Reflections on WA State Election 2017 : evidence-based campaigning

Chilla Bulbeck

Near the conclusion of the WA 2013 State Election campaign, I asked then Campaign Treasurer Jess McColl, if we could spend $10,000 on an exit poll. Several members of the Election Campaign Committee felt the money could be better spent in additional advertising. I said I would find the money to pay for the poll myself. Jess said, 'You really believe this is valuable, dont you? We found the money in the campaign budget (luckily!). That poll established, among other things, that one third of Greens voters decided to vote for us on election day, even up to the moment they entered the polling booth. From this we developed 'why to vote cards, including reasons for voting Green on our how to vote cards. This innovation was soon adopted by other parties.

The WA State 2017 campaign had five key goals, the first of which was 'Develop and implement an evidence-based strategic plan for conducting the 2017 State elections. The other goals covered an organising model for running elections, an outreach and visibility strategy, identifying and training campaign managers and candidates, and establishing and supporting the regional upper house campaign teams. 

In 2017 we spent 6% of our budget on surveying voters (where the industry generally suggests around 10% of the budget should go to research). I didnt have to beg for a single cent of this. We now routinely conduct focus groups to find out what persuadable voters care about. This research informs the development of our messaging. After the Federal 2014 Senate by-election we commissioned a survey of voters contacted by our doorknocking and phone calling campaign to assess the impact of field campaigning. Lonergan is commissioned after our election campaigns to conduct an online exit poll on our behalf, which informs the campaign plan for the next election.  

At no cost to the party, we benefit from the brilliant computing and statistical skills of Grahame Bowland and Norman Stomski, for example in creating our voter outcome maps and regression analyses to predict what the Greens voters should be in any locality (compared with what it is).

How useful, accurate or helpful is all of this research and evaluation?

Lesson 1: Talking to voters does work

Exit polls consistently find an inverse relationship between the reach of our message and the impact of our message. People most likely to be persuaded are voters who come to us: come to a meet the candidate event, visit our website or our Facebook page. Beyond the 'visitors, television ads reach the most people but have the least impact. At the other end of the scale, the most effective communication is one on one, either by doorknocking or telephoning. Surprisingly, our data suggests door knocking is no more effective than telephone calling.

Comparison of voter contact and the difference in the Greens vote between the 2013 and 2017 state elections shows a strong relationship in Maylands, where a massive voter contact led by Mary OByrne  achieved a swing of 4.86% for our candidate Caroline Perks, the largest swing among the East Metro lower house seats. In Scarborough, Jude Cullity and her team have been doorknocking every election since 2013 and achieved a swing of 3.37%. There was success in more localised field campaigns in places like Joondalup, South Perth, along the Midland Line and Dunsborough – to name a few.  

North Metro was the site of the largest swing in the State: 5.59% in Churchlands. The candidate, Jo Gurak, ran a strong field campaign but North Metros major field campaign was in Perth for Hannah Milligan, where the swing was only 1.59%.   

Lesson 2: External factors can obliterate the field campaign

The Perth campaign team was up against a very popular ALP candidate, John Carey - as they knew. The largest negative swing in the State was in the seat of Kimberley, where Broome is the major population centre. In 2013, the Greens almost won this as we were the only party opposed to the proposed James Price Point gas hub. In 2017, the Greens vote fell by 14.3%. Was there any way of minimising this decline?

Table 1 below shows that the North Metropolitan campaign was the most successful as measured by the final vote and the swing since 2013. Yet the largest field campaign was in East Metro and the percentage of contact attempts as a percentage of votes cast was second highest in Mining and Pastoral (data not shown in table). East Metro achieved the second highest positive swing and Mining and Pastoral the largest negative swing (because of Broome).

In fact, the most pervasive external predictor of the Greens vote in any electorate or region appears to be the extent of the One Nation vote. The lowest Pauline Hanson One Nation (PHON) quota is in North Metro where the Greens quota was the highest. There is almost an inverse relationship between the two quotas.

The PHON effect is due to both arithmetic and intervening variables. Arithmetically, the higher the PHON quota, all things being equal, the lower the Greens quota. Secondly, because Greens voters and One Nation voters are at opposite ends of many values scales, PHON was more likely to field lower house candidates where voters were more exposed to the economic downturn, i.e. not the leafy western suburbs (see box). Only 28% of North Metros lower house seats had a One Nation candidate, compared with around 60% in East Metro and South Metro, and between 88% and 100% in the non-metropolitan upper house regions.

