Poll-itics
Interpreting opinion polls in Australian democracy
by Jacob Deem
Australia is once again in the grip of poll mania after Malcolm Turnbull’s thirtieth consecutive negative Newspoll. While this poll has particular significance, given Turnbull’s use of the ’thirty-in-a-row’ threshold as part of his justification for challenging Tony Abbott’s leadership in 2015, the reality is that these barometers of public opinion are never far from the surface in discussions of Australian politics.
Apart from Turnbull’s use of the negative Newspolls as a justification for ousting Abbott, politicians generally claim not to place too much weight on the polls — naturally, when they receive an unfavourable poll, they argue the survey is flawed in some way, but most politicians are astute enough to avoid making a big deal about positive polls, lest it come around to bite them when the tide of public opinion changes. But whatever they may say, there is no denying that politicians take great note of the polls. Political parties wouldn’t conduct their own internal polling if they didn’t think it meant something. You would need to go back to John Howard in 2000-2001 to find a Prime Minister who wasn’t replaced when the polls soured.
Clearly, polls are an increasingly important element of Australian politics. The real question is how these polls should be interpreted and used to inform the public debate. To answer that question, one must first examine how they are currently used. The popular impulse, from both media and politicians, is to use polls as predictors of electoral success — they ask: ‘based on the latest poll, how likely is it that Party A will win the next election?’ This approach was clear when Turnbull outlined his reasons for challenging Abbott for the leadership in 2015:
The implication was that the trajectory of negative Newspolls indicated that the Liberals would lose the next election if Abbott were not replaced. It’s easy to see why they are used in this way. The polls to which Turnbull referred were based on the question: ‘if an election were to be held today, which party would you vote for?’ But while reputable opinion polls have a reasonable chance of predicting the results of an imminent election, they become less useful when election day is in the distant future. There are several reasons for this.
First, opinion polls are fundamentally designed to capture snapshots of contemporary views, not to predict future behaviour. In the same way that my preference for one type of breakfast cereal today does not guarantee I will still prefer that brand in a few months’ time (or longer), people can and do change their minds about politics and their leaders. While tracking trends in preferences over time does tell us something (you don’t lose thirty consecutive Newspolls by accident), history shows that it is not too late for Turnbull and the Coalition to turn things around and win the next election despite their current poor form.
Second, opinion polls can tell us which party someone says they will vote for, or who they would prefer as Prime Minister, but they don’t tell us why they have expressed that preference. Research suggests that a substantial proportion of the electorate are settled in their voting patterns, and will invariably vote for the same party at each election, but this does not guarantee that they will respond accordingly in a poll. Given the increasing influence of polls in the political process, respondents can use an opinion survey to send a message to the government, rather than answer in a way that accurately reflects their voting preferences. For example, a die-hard Liberal might tell a pollster they would vote Labor, even though they have no intention of doing so, to voice their displeasure about the latest policy proposal or parliamentary scandal.
This leads to the third problem — polls are no longer passive measures of political trends. Instead, they are active participants in the process. Scholars have long theorised and attempted to pinpoint the influence of polls on voters’ attitudes, highlighting bandwagon effects (where success in the polls leads to future successes because voters gravitate to ‘winners’).
Additionally, polls can lead to strategic and contagion voting, where citizens make decisions about who to vote for, based not on their preferred candidate, but on their perceptions of that candidate’s chance of success (which is itself informed by polling results).
Turnbull’s use of the thirty negative Newspolls adds a new dimension here — polls don’t just influence voters’ behaviour, they now apparently tell us when it is time for a new Prime Minister. Even if given the chance to stay on as PM, Turnbull’s words in 2015 opened the way for polling to take on an even more significant role in leadership debates in Australia. If replaced — which currently seems unlikely, but one should never rule anything out in Australian politics — the risk that the ’thirty negative Newspolls’ will become a rule of thumb or even a precedent for deciding when a Prime Minister’s time is up. If Turnbull is not replaced, the key challenge for him will be explaining why the standard he set for Abbott does not apply to him, and why future negative polls do not reinforce the idea that he has passed his expiration date. Either way, we have entered a dangerous new frontier in the way polling results influence the shape and priorities in Australian political debate.
It is important that the public and the media avoid buying into the use of polls, especially mid-term polls, as a basis for deciding who should lead this country. Instead, they should be interpreted and used to tell parties what they are doing right, and where there is room for improvement. Rather than seeing a mid-term poll as an indicator of future electoral success or failure, it is more helpful to view these polls as feedback on the government’s performance; they are interim report cards, not the final grade.
Certainly, the thirty consecutive negative Newspolls should be a wake-up call for the Coalition. Turnbull in particular has plenty of soul-searching to do, especially as it was he that set this absurd standard. Turnbull must convince both the Liberal Party and the voters that he is the right leader for Australia. Equally, the Liberal Party will want to look at what it can do to improve its reputation, using future polls to gauge the relative popularity of its economic and policy decisions. They should do so because results reflect broader discontent with the government’s current performance. If Turnbull is replaced as leader, it should be because the members of the Liberal Party believe there is someone who can do a better job as Prime Minister, not because they want to try and improve the polls.
It would be a dangerous step to grant opinion polls even more influence than they currently possess by continuing to use them as a basis to decide when and how we should debate changes in Prime Minister. Yes, politicians should take note of the polls, but they must view them as general indicators of current performance and measures of where improvements can be made, not the only criteria for success. Ultimately, voters respond best to policy, not poll-icy.
ABOUT THE AUTHOR
Jacob Deem
Jacob is a sessional academic in Griffith University’s School of Government and International Relations, teaching a range of undergraduate courses included ‘Australian Politics’ and ‘Political Leadership’. Jacob is a PhD candidate and Research Assistant in Griffith’s Centre for Governance and Public Policy. His research focuses on federalism and the principle of subsidiarity, focussing in particular on public attitudes and values in Australia and overseas.
Follow The Machinery of Government
