Take a guided tour of what the polls tell us about the same-sex marriage survey — then take the reins yourself to explore how the result might pan out.
Let us know what you think the result will be in the comments.
This chart models possible results of the same-sex marriage survey based on two key factors: turnout and level of support.
1. Turnout: If a higher proportion of voters fill out their surveys and send them back on time, the marker moves higher on the chart.
2. Level of support: If more voters choose 'yes' to support same-sex marriage, the marker appears further to the right.
The ABS estimates it has received 9.2 million survey forms, or 57.5 per cent.
But we don't know anything about who has voted, or how they voted.
The most recent Essential poll finds 89 per cent of Australian voters say they have already voted or that they will 'definitely' or 'probably' vote in the survey.
And it finds 58 per cent of voters support changing the law to allow same-sex couples to marry.
But the story is not that simple.
Young and old people have very different views about same-sex marriage, so let's take a look at what happens to the model if we split voters into three separate age groups — each with its own marker.
The data seen here are based on an average of the last three Essential polls on same-sex marriage.
Around 10 per cent of people said they 'Don't know' whether the law should be changed, so we've assumed they won't vote and have taken them out of the totals and recalculated the proportion of 'Yes' and 'No' votes.
People aged between 18 and 34 are the most in favour of same-sex marriage, while older Australians are the least supportive.
This could affect the result — because there are more older voters than younger voters in Australia, and because some age groups are more inclined to participate in the postal survey.
At the moment, polls suggest older Australians are more likely than young voters to have their say in the postal survey.
Still, if turnout rates and the yes vote are exactly the same as the Essential polls, the overall national result would be 62 per cent for 'Yes'.
In this scenario, the over-55s would have returned 1.7 million more survey forms than the 18 to 34 age group.
Saying you're going to vote is one thing, but how many people will actually send back their survey forms?
What happens if only those who say they have already responded or will 'definitely' respond to the survey are considered?
The result is the same — 62 per cent in favour of same-sex marriage.
What happens if we bring those people who told pollsters they 'Don't know' what they think back into the equation?
Let's assume, as an extreme example, that the same-sex marriage campaign has swayed all of those people to not only return their surveys but to all vote 'No'. How would that change the result?
The push to change the law would still be supported, but with a reduced majority of 56 per cent 'Yes' votes.
Alright, we've seen what the result might look like if the polls are right. But, well, polls aren't getting everything right these days, are they?
What if the polls are wrong?
Here's an extreme example, where none of those aged 18 to 34 bother to send back their survey forms.
Using those in the other two age groups who said they would 'Definitely' or 'Probably' vote, the 'Yes' vote still comes out ahead at 59 per cent.
Keep in mind, we've returned to our original assumption here, that all the people who responded 'don't know' in the poll don't fill out and return a survey.
But what if the polls are wrong in a different way, and there are a whole group of silent 'no' voters out there who aren't showing up in the polls.
Dropping 'Yes' vote support by 13 percentage points below the poll numbers across all age groups sees the 'No' vote come out ahead, with 52 per cent against a change.
Again, in this scenario, the people who said they 'Don't know' have been excluded from the total.
Now it's your turn
Now we're handing the reins of the graphic over to you.
In this interactive version of the chart, the Don't Know contingent is once again included.
This allows you to choose the proportion of the 'Yes' vote out of all enrolled voters, not just those who have already made up their minds.
There are a lot of unknowns in all these examples — will turnout or the way people vote differ significantly from the polls?
We just don't know.
Try changing the voter turnout or support for same-sex marriage percentages to see what effect it has on the result.
Maybe you can even come up with an accurate prediction.
About this story
- Data on opinions about changing the law around same-sex marriage were taken from The Guardian Essential Research poll published on September 5, 19 and 26, 2017 and averaged across the three polls
- The group of voters who said they would 'definitely' or 'probably' vote on September 26 included the voters who said they had already voted
- The number of electors in each age group was taken from the Australian Electoral Commission data on the number of electors eligible to participate in the Marriage Law Survey, published on August 30, 2017
- Reporter: Cath Hanrahan
- Designer: Ben Spraggon
- Developer: Colin Gourlay
Editor: Matt Liddy