Saturday, April 1, 2017

Postal Plebiscite: Australia's Biggest Bad Elector Survey

(Note added Aug 2017: This article from April has received much interest lately.  For new comments go to Electoral Process, But Not As We Know It: Postal Plebiscite V2).

The federal Coalition went to the 2016 federal election with a commitment to hold a national non-binding plebiscite on marriage equality (aka "same-sex marriage") prior to any further parliamentary vote on the issue.  The plebiscite was, as noted here before, a bad idea in policy terms, though it was mostly successful in neutralising marriage equality as a campaign issue.  The plebiscite plan was voted down in the Senate, leaving the whole issue apparently unable to progress within this term.

The option of simply changing the law seems impossible because religious reactionaries within and supporting the Coalition won't allow it.  (They're the ones who don't understand why people keep talking about marriage equality, but would bring down the Prime Minister and/or destroy their own party even in a failed attempt to stop it.) Leaving the issue as a festering distraction til the next election isn't too attractive either, so along comes Peter Dutton with a proposal to have the plebiscite anyway, but to do it by post.  Voting would be optional.

The idea of holding a voluntary ballot that does not need the approval of parliament is not new; this option of a "fee-for-service" ballot under Section 7A of the Electoral Act was discussed in the Senate plebiscite report. The option was not costed at the time because there was no proposed legislation to implement it.

Before I discuss the drawbacks of this option, I should first say if there are any advantages compared to the previous model.  None of the advantages provide a reason to actually do it, but I mention them all the same; it won't take long.

Firstly with the "voting" less focused on a particular date but spread over some time, and without polling booths to act as a focus for hostile campaigning and clashes, there's a strong chance a postal vote would be less intense in terms of the potential for hate-related incidents.  On the other hand, experience with all-postal elections (such as Tasmanian local government votes) suggests there could be an increase in letterboxing.  It would be surprising if the vote finished without at least some instances of letterboxing of gay-hate fliers.  In my home state these have sometimes occurred in recent years even in connection with elections at which same-sex marriage wasn't an obvious issue.

Secondly, it wouldn't be compulsory.  This cuts both ways.  While dragging everyone to the proverbial ballot booth would have been more representative, it would have meant some of those voting would have been there against their will.  It's bad enough that people are voting on minority rights when those rights are none of their business, and it's possibly worse when some of them don't even want to be voting.  What they would actually base their decision on in that circumstance is anyone's guess.

Thirdly, I expect it would be cheaper, but nowhere near as cheap as just changing the law.

The Bottom Of The Polling Pyramid

Bill Shorten has said "A postal vote, which is non-binding - it is an opinion poll".  This is in fact a serious insult to opinion polling.  A postal plebiscite would not be even that.

If we think of polls as a quality pyramid, at the top are the public voting intention polls put out by the most reliable companies, often as part of a long-running stable arrangement with a media source.  They're not absolutely scientific, because full method details are not published and hence can't be reproduced, but we know these things work pretty well because we can look at their predictive record.

Below them come the issue polls, which are conducted the same way but we have no way to test what the "right" wording is against actual results. Below those then come what I call the poll-shaped objects - things that look like polls and are done by pollsters but have skewed or suspect wording, disclosure or information deficits, are misinterpreted by those touting them or are otherwise suspect.

Down near the bottom of the pile are two forms of "polling" that are unscientific junk.  The first is the opt-in reader or viewer poll that many media outlets love.  The second is the "elector survey" in which an MP sends a letter out to his/her voters asking them to post back an answer to some questions.

"Elector surveys" are awful firstly because they only capture the views of those voters who choose to make effort to participate, and these voters aren't representative.  People who feel strongly about an issue, or have time on their hands, are more likely to reply.  These surveys are awful secondly because voters who share the political views of their MP will think that by responding they are helping their MP, while those who do not will more likely assume they are at best wasting their time.

What all levels of polling other than the worthless forms have in common - even the most mendaciously designed commissioned skew-polls have this - is the use of scaling.  The pollster knows that the sample of people who they can contact and who agree to be polled is not fully representative.  Twice as many women as men answered one Lonergan poll that actually published unscaled reference data (a great rarity in Australia).  So the pollsters ask demographic questions like age, gender, income and sometimes education level, and use these to scale their raw data to produce the best possible estimate of what the population on the whole would say if you could ask them.

