Dec 6, 2010

Public Opinion on same sex marriage

With three separate polls having been released over the last couple of months gauging public opinion on same sex marriage, it’s worth taking a squiz at how opinion has changed over th

Possum Comitatus — Editor of Pollytics

Possum Comitatus

Editor of Pollytics

With three separate polls having been released over the last couple of months gauging public opinion on same sex marriage, it’s worth taking a squiz at how opinion has changed over the last few years, as well as breaking the results down to look at how there is a fairly wide variation on views among demographic cohorts.

First up, this is what the polling looks like going back to 2004 – the lines are the historical time series, the column charts are the polls from this year.


Over the longer term, we’ve see support for same sex marriage increase throughout the community – however, over the last two months, something interesting may have happened.

When it comes to issue polling in Australia, there’s a substantial proportion of the population whose mental autonomy on any given issue is fairly questionable, where they believe (or at least, say they believe) whatever their position is of the party they vote for.  If a party changes its policy from supporting some particular issue to opposing it, so too does a significant chunk of the population. We witnessed that with the CPRS debate, where we saw a decent level of support for it from Liberal Party voters under Turnbull, but then  a few weeks later under Abbott’s leadership, that support collapsed.

What we’ve seen over the last few months is the issue of same sex marriage getting a much higher profile in the media and politics than usual – so I wonder if what looks to be the recent, short term decline in support is simply a function of a number of politicians being loud in their public  opposition to the idea of same sex marriage?

We can slice and dice these polls in a number of ways to see what is happening underneath the top line numbers. Firstly, we can pool all three polls to look at support by party vote:


We have a somewhat expected result of the ALP vote being moderately for the issue, the Coalition vote being weakly against the issue and the Greens vote being very strong on the issue. If we pool the Nielsen and Galaxy polls, we can also get some breakdowns by geography and gender:


Queensland is the weakest state on the issue – which goes a long way to explaining the politics involved by both sides. Regional marginal Qld seats particularly are somewhat vulnerable over the issue as they have long been among the seats in Australia most opposed to the same sex marriage.

Note too the large gender split on the issue, with men being 13 points behind women on the level of support.

Finally, it’s worth having a closer look at the Galaxy results over 2009-10 and how they changed, as they provide us with all sorts of income, age and socioeconomic cross-tabs. I’ve added the approximate margins of error for each cohort in the table below, just to give an idea of the uncertainty behind any result.


What we see is that those folks less educated, older, childless, male and on lower incomes either not working or working in blue collar occupations are more likely to oppose, while those with higher education, younger, female and on higher incomes working in white collar jobs are more likely to support same sex marriage.

(Visited 81 times, 1 visits today)


Leave a comment

55 thoughts on “Public Opinion on same sex marriage

  1. Gay marriage conscience vote only first step « CDU Law and Business Online

    […] would be extremely unlikely to favour a liberal interpretation of the marriage power. With a fairly strong majority of Australians favouring gay marriage (although it’s a tad weaker among religious people) and one of our two major parties adopting […]

  2. Public opinion on same-sex marriage | unsane.info

    […] Over at Crikey in December, @pollytics wrote yet another one of his fascinating summaries of polling information – this time on the topic of same-sex marriage. […]

  3. BUTSeriously

    Same sex marraige is not a moral/ethical issue but an existential one. If 20% of the pop goes gay humanity would go pop! and disappear within a few generations. This makes it so problematic, in a sense more so than incest. Nothing whatsoever to do with bigotry.

  4. oedipus

    Same sex marriage seems to be the only way that marriage can be guaranteed to continue given the reluctance of opposite sex partners to honour the institution.

  5. Kevin Bonham

    [More straw clutching, I see, KB, ]

    If you’re seeing any , that would be the most likely example of it. I’m a bit jaded with the way you’ve felt obliged two posts in a row to make some kind of insulting cheapo wannabe-dismissive summary of my posts at the start, without making any remotely adequate attempt to connect that summary with any evidence that my posts taken as a whole deserve those sorts of insults. Please stop it and stick to debating the issues.

    [You ask if I’d be just as happy with a question that asked if Australia should legalise same sex marriage even though a vast majority of countries have not done so, on the grounds that that is equally factual, and quite simply: yes, I would.]

