Opinion polls incorrectly called the Labor Party to win the 2019 federal election due to a "polling failure" caused by samples overrepresenting the party, an inquiry into polling performance has revealed.
Subscribe now for unlimited access.
$0/
(min cost $0)
or signup to continue reading
The Association of Market and Social Research Organisations launched the inquiry in June last year to determine what went wrong with polling predictions in the 2019 election.
The Coalition won the election with 51 per cent of the vote to Labor's 48 per cent, an outcome entirely opposite to what polls had predicted.
"Almost the mirror opposite of what the final polls found; all missing the result in the same direction and by a similar margin," the report said.
The report stated the first preference vote for Labor was overestimated because samples were "unrepresentative and inadequately adjusted".
"The polls were likely to have been skewed towards the more politically engaged and better educated voters with this bias not corrected," inquiry chair Darren Pennay said.
"As a result, the polls overrepresented Labor voters."
READ MORE:
It noted the lean had occurred in almost 70 per cent of final poll results since 2010 overestimating two-party preferred support for Labor.
The report found the incorrect result could not be attributed to 'shy conservatives' or respondents misleading pollsters and early voting.
"The allocation of preferences led to a slight increase in overall poll error in some estimates of the two-party-preferred vote, but was not a major contributing factor to the failure overall," it read.
Mr Pennay said participants had become increasingly disengaged with polling in recent years, with respondents less likely to answer the phone or participate if they did.
The report recommended pollsters look to new strategies and better understand aspects including the education of respondents to address bias.
Mr Pennay said probability sampling, achieved by cold-calling participants, was the "gold standard" for pollsters to get a broad sample.
"That's trying to contact people at random, rather than have them opt in," he said.
"Everyone has to have a chance of being selected in a sample, and you have to know that chance of being selected," he said.
It was also recommended polls report the number of respondents who were undecided or hadn't definitively made up their mind. It also called for a code of conduct to be established to provide regulatory oversight to the sector.
The role of the media was key, the inquiry found, with some reporting failing to meet disclosure standards set by the Press Council.
"We think there's a role for the pollsters and the industry more generally to provide more support and education resources to media ... reporting on the polls so they can put appropriate caveats on that," Mr Pennay said.
He said although the polls were a good guide to the election results, they were not an infallible source.
The inquiry sought data and cooperation from the four pollsters behind the national election poll results: Essential Research, Ipsos, Roy Morgan Research and YouGov.