the spectrum (Labor and Greens).

Unsurprisingly a larger than average number identifying as being in this cohort also declined to tell us who they voted for in the last election.

Does social media hold the key?

One accusation levelled at the media was that it listened too closely to social media sentiment when it came to deciding what was likely to happen at the last federal election.

We wanted to sample how people behaved when it came to social media, to see whether this may have made them more susceptible to believing Labor were going to take victory.

Of the 1000 respondents 409 said they ‘never’ posted about politics on social media – suggesting a large number of people were silent about their political intentions during the election campaign.

Again the data shows those who voted for more right of centre parties were less likely to post about politics on social platforms, with half (50.1%) identifying as having voted that way in the federal election.

By contrast, 32% of those who said they never posted identified as voting for more leftist parties (Labor and the Greens).

This skew may be caused in part by age – traditionally right of centre voters skew older and therefore may not use social platforms as regularly, or feel less comfortable talking about political views on them. This survey did not ask for demographic data on age, location or gender.

Conversely those who identified as posting ‘very regularly’ about politics on social media (five times per month or more) skewed heavily towards more leftist parties – 49.4% of respondents compared to just 41% or right of centre respondents.

This cohort represented just 8% of respondents.

The number of respondents who said they would be uncomfortable admitting their political affiliations to a polling company who said they never posted on social media was slightly higher than the survey average at 17.6% compared to 15.5%.

What is the influence of polls?

Ultimately, what influence do polls have when it comes to affecting people’s voting habits?

We asked the panel to rate on a scale of 1 (not at all) to 10 (highly influential) ‘How much do political polls influence your decision making at the polling booth?’.

The vast majority of people, 58.8%, rated this question a 1, suggesting they do not put much stock in the polls. Just 6% of people rated it 8 or higher, suggesting they are highly influenced by what the polls are predicting ahead of election day.

Conclusion

While we can’t say categorically whether the polls prior to the last election had any impact on voting intentions, it is clear there is a lot of work to do for pollsters to gain a position of influence and authority in the minds of the Australian public.

The recently-announced review of practices may go some way to addressing this issue, but we would suggest there is a long road back and a robust overhaul of political polling practices needed to regain that public trust.

The fallout for pollsters in the wake of last month’s federal election has been nothing short of seismic. In a sign that tolerance for getting predictions wrong is likely to spell bad news in future, Nine’s newspapers, The Sydney Morning Herald and The Age, last month severed ties with long time polling partners Ipsos.

And it’s not just polling companies, but experienced social researchers who are also going back to the drawing board after the election results failed to match the predictions. Here, social researcher Dr Rebecca Huntley, writing in the Sydney Morning Herald, explains why it might be time to stop polling and to start listening.

Read more on this story

Welcome to EKAS’s inaugural newsletter. From this issue forward, we will bring you the latest news from our company and keep you up to date with the most interesting trends in research, marketing and surveying.

Each month we’ll feature bespoke polls, covering a wide range of industries, providing you with insights you won’t get anywhere else. Our first issue provides an in-depth look at how pollsters got voter sentiment so wrong in the lead up to last month’s federal election.

You can also get in touch with the EKAS team at any time by contacting Jaxon (jaxon@ekas.com.au) or Matt (matt@ekas.com.au). Our third managing partner Andrea (andrea@ekas.com.au) is contactable too, but may not reply as speedily as she’s currently on maternity leave.

We’d also like to take this opportunity to welcome our newest hire, Patty Rezzonico, who joined us this month.

Patty joins EKAS Marketing Research Services as the Qualitative Manager – managing and overseeing all project and operational aspects of the Qualitative Recruitment services.

Patty has more than 15 years of experience across a range of management and business disciplines, specialising in Healthcare and the Market Research Industry. She brings experience in all facets of field work, project management, and professional development of operations, streamlining sampling and recruitment increasing interviewing capacity and productivity.

With the advent of automated checkouts, companies have begun to shift their purchasing habits into the virtual space; reducing overheads, facilitating comparison shopping, and allowing consumer goods companies to collect vital information on purchasing habits and demographics, all at the click of a button. Both companies and consumers have begun the shift into eCommerce, but how can companies maximise the benefits of the virtual checkout?

To solve this problem, Ekas Marketing and Research Services have developed an exciting survey tool to directly observe the effects of the virtual checkout on consumer behaviour.
Introducing the Virtual Shelf.

With this innovative tool, companies can evaluate brand loyalty and explore consumer behaviour in a confidential laboratory setting, reducing the need for costly rollouts and the trial and error usually associated with implementing virtual checkouts into a company’s eCommerce repertoire.

Are consumers more willing to buy products on discount? What types of consumers are more willing to buy generic products? Given a budget, which brands would the consumer prefer to buy? Does the placement of items in a virtual checkout affect consumer behaviours? All these questions and more can be answered with the Virtual Shelf.