We tend to assume that we do politics differently in Australia.  Photos of dogs in polling booths, democracy sausages and primary school fetes get shared on social media, celebrating the novelty of Australian politics.  Certainly, as political scientists we assume that Australian voters are different from voters elsewhere, and we tend to attribute that to compulsory voting. 

By being made to vote, Australians, we assume, weigh up their options differently from voters who choose to turn out to vote, as in the US, UK or next door in New Zealand. Moreover, our parties behave differently: they don’t need to motivate us to vote (the prospect of a $20 fine does that well enough), so we don’t see any real campaign theatrics. Our politics, at least in terms of campaigning and elections, are kind of dull.

It is easy to rely on assumptions, though. And it is just as easy to overstate how ‘exceptional’ we are.  In a paper recently published in the Australian Journal of Political Science, Clive Bean, Ian McAllister and I test two competing models of vote choice to explain how Australian voters make up their minds. The first of those models is the most widely known, and the one we rely most heavily on in explaining voting behaviour: spatial politics. 

Spatial politics describes the concept of a political ‘left’ and ‘right’, and the way in which many voters think of ourselves as ‘closer’ to one side or another.  To put it another way, spatial politics is the way that we conceive of politics as ‘happening’ along a two-dimensional spectrum; we give politics and ideology shape and space. 

The spatial model – which we can also conceive in three dimensions, per the kinds of ‘ideological spectrum’ tests that proliferate online – was first described by Downs in the 1950s and really dominates how both researchers and citizens think about politics and voting. 

The other core component of spatial politics is party identification.  Party ID describes how we situate ourselves on the left-right spectrum with reference to the parties or candidates in a political system.  For instance, if a voter considers themselves fairly left-wing on the two-dimensional spectrum, and assesses the Australian Labor Party as also being left-wing on that spectrum, they are more likely than not to feel closer to the ALP than to the Liberal Party.  In other words, they are closer in ideological space to the ALP than to the Liberals.

Of course, a large number of us never think that hard about where we sit, or where the parties sit.  We take our relative positive on the spectrum as a given, handed down from our parents and other socialising influences, and adopt our party ID accordingly. 

None of this is ground-breaking stuff.  It underpins most political reporting and conventional wisdom about voting, elections and politics generally.

The second model that we tested is of more interest to us as researchers, as it is not as commonly discussed.  Valence issues are those things on which most voters agree: we want a thriving economy; border protection is important; health and education should be widely available and as inexpensive as possible.  Across left and right, these positions are almost universal.

Where voters tend to disagree is on the best ways to deliver these outcomes.  Two Australians can agree that economic management is the most important issue facing the country, but disagree vociferously on which party (or Prime Minister) is best able to manage the economy.

This framework for explaining vote choice has received considerably less attention than spatial frameworks, but research by Harold Clarke and colleagues has found that valence issues have stronger predictive power than spatial models on voters’ decisions in recent elections in the UK, Canada and the US.

We applied as similar a model to Clarke et al.’s as possible, given some variation in the available data and measures. Focusing on the 2013 Australian federal election, we analysed data from the 2013 Australian Election Study.  Importantly, we followed earlier studies by modelling valence politics (i.e. which party survey respondents believed was closer to their own views on whichever issue they identified as the most important to the country) alongside spatial politics (i.e. whether they feel closer to either the ALP or Liberal Party, and where they place themselves on an 11-point left-right scale).

We found that, in contrast to those similar studies from the UK, Canada and US, Australian voters rely more heavily on spatial politics – where they place themselves on the left-right spectrum and which party they identify with most closely – rather than on their beliefs about which party best reflects their own views on the issue most important to them.

At face value, this finding can seem counter-intuitive. It is appealing to believe that voters consider salient issues and assess the parties rationally, voting for the one that can best manage those issues. 

However, think back to what makes Australia different.  So many of us vote because we are made to (or rather, because we want to avoid a fine).  Compulsory voting is a burden on citizens who would otherwise not vote. And for these citizens, the most rational action is not to weigh up the issues and vote accordingly.  It makes more sense for those voters to draw on informational shortcuts wherever possible.  For many of us, the easiest shortcut is which party we identify with – the same party that our parents identified with, in most cases.