Strategic voting didn’t defeat Harper. Voter turnout did.
History rarely repeats itself, but it does rhyme.
Long before Justin Trudeau ever contemplated a life in politics, his father stoked the embers of Western alienation that would give rise to the Reform movement. Driven by a fiery hatred for the legacy of Pierre Elliott Trudeau, Reform MPs like Stephen Harper led the unified Conservative Party that governed Canada until Monday night.
Justin Trudeau turned the tables on Harper in the final weeks of the election campaign, but the groundwork for change was built meticulously over a period of years. Nowhere was this grassroots organizing more concentrated than British Columbia, where many voters carried the same bitterness toward Harper’s government once felt by Albertans toward Pierre Trudeau.
When the election finally arrived, turnout in B.C. surged from 60.4 to 70.4 per cent, outpacing Alberta, Ontario and Quebec. That’s what defeated the Conservative Party in B.C. – not “strategic voting”, but the appearance at polling stations of 471,397 citizens who were too young, not registered, or simply stayed home the last time. Together we elected the country’s most eclectic mix of MPs, including 17 Liberals, 14 New Democrats, 10 Conservatives and one Green.
Like in other parts of the country, thousands of British Columbians worked not just with parties, but with comedians, musical acts, trade unions, First Nations leaders, news outlets, community groups, campus associations, environmental, civil liberties and other issue-driven organizations – including Dogwood Initiative – to encourage people to register and cast a ballot.
At the same time, the Conservatives’ policy choices came back to haunt them.
This is the story of Dogwood Initiative’s role in that much broader push to get local people involved in the political decisions that affect them. First I want to outline the significance of the Conservative collapse in B.C., which has so far been overlooked by national media. Next we can delve into the “strategic voting” narrative that dominated the campaign and look at the role of opinion polling. Finally we need to talk about the art and science of getting out the vote – because that’s what truly defeated Stephen Harper.
The map looked very different four years ago. In 2011 the federal Conservative Party won 853,272 votes in British Columbia, claiming 21 out of 36 seats and finishing second in 13 more. With overall turnout low, the Conservative share of the popular vote was 46 per cent. If you transpose those votes to the new riding boundaries, the Conservatives would have won 28 seats. What happened?
On Monday the Conservatives actually did a remarkable job in most parts of the country getting their base out to vote. Outside of B.C. the party experienced a net loss of just 85,669 votes compared to 2011. But here in B.C. the collapse was dire: a net loss of 149,075 votes, or 64 per cent of the national total. In the 19 ridings where Dogwood Initiative worked in this election, Conservative candidates lost 82,257 votes – nearly as many as in the rest of the country put together. Swamped by the rising tide of voter turnout, the Conservative vote was diluted. They lost all but 10 seats out of 42.
Harper’s defeat was also fuelled by his own policy choices. Significant among those was his party’s determination to ram crude oil pipelines through B.C. without due process or local consent. In coastal ridings held by Tory incumbents, our polling found up to 44 per cent of people who had voted Conservative in 2011 disagreed with the government’s approach on the oil tanker issue.
By my count, half the ridings in the province border on salt water. In those 21 coastal seats the Conservative vote dropped by 96,524 votes – more than the net loss in the other nine provinces and three territories combined.
Many of these “Green Tories” have become Dogwood supporters, according to our internal surveys. That’s part of the reason we focus on shared values and government policies, not partisan attacks. In the election we knew we had to keep our strategy simple: give out information and get people to the polls, but trust them to make their own decision at the ballot box.
We surveyed candidates across the province, built maps of past election results, and conducted riding-level polls on issues and voting intentions, publishing everything online for anyone who pledged to vote. We held seven live candidate debates, put up lawn signs and billboards, ran targeted radio and web ads, canvassed door to door and signed up voters by mobile phone at concerts and public events.
In between organizing 475 local events, our teams even found time to make this fun movie trailer-style recruitment video, viewed more than 30,000 times on Facebook:
We sent emails to 179,683 B.C. supporters throughout the campaign. As advance polling dates approached, 854 of our volunteers hit the phones. We had 42,551 live calls with voters to talk them through the identification requirements. On election day and the evening before, we sent 35,288 text message reminders.
In short, we did everything we could to get British Columbians fired up to cast their ballot. What we didn’t do is tell anyone who they should vote for – even when they asked.
Prime Minister-designate Justin Trudeau vowed again on Tuesday that 2015 will be the last election to be fought under our single member plurality system, popularly known as “first-past-the-post”. The question of how to make the voting system more fair and representative will be the subject of study and debate over the next 18 months. Whatever the outcome, I’m sure many of you – as we do – hope this is the last time a campaign is so dominated by talk of “strategic voting”.
There are evidently many definitions of strategic voting, but at its core the exercise is based on blocking the candidate you don’t want by voting for another candidate who is not your first choice. To do so at scale requires telling large numbers of people how to vote. In addition to being contrary to Dogwood’s mandate as a nonpartisan, issue-focused group, research shows it simply doesn’t work, and often is counterproductive.
Nonetheless at the national level a majority of voters clearly wanted Harper out, and for many that desired outcome trumped loyalty to whichever party might be their first choice.