    You might be happy with it but I’d be expecting such a poll to show a lower % in favour of gay marriage than either the Galaxy poll with its preamble or the Nielsen poll without, and would be willing to bet at two or three to one that it would (the reason I wouldn’t make the odds even more against myself is random variation). We’ll probably never find out for sure though unless some homophobic organisation wants to commission one asking the question in that way.

    [I’m not interested in which question that gives the highest possible support, I’m interested in having questions that find out what people’s opinions are as accurately as possible. I think that questions that make explicit the facts of the issue they’re asking about are superior to questions that let a possibly uninformed respondent give their opinion on their possibly different interpretation of the issue – that’s just adding white noise for no reason, as I see it.]

    I find these two sentences (the first of which I completely agree with) incompatible with each other, largely for reasons I have already explained. People’s opinions on issues are the way they are partly because many people *are* uninformed and some people will interpret the issues in a way that influences their opinion of those issues, but that may not be an entirely correct way to interpret the issues. You can call these factors white noise but they are a contributing factor to the state of actual public opinion, which is what you (by sentence 1) are aiming to measure.

    I object to supplying information before asking a poll question even if the choice of information to be supplied really is completely neutral (ie cannot be seen as giving weight to a supporting argument for either side). The reason I object to this is that sometimes one side of a debate is more misinformed than the other. It’s a particularly important consideration on this issue because the anti-gay lobby is one of the more spectacularly misinformed movements that I’ve ever encountered.

    [You seem to think there are people that would answer “yes” the question “should we follow the Netherlands, Canada, etc and legalise same-sex marriage” who might answer “no” to a question without preamble, possibly because they think we’d be the first in the world to do so.]

    Or they may think it is only legal in minor European countries with reputations for radical and permissive social attitudes (like Sweden, Belgium, Iceland – all omitted from the list). Or they may just be unsure whether it is legal anywhere else or not. Or they may be waverers who are easily prodded one way or the other by the mention of an item of information. There are many possible reasons why the responses could differ in some cases. Even if each possible reason contributes only a point or so of difference, the points will add up.

    [ Given that the second scenario (us acting first) does not relate to a situation that actually exists in the real world, I’m totally uninterested in people’s opinion on it.]

    Again, this is inconsistent with what you said earlier. Firstly you say that you want to find out “what people’s opinions [on gay marriage] are as accurately as possible”. Now you say that if there are some people out there who are under the false impression that we would be the first in the world to legalise gay marriage, you’re just not interested in their opinion on it.

    [Put another way: if people’s idea of the worldwide state of same-sex marriage law is so far from reality that hearing an accurate description would change their opinion, then I’m more interested in their opinion that they’ve based on a statement of how things actually are in the world I live in, than the different opinion they previously held on what Australia should do in some parallel universe with different facts.]

    OK, you’re clear enough on that now that I’m taking that as your main position and dismissing “I’m interested in having questions that find out what people’s opinions are as accurately as possible.” as an aberration or miswording that doesn’t reflect your real views.

    Well, there’s nothing wrong with being interested in what their opinion would be in that hypothetical case (except that if you’re going to try to find that you should give them a comprehensive summary of both sides of any relevant facts, and maybe give them a while to think about it).

    But it’s interesting you talk about parallel universes, because the result you are looking for is one that tells you what opinions about gay marriage might be in a parallel universe where everybody was informed about the issue.

    If we’re going to be in that parallel universe, that’s fine, but more likely there will always be some portion of the voting public that is misinformed. The main purpose of neutral opinion polls is to examine actual sentiment out there, informed or otherwise, because that is the sentiment that can influence election and/or referendum outcomes.

    From that perspective, I’m far more interested in opinion in the real universe where some people have their heads in parallel universes, than in opinion in a parallel universe where nobody does that.

    Fiddling with how the question is asked can be useful, because it might tell you (as Silver argues) that if you frame the result a certain way opinions change, and that might give activists information on how they can make their case more effectively. That’s fine. But when the question is fiddled to the extent that it can no longer measure existing public opinion accurately, then it shouldn’t be passed off as X-%-of-Australians-believe-Y.

    Thus I reject Galaxy 2010 as a measure of public opinion in this universe. If people find it a useful measure of the state of opinion in some other universe, they’re welcome.

Leave a comment

Share this article with a friend

Just fill out the fields below and we'll send your friend a link to this article along with a message from you.

Your details

Your friend's details