Opinion polling for the 2023 Australian Indigenous Voice referendum

Opinion polling on whether to change the Australian Constitution to establish an Indigenous Voice has been conducted since 2017, when Aboriginal and Torres Strait Islander leaders petitioned for such an amendment as part of the Uluru Statement from the Heart. The number of these polls conducted grew substantially following Labor's victory in the 2022 federal election; the party had committed to holding the referendum required for this constitutional change in its first term of government.

At least ten firms polled Australians on the proposed amendment, greater than the number who have polled party support for any previous Australian election. Some firms were commissioned by media organisations, think tanks, advocacy groups or university foundations. Other firms self-initiated their polls for market research or strategic communications purposes. Considering methodologies, polls were almost exclusively conducted online, with only one firm using SMS. Pollsters differed on whether to give a forced-choice question, as is done in actual Australian referendums, or allow respondents to express indecision or lack of knowledge. Some pollsters also used Likert-style questions to allow respondents to express how strong their opinion is.

The Australian Constitution requires a proposed amendment to attain a double majority in the referendum – not only a majority of votes nationwide, but also a majority in at least four of the six states. Because of this requirement, the level of support in each state was of special interest. One way pollsters investigated state-level support was to break down results from national polls. However, these polls sometimes did not survey enough people from each state to give reliable results about state-level support, especially for smaller states like Tasmania and South Australia. Another way pollsters investigated was by specifically surveying people from a particular state.

Pollsters also often broke down their results by age, gender, and party affiliation. The last was of particular interest because bipartisan support is often considered necessary for an Australian referendum to pass, though it is not a formal requirement.

The extent of support for the Voice among Aboriginal and Torres Strait Islander Australians was a key point of discussion. All publicly available polling indicated absolute majorities among this group favouring the Voice. Important caveats include the small sample size associated with certain polls, the length of time elapsed since the polls were conducted, and the lack of publicly available results and methodologies for certain polls.

National poll results

 * Notes

Aboriginal and Torres Strait Islander Australians
Two polls surveying Aboriginal and Torres Strait Islander people's views on the Indigenous Voice commissioned in early 2023 by The Uluru Dialogue, a pro-Voice lobby group, were conducted online by Ipsos and YouGov. Both found broad support for the Voice: 80% in the Ipsos survey and 83% in the YouGov survey. The YouGov result compared to 51% support in the broader population. These results have been cited by Yes campaign figures, including Prime Minister Anthony Albanese, as evidence for broad Indigenous support for the Voice.

The ABC conducted a review of Indigenous-specific polling in August 2023, consulting with polling experts Kevin Bonham, William Bowe, and Simon Jackman. All agreed there was nothing to suggest "anything other than broad support among First Nations Australians" and particularly endorsed the YouGov poll. Caveats included the small sample size associated with the Ipsos poll, the recency of the polls (both having been months out of date at the time of the review) and questions over whether online samples adequately represented remote Indigenous communities, although this was noted as a lesser concern given that the proportion of Indigenous people living in remote areas is often overestimated.

Late in the campaign, Resolve Strategic found 59% of Indigenous Australians in support, compared to 44% support in the broader population. This was a considerable decrease from the near-80% support in the Ipsos and YouGov polls, but still represented a majority. Another late poll from Focaldata found about 70% in support, but the firm urged caution in interpreting the figure because it was based on a sample of about 250 Indigenous Australians.

Passing the Message Stick, a research group aiming to discover strategies for a Yes victory in the referendum, found in a telephone survey 24% of Indigenous people opposed the Voice and 42% either did not know about the referendum or had heard very little. These results was initially presented in a webinar with a limited audience (400 people), but later more widely reported. The results have been cited by No campaigners to dispute the Yes campaign's perspective of broad Indigenous support. However, the ABC excluded this survey from consideration in their review of Indigenous-specific polling, because not enough information about its results and methodologies was publicly available.

The No campaign conducted three internal polls on Indigenous support from February to May 2023, though only two results were publicly released, showing a fall in support from 60% in February to 57% in May. These results compared with 59% (February) and 54% (May) support among the broader population. These polls were not publicly reported until early October. No details were provided regarding sample sizes, precise questions asked, or methodologies.

Weighted media audience surveys
Some media organisations have surveyed their audiences about their views, then weighted the results by various demographic factors in an attempt to make the results nationally representative.

Australian Reconciliation Barometer
Since 2018, Reconciliation Australia has included a question in its biennial poll, the Australian Reconciliation Barometer, on whether it is important to "protect a First Nations Body in the Constitution". Although this poll has drawn academic attention, there is some dispute over whether the data from this question is suitable for discovering public opinion on the Indigenous Voice. Francis Markham and William Sanders included the question in their analysis, taking respondents indicating importance to be expressing support for the Voice and respondents indicating unimportance, opposition. Murray Goot, terming the question "quite general", did not believe it specifically referred to the Voice.