Nathan and I both filled out our same-sex marriage surveys and have posted them back to the Australian Bureau of Statistics (ABS).
As a scientist, I see so many design flaws with the process that, whichever way the decision goes, it's legitimacy is vulnerable to being questioned.
We all know that the compulsory plebiscite was blocked by the Senate and the Government has decided to go for a non compulsory postal survey instead.
Naturally, this means that only the views of a portion of the population will be sampled because not everyone is going to return their postal survey. Gathering data on only a 'subsample' of the population is widely and legitimately used in science – provided that the design of the subsampling is scientifically sound. Herein lies my biggest issue. Those who choose to participate and return their surveys will not necessarily be representative of the population as a whole because they will be a self-selecting bunch of individuals, i.e., those who feel strongly enough about the issue to vote. If one is going to subsample a population, then one needs to randomise the sampling process. The ABS should know this. It's one of the basics of good scientific design.
The question being asked is also being confused. The 'no' campaign has been raising a bunch of questions that have very little to do with the specific question at hand, which is about whether the law should be changed to allow same sex couples to marry. Thus, if some people are answering 'no' because they don't agree with some of the other issues being raised (boys wearing dresses; safe schools program), then this potentially also calls into question the validity of some of the returned surveys.
I've only raised some of the scientific issues. There are, in fact, many problems that could affect the credibility of the postal survey, from undelivered survey forms to people filling out multiple surveys.
And that's the worry!