What UK pollsters can teach CSSA members about the 2019 Federal Election result
By Brenton Prosser PhD, Director Research CSSA
Why didn’t we see it coming? How did the pollsters get it so wrong? What was behind the polls that consistently put Labor ahead over the last 12 months? How could political parties ever change their Prime Ministers because of polls again?
These are now but a few of the questions being asked in the wake of last Saturday’s unexpected Federal Election result. But the key message for CSSA members is that it is business as usual for the work of CSSA.
I vividly recall a sunny stroll through the stonewalled streets of upper Sheffield the day before the 2016 Brexit result. I was returning to the politics research impact centre where I was working with a couple of academic colleagues. Based on the polls, they both assumed a ‘stay’ result. I had a similar experience previously with the 2015 British General Election where the polls vastly underestimated the strength of the Conservative vote. Together these events led to a British Polling Council Inquiry (the BPCI) into what went wrong with the polls.
Over the long term, polls have shown themselves to be very reliable. But, some of the findings of the British Polling Inquiry, recent UK parliamentary inquiries and subsequent research (including into electoral change) might help us make sense of what happened here in Australia last Saturday.
One of the observations from the UK was the growing complexity of getting accurate results through polling. Historically polls were built around targeting demographics by calling people at home on landlines. But increasingly people are on mobiles or online and harder to catch. This has introduced new challenges around getting representative samples, delivering rigorous probabilities and including harder to reach groups (such as the marginalised and youth). The other important factor identified in the UK reviews is the increase in potential poll participants declining to participate. In short, our changing society is producing a larger invisible vote.
Associated with this is a new phenomenon from a dramatically expanding media and social media landscape. Polling on the popularity of leaders is a regular feature of news reporting and contributes to prominent narratives around the plight of parties. The UK inquiries found that one implication of this is that people do not report their voting intention accurately when it contradicts what they see as the dominant media view. To put this another way, media commentary around polls is producing a larger silent vote.
Another theme that emerged in the UK was around political convention and elites (including pollsters) losing touch with voters. Admittedly, the culture of politics is much more egalitarian in Australia than the UK, but recent research drawing on the British Electoral Study points to an important electoral change in the UK. Rather than traditional ideological categories shaping voting intention, people increasingly allocate their vote according to identity and local issues. In the UK, significant local variation occurred around cosmopolitan, provincial and marginalised (rather than class) identities. This shift was missed by major parties, but also presents a complex challenge for parties both to capture the votes of a fragmented electorate and to maintain a nationally consistent party narrative. Importantly, the UK inquiries suggest that while pollsters accurately capture views on different issues, they have not adequately adjusted for this shift in their political polling. In other words, when data comes in that contradicts what pollsters expect, it can be reweighted as an aberration rather than a revelation. Of course, it is too early to say if this has been a factor in the 2019 Federal Election.
Back in the Australian context, highly respected pollster Rebecca Huntley has highlighted a similar complexity in the relationship between polling, policy and politics. In her recent Quarterly Essay, Huntley contradicts one of the conventions of the Australian politics – namely that the major political parties are too driven by the polls. She points to decades of strong public support for changes in areas that the major parties have been afraid to reform (such as immigration, indigenous recognition and climate change). Huntley’s challenge to the major parties is to reflect these polling results and present a bold policy agenda to the Australian people.
One of the unconventional features of the 2019 Federal Election campaign was that an opposition party pre-released a detailed policy platform. Supporters of Labor will claim that this was a bold policy agenda. Commentators are already arguing that their loss will ensure ‘small target’ strategies in opposition will return as the new norm. Perhaps a fragmentation of the Australian electorate also means that major parties are struggling to balance contradictory national and marginal demands. But perhaps it may be, as Huntley suggests, that the major parties are not directing their boldness toward the sorts of issues that have wide public support and could win a first preference majority. While Australian political convention states you cannot be too bold if you want to be elected, could it be that the reason voters are disengaged is because the major parties are not bold enough?
Some of the issues raised here have currency for Catholic Social Services Australia in its role as a national advocate that informs public opinion and assists lawmakers. As a public voice that helps improve the lives of the poor and vulnerable in Australia, we speak across the political spectrum. But just as importantly, if the political spectrum is changing so that the public is more issues and identity orientated, we also need to adapt to speak in ways that connect.
That is why CSSA has made an important pivot in its work over the last twelve months. Under the leadership of our CEO, Fr Frank Brennan, and the CSSA Board, we have built the capacity for research and evidence to underpin all our work. What this means in practice is important.
Often government departments or sector peaks are accused of starting with a preferred view of a minister or with a set lobbying position only to seek out information that supports it. By contrast, a research-first approach is exactly that, one that starts by collating evidence about demands on members, collecting data on the needs of the disadvantaged and reviewing research on socially inclusive change. Also, the reach and diversity across our network perfectly places us to produce case studies and conduct place-based trials.
It is also common around election time for some peaks and lobbyists to endorse a particular party or take a partisan approach based on advocacy rather than evidence. The risk with this is they can alienate themselves in the future.
But if debate is evidence-based, then there is a common foundation for discussion, as well as freedom for genuine disagreement on what that evidence might mean for government action (i.e., policy). And when research and evidence is rigorously produced and translated with impact, there is the potential to rise above the cut-and-thrust of party politics and drive powerful public discourse.
This is a non-partisan approach that can work well across political parties and with public service leaders. It is also why, despite the ups and downs of election results, the work of CSSA does not change. Working with our members and by building a strong evidence base, we remain well placed to continue to help improve the lives of the poor and vulnerable.