All the opposition parties seized on this meme of strategic voting, cherry-picking national or regional polls, algorithmic poll aggregators, past election results or riding-level polling of varying quality to make the case that in one place or another, they were the best-positioned to defeat a Conservative. The media amplified this narrative, and it soon became clear that at the ground level, many “strategic voters” were deeply confused.
In Victoria, voter Jean Oliver was quoted on the CBC talking about her decision to vote strategically for the Liberals, because Justin Trudeau was the best-positioned to defeat Stephen Harper. The problem was that her local Liberal candidate, Cheryl Thomas, had dropped out of the race due to controversial comments on social media. Thomas’s name remained on the ballot and she received 8,482 votes – more than the party received in 2011 or a by-election in 2012.
In Vancouver Granville, strategic voter David Laing told the Globe and Mail, “As a white male, I feel like I could certainly do a lot more to address the First Nations issues in our country. I’d love to vote for Jody,” referring to Liberal candidate Jody Wilson-Raybould. “However, having her as the representative for my riding doesn’t really mean much if Harper remains in power.”
The quote belies a misunderstanding shared by many Canadians that we elect a Prime Minister, rather than one out of 338 Members of Parliament. If Mr. Laing followed the recommendation of Leadnow’s Vote Together campaign in Vancouver Granville, then he would have voted for NDP candidate Mira Oreck – who lost. Wilson-Raybould won by 9,181 votes as the Liberals rode a wave from coast to coast.
Indeed more than a million votes migrated from the NDP to Liberals as momentum built behind Trudeau. 34 per cent of respondents on a post-election poll conducted online in B.C. by Insights West said they supported the candidate in their riding who had the best chance of defeating a party they disliked. But did they?
It’s difficult to quantify due to the practice in Canada of voting by secret ballot, and the inability of third-party campaigns to access the voters’ list or place scrutineers at polls. In the end, every person goes behind a cardboard screen and votes their conscience, based on a complex blend of criteria. That brings up another question: where are they getting their information?
Dogwood certainly has many supporters who expressed a willingness to switch their vote to defeat a pro-tanker Conservative candidate. We faced intense pressure at times to make local endorsements or put our thumb on the scale for one candidate or another. We refused, saying again and again that we trusted British Columbians to make up their own minds. As a compromise we decided to conduct local, riding-level polling in places where crowdfunding could cover the costs.
The goal was to offer data that we knew media outlets – especially in small markets – were unlikely to have. Indeed news reporters regularly confused algorithmic projection models with riding-level polling, despite disclaimers like this one on threehundredeight.com: “These riding projections are not polls and are not necessarily an accurate reflection of current voting intentions in each riding.” We wanted to offer some clarity amid the disinformation and partisan spin.
We engaged Insights West, a B.C.-based market research company, to conduct independent professional polling – not just once but three times in as many ridings as we could afford, to provide a sense of how voting intentions were shifting over time. These polls correctly predicted the winner in eight out of nine ridings we looked at, but failed to pick up the late Liberal surge in Burnaby North-Seymour.
That’s in large part because we wrapped up interviews on the final wave October 10, nine days before the election. There were two reasons for this. We chose to use live telephone interviews rather than a pre-recorded touchtone survey. It cost far more and resulted in smaller sample sizes, but we knew we were talking to real voters. It also took many nights to conduct each wave of interviews, making us less nimble in terms of getting the information analyzed and distributed quickly.
If we did this again, we would compress the schedule and move the final wave of polling later.
We also made a bad call in the final week, and that was to send an email with a graphic showing candidate support as a function of “decided voters” rather than highlighting the large number that were still undecided. As it turned out, those undecided voters were critical to the outcome in many places. Our website, VoteBC.ca, always displayed undecided voters as a cohort of their own, but that one email gave the impression voters were locking in their choices when in fact many waited until the final weekend.
On top of calling eight ridings correctly, we’re proud to note that Insights West predicted the province-wide popular vote to within a single percentage point for all four parties. However, that poll used an online panel and went right down to the wire on the Sunday before election night. It was not possible to offer the same methodology, sample size or speedy turnaround at the riding level.
The bottom line is that polls are snapshots in time. Multiple snapshots can point to trends. If people were determined to vote based on polls, we wanted them to have the best local information we could provide about how their neighbours were leaning ahead of election day. At no point did we suggest that anyone forego their own democratic choice or vote against their values.
Getting out the vote
Long before this election was called, we knew we wanted to work in ridings where pro-tanker Conservative candidates were offside with a majority of their constituents on our key campaign issue. We gave them a clear message: either get in line with B.C. values or take your chances with a highly motivated electorate.
Elections Canada’s transposition of the 2011 election results over to B.C.’s new 42-riding map allowed us to zero in on the places where we had teams working on the ground. We chose 19 ridings, including 10 that would have gone Conservative in the last election, and began planning the campaign.
Before the writ dropped I published an opinion piece in the National Post that served as a final warning: “Sooner or later, Enbridge will be defeated in B.C. The only question is whether the current government goes down with it.”