As an indicator of what Australians generally believe, a postal plebiscite would be unscaled, and therefore much less reliable than many of the polls conducted on the issue.  An unscaled sample of, say, seven million is less reliable than a scaled poll with a sample of 1000.  So, unless there was a very high participation rate or a landslide win to one side or the other, a postal plebiscite wouldn't tell us whether Australians support same-sex marriage.  Not that we should need to be told, since polling - whoever the sponsor, whatever the question - has so persistently showed that they do.

Optional voting and commitment

The most likely way in which a postal plebiscite would be skewed would be if opponents were more likely to feel strongly enough to vote but many who supported marriage equality did not care enough to do so.

Some early polls supported this view.  For instance this Nielsen in 2010 found that 41% of supporters were strong supporters but 57% of opponents were strong opponents.  More recent commissioned polls show wide variation depending on who commissioned them - eg Crosby-Textor 2014 (67% vs 67%) but an under-documented Sexton poll commissioned by opponents supposedly has a 33-19 breakdown of strong vs "somewhat" support (58%) but a 20-3 breakdown of strong vs "somewhat" oppose (86%).  (This unpublished poll is being crowed about by Miranda Devine despite showing a 52-23 support-oppose response, but that's Miranda Devine for you.)  A neutral Newspoll in 2014 showed the share of a view that is strong at 76% of supporters vs 70% of opponents if you exclude those "leaning" one way or the other, but 57% vs 58% if you include "leaners".

So overall there doesn't seem to be solid polling evidence that supporters of marriage equality are much less committed than opponents.  Even if there was, in most recent polling, strong supporters outnumber all the opponents combined.  There is, however, a fear that virtually every opponent (including those weakly opposed) would vote in a plebiscite while weak supporters might mostly not bother and even some strong supporters might not get their act together.

Would the medium skew the message?

One key focus for opponents of the postal plebiscite has been the perception that the use of entirely postal voting would skew the outcome, even beyond the skew caused by voting being optional in the first place.

What do we know about this? Firstly, elderly voters more often use postal voting than younger voters in elections, because it is the most convenient voting form for those who lack physical mobility.  Secondly, optional all-postal votes in local government elections in my home state show a massive skew in participation towards elderly voters.  Only 30% of 20-24 year old voters vote while over 70% of over-65s do.

Voting in Tasmanian local government elections by age and gender 2014 Source: Tasmanian Electoral Commission annual report

However just because older voters are more likely to choose postal voting from a range of options in federal elections doesn't mean younger voters wouldn't use it if there was nothing else.  Also, most probably older voters would be more likely to vote than younger voters in Council elections even if a range of voting methods were available, since older voters are more aware of local government issues and more likely to be affected by them and to know the candidates.  What we really need is before-and-after samples for a change from postal voting to multi-mode voting, and vice versa, in optional voting environments.

In the USA, Oregon has predominantly postal voting and has high turnout rates as a result but the evidence does not suggest (see table p. 10) that young voters are unlikely to vote compared to other states.  However this is not a before-and-after sample.  I'd welcome links to any material that examines this question more fully.  We also can't necessarily translate conclusions about voting in elections to conclusions about voting in a one-off single-issue exercise that isn't binding.  The non-binding nature may itself make some people think there's no point voting.

While there is a general view that younger voters don't relate to putting letters in the post, I'm not sure how predictive that would be for this specific vote.  Young voters, probably sick of anti-"millenial" stereotypes in the media, might even be especially keen to vote to show the older generations that they can actually put a letter in the postbox when they have to.  Social media heavily used by young voters could easily build expectation that other young voters will vote.  It is also among younger voters that this issue is most likely to be named as important from among a bunch of issues.  At elections this has little impact because the young voters so naming it tend to vote for the left anyway, but for a postal plebiscite it shouldn't be dismissed so lightly.

The area in which young voters would certainly be disadvantaged in an all-postal vote is that they are more likely to be so transient (in or out of the country) that they just don't get their voting papers at the mailed address in the first place.  A conservative parent who is deeply opposed to marriage equality might not be in any hurry to forward the envelope to their same-sex-attracted (or just generally left-wing) son or daughter.