One Nation supporters are overwhelmingly Australian born (98% in the 2016 election AES Survey, although interestingly 82% of Greens voters are Australian-born), identify as working class (66% compared with 24% of Greens) and have post-secondary but not university qualifications (57% compared with 33% of the Australian population, c.f. 20% have university qualifications compared with 42% of Australians). They are gloomy about their household and the nations financial prospects. Immigrants and refugees have destroyed the Australia that existed in the 1950s (83% want immigration numbers cut 'a lot compared with 7% of Greens voters; 90% agree with boat turn-backs compared with 10% of Greens voters). They are tough on crime (88% want the death penalty, compared with 15% of Greens voters). Greens voters are the most likely to have university qualifications, disproportionately live in inner city areas and our main abiding policy distinction with the ALP is our approach to asylum seekers. 

Door knockers and phone callers found fertile ground in North Metro with the message that PHON was racist. I spent some time with a voter in Scarborough who told me that Sharia law had been introduced into places in England and soon would be in Australia. I was about to say, 'thank you for your time and walk away when she said, 'But Im not a racist and Im not voting for Pauline Hanson. Normally a Liberal voter, she was open to the proposition that voting Greens, especially in the Upper House, would stop both One Nation and their preference bedfellows the Liberals.

It could also be argued that another minor party on the right robbed us of Lynn McLarens return in South Metro. The Liberal Democrats were positioned to the left of the Liberals on the upper house voting ballot and their vote of 2.5-6% in most booths was double that of around 0.5-1.5% in the other upper house regions – and sufficient to give them the seat on preference harvesting.

Lesson 3: Known unknowns

When people are asked what influences their vote, they usually identify party policies, party values or the candidates. Some voters vote for the Greens because they want a change or are sick of the major parties.  In this election, Greens voters identified as our key messages: the renewable energy economy; protecting precious places; tackling inequality – and saving the Great Barrier Reef!  Most voters got most of our messaging right, but they did not warm to integrity in government, while we did not campaign on saving the Reef. 

Even more questionable is voters self-reporting on the source of the message that influences them. The research suggests it takes five exposures to swing a voter. Exposure includes the full range of messages: seeing a billboard in a neighbours garden, talking to friends or family, receiving a friends Facebook share, or a conversation with a Greens volunteer handing out leaflets at the train station. Given this, the voters answer to the question, 'How much did this contact influence your vote? must be somewhat speculative. And voters generally 'know that only the gullible are influenced by advertising and so downgrade this message source.

Conclusion: a worthy goal – but is it realistic?

In 2013, Giz Watson, the unofficial parliamentary leader of the Greens WA, explained our loss of two seats as due the 'tsunami of the conservative vote. In 2017, we were hampered by the massive swing to Labor – delivering their largest ever majority in the Legislative Assembly. The Labor swing (and messaging) swamped Lynn MacLarens persistent support for the Beeliar Wetlands during and before the election campaign. At least Labor got onto the bandwagon and Lynns work paid off. Losing Lynn tempered our joy at welcoming Robin Chapple and Alison Xamon back (Alison in North Metro this time), to be joined by Tim Clifford and Diane Evers.

Because none of the candidates secured a quota in their own right, 136 scrutineers, led by the formidable Ozzie Coghlan and assisted by Zia Hakimi, delivered the goods secured by our hard-working preferences team convened by Luke Edmunds. Only once in recent years have we had a quota surplus to bequeath: in the extraordinary 2014 Senate by-election. 

A quota in every election should be our goal. This means more or less doubling our vote from around 8-9% to around 14-15%.  If we could achieve this, we can build on this with the aim of making 15% of the population mossed-on Greens: voters who always vote for a fair sustainable future for everyone. We can start working on pulling in additional 'churn voters, those who swing to us in one election and then away in the next. Before long, we will hold the balance of power in both State houses.

How do we do this?  Does this mean a community outreach program in every region, including a Greens presence at every country fair from now until the next election, a listening doorknocking and phone calling campaign in key metro electorates, a concerted program of engagement with community groups to explain the significance of preferential voting and the value of voting Greens? Does it mean a massive social media push through Facebook pages and websites across the State? How can our four new State politicians and our two senators best work with all our amazing volunteers co-ordinated by Sophie Greer, our new State Director, to ensure that 14% of Western Australians come to realise that we really are The Genuine Alternative?

Photo: Kalgoorlie doorknocking team February 2017. Chilla Bulbeck