Monday night Dianne Watts held South Surrey-White Rock for the blue team by a tiny margin. In all other 18 ridings the Conservative vote share declined, voter turnout increased and the pro-tanker candidate was defeated.
In Courtenay-Alberni, veteran cabinet minister John Duncan saw his vote share slide by 23.9 per cent. Voter turnout jumped to 76.62 per cent, third-highest in the province, and Duncan lost his seat. Next door in North Island-Powell River the Conservative vote collapsed to the tune of 36.1 per cent, costing Duncan’s former advisor Laura Smith what should have been a safe riding.
In North Vancouver, Conservative MP Andrew Saxton lost 27.7 per cent of his 2011 vote. Voter turnout climbed to 76.4 per cent, fourth-highest in the province. Saxton lost his seat. So did John Weston next door in West Vancouver-Sunshine Coast-Sea-to-Sky Country, whose voted dropped by 27.9 per cent compared to the previous election. Turnout in that riding was 73.2 per cent.
Vancouver South was a bit different. MP Wai Young was infamous among our supporters for her private member’s bill threatening to jail pipeline protesters for up to 10 years. On election night she held on to her vote share, down just 463 votes compared to 2011. However, last time turnout in the riding was 55.25 per cent. This year participation climbed by 15.3 per cent and Young was defeated.
In Delta, Tory revenue minister Kerry-Lynne Findlay faced voter turnout of 75.11 per cent, losing her seat by 9,247 votes.
In six ridings where our organizers worked, the number of local Dogwood Initiative supporters exceeded the margin of victory. Nowhere was that ratio more dramatic than in Kootenay-Columbia. Dogwood organizers partnered with the West Kootenay EcoSociety, contacting 5,573 supporters to get out to the polls. The seat was won by Wayne Stetski of the NDP by just 285 votes. Voter turnout there was 73.79 per cent.
The most dramatic pro-tanker collapse we saw was in Esquimalt-Saanich-Sooke, where the Conservatives came within 406 votes of a win in 2011 (under slightly different boundaries). This time candidate Shari Lukens witnessed a 44.1 per cent crash in Conservative support – a loss of 9,393 votes. At the same time voter turnout spiked, hitting 75.94 per cent. Lukens finished a distant fourth place.
As I wrote in June, “Enbridge has become a liability for Conservative candidates in every B.C. riding touching salt water.” Today, aside from Dianne Watts in White Rock, Richmond MP Alice Wong’s is the only blue riding bordering the Pacific Ocean. Wong held on by 463 votes.
This would be easy to read as Dogwood claiming credit for taking down a bunch of Conservative MPs. That’s not the case. British Columbians decided they’d had enough of this government and voted them out. Our role consisted of identifying citizens that shared a common set of values, then offering tools and training to get their friends and neighbours out to vote.
We don’t encourage single-issue voting, and it’s not our belief that oil tankers were the ballot box question for all British Columbians. Rather, the federal government’s belligerence on our core campaign issue provided an excuse to organize. It also isolated the Conservatives, because the opposition parties saw the backlash in B.C. and positioned themselves accordingly.
In most places our numbers were dwarfed by dynamics totally out of our control – the national appetite for change, the rise of Justin Trudeau, the surge in voter turnout across the country. What I like to think is that along with all the other groups mentioned at the beginning, we were pushing in the right direction. Together we got a flywheel moving that drew millions of Canadians to cast a ballot for the first time.
Across the province, our organizers managed a wave of new volunteers who had never worked on a political campaign before, but were inspired by the chance to contribute, in a nonpartisan way, to historic change for Canada.
Together their dedication added up. When a team of three volunteers in Vancouver found themselves short two phones and two laptops, they shared one call station and phoned as a team, rather than giving up the opportunity to talk to their neighbours.
One volunteer called a man in his 60s who confessed he’d never voted in his life because he didn’t believe it made any difference. She talked him around and he agreed to give it a try. Another young volunteer from Surrey stayed on the phone as long as she had to, guiding a confused voter by satellite map to the right polling place in Prince George.
On the last day of advance polling, I remember the volunteer in Burnaby who got a woman to her polling place with 24 minutes to spare – right before she left on an international trip. She would have missed voting day.
Multiply that by 42,551 election phone calls, 179,683 email addresses and literally countless face-to-face conversations with British Columbians over the last four years.
Without access to the voter’s list it’s impossible to confirm the effects of specific voter turnout methods on individual supporters, as we were able to do in the 2014 municipal elections. What we do know is that face-to-face conversations are most motivating. Phone calls are second best. For most people email is third. We organize on the belief that every interaction is cumulative, and every conversation has ripple effects that carry out into the community.
In the ridings where Dogwood teams did not work, average turnout was 69.25 per cent and the decline in Conservative vote share was 12.87 per cent. In the ridings where Dogwood teams did work, average turnout was 71.64 per cent while the pro-tanker party saw its vote decline by an average of 22.9 per cent.
The Liberals won a majority by 14 seats this week. The party picked up 15 seats in B.C. If Justin Trudeau plans to govern for a second term, his team would be wise to study this unique electoral battleground.