The Lack Of Oversight

A serious oddness of any kind of fee-for-service plebiscite would be the lack of normal electoral enforcement protections.  For starters, there would be no enforceable requirement to authorise electoral material, and hence nothing to prevent the letterboxing of unauthorised gay-hate material (though in some states this would be a breach of state law).  There would be no offence of publishing false or misleading material about the mechanics of the vote (such as incorrect voting dates or wrongly telling people voting was compulsory).  There would be no legal process for challenging irregularities or errors and having an official correction declared (though that said, the result would not be binding anyway, but some politicians might base a conscience vote on who "won" or "lost" either nationally or in their electorate).  There would be no parliamentary oversight of the question wording to ensure neutrality, so the question might either not be neutral or might not be perceived (by either side) as neutral.

There would not even be any clear mechanism to prohibit vote-buying, though there would be protection in laws related to postal services against stealing votes to engage in ballot-stuffing.

Why Bother?

The reason canvassed for actually considering a postal plebiscite at all is that it would be considered to fulfill the government's election promise to give all Australians a say ("every single Australian will have a say").  However given the unlikeliness of an all-postal plebiscite effectively reaching all voters, or of the vote being adequately protected from what would usually be electoral breaches, it is doubtful whether a postal fee-for-service plebiscite really fulfills that promise.

The current law seems almost certainly doomed; one of these days Labor will win an election comfortably, or perhaps even narrowly, and that will be the end of it.  Even with its defects, there doesn't seem to be much prospect that a postal plebiscite will actually deliver a no vote, and if it does the turnout will probably be so small that the outcome won't be taken very seriously.  In my view what this heel-dragging is about isn't delaying the advent of same-sex marriage by perhaps a few years, but legitimising and rewarding certain views.

A national plebiscite - whatever the outcome - is effectively a statement by a government that an issue is genuinely difficult and that both sides of the issue are equally valid.  It tells those who are opposed to equal marriage that their concerns were so legitimate that it was acceptable to ask the whole nation to express a view.  It shows "the base" they have been well and truly listened to and given a platform from which to lecture society.  Moreover, the vote is bound to equip opponents with endless grievances so they can claim that they were robbed, and we can see this already with the complaints about the widespread public support for change by major Australian companies.

What message it sends to left-out voters who are flirting with One Nation is another question. These voters, alienated from the political system, resent that politicians are preoccupied with same-sex marriage instead of dealing with the issues that affect them, and some may even oppose the measure for this reason.  But does a plebiscite really send a message to these people that their concerns are being listened to, or does it have the opposite effect?  Or does it show that the conservative "elites" and the progressive "elites" are one and the same in being so focused on such an issue and not on the impacts of the current economy?

My own view of course: the objections to marriage equality should not be legitimised with a plebiscite of any kind, or in any other way at all.  They belong in the same bin with laws against interracial marriage, laws against women voting, laws against shopping on Sundays, apartheid-style laws and other such forms of unjust discrimination States have sought to impose down the years.  The arguments advanced against allowing same-sex marriage have been woeful, and the State has a duty to denounce and dismantle avoidable policy stupidity in all its forms, and not to give it the pat on the back that it is craving.

But I doubt Peter Dutton sees it that way.


PS: Note especially for those waiting for Tasmanian LegCo coverage: I have been severely distracted by a chess-political matter but will try to step up activity on this site over the next week or so.


  1. Yes we need an amendment to s 116 of the Constitution so that after "The Commonwealth shall not make any law for establishing any religion, or for imposing any religious observance" we add "or for implementing any religious taboo". Even as it stands, it's arguable that _the Commonwealth_ could not make a law for Sunday closing (Henry Higgins thought so in the Convention Debates) but it should be made clear that it applies to things like marriage. In fact, let's have the French system where civil marriage is the official one and a church ceremony is an optional extra that adds nothing more to the couple's legal status.

  2. Let's add religious memberships, affiliations etc to the Members Register of Interests

  3. The bigger the vote is, the less bad the survey will be, unless of course potential Yes voters skew it by getting in a huff and not voting (I'm looking at you Michael Kirby). Everyone should take one minute to vote and 10 minutes walking to the nearest letter-box _immediately_ after receiving it in the mail. (With a time-extension for country people of course.)


The comment system is unreliable. If you cannot submit comments you can email me a comment (via email link in profile) - email must be entitled: Comment for publication, followed by the name of the article you wish to comment on. Comments are accepted in full or not at all. Comments will be published under the name the email is sent from unless an alias is clearly requested and stated. If you submit a comment which is not accepted within a few days you can also email me and I will check if it has been received.