The Case for Canvass -Analyst Group Digest: 06/08/2018
Canvassing is one of the tried and true tactics that most campaigns have employed at one point or another. Sending canvassers to knock on voters doors—whether volunteer, paid, or the candidate themselves—is almost synonymous with political campaigning. The progressive community has made great strides in improving the quality of canvasses in recent years. Read below for ways our community has taken advantage of canvass’ interactivity.
The Strengths of Canvassing
Canvassing conversations can have a large effect. At the dawn of the age of randomized experiments in politics, the first overarching hypothesis from Don Green and Alan Gerber was that face-to-face interaction was key to shifting voter behavior. Two decades later, we continue to find strong effects from in-person conversations. For instance, last cycle, canvasses across three states increased the vote margin for Democratic senate candidates by about 7-8 percentage points (short-term impact among voters who engaged in conversation).
In fact, likely the largest persuasion effects ever measured by experiments are two tests of candidates canvassing voters her/himself. The impact of these canvasses was substantial - they increased candidate support by between 20 and 40 percentage points among those contacted! Thus, local candidates, who can plausibly reach a large percent of the electorate by door knocking should strongly consider dedicating time to canvassing voters directly.
Get direct feedback on your message. The strength of canvassing likely comes from engaging in meaningful, responsive conversations with voters (possibly even making a personal connection). For instance, volunteers and organizers can ask a voter what their most important issue is and react in real time. Both the League of Conservation Voters and Working America—two of the premier progressive organizations with significant canvassing programs—begin canvass conversations by asking voters what is important to them, generating new data on what’s on voters’ minds.
This type of contact could also be used to understand voters’ priorities writ large. For example, Working America turns canvass conversations into “front porch focus groups” —better understanding the emotional state of the electorate.
Volunteer canvasses. While this digest series has focused on choices between paid modes of outreach, we should note that volunteer canvassing is a “no brainer” for political campaigns. Activists who are inspired by candidates will volunteer to help the campaign (for free!), and canvassing is an effective outlet for that passion. Remember that training is key to ensure high quality contacts. For persuasion canvasses, make sure voters know at least the candidate’s biography, values, and main issues. For GOTV canvasses, Analyst Institute’s evidence-based recommendations are tried and true training tools. And remember that asking volunteers to reach out to their friends in addition to traditional door knocking can help net votes!
A strong ground game is hard to blindside. If a Super PAC suddenly decides to attack your campaign or candidate, it’s easy to get blindsided by a wave of television, paid digital, or mail. It’s much more difficult to suddenly parachute in enough canvassers to make an impact. Planning ahead and building up a strong ground game before the opposition can provide a big advantage. Additionally, if a campaign finds itself with an abundance of volunteers or an IE has more resources than normal (both of which will hopefully happen often in 2018), there are usually more doors available for canvassing. The canvassing organization can set up additional field offices and target more doors in rural areas.
The effect could be more durable. Although previous research shows that the effects of canvassing usually decay over time—just as with pretty much all interventions and modes—longer, meaningful conversations carried out through canvassing might have longer-term impacts. In a 2016 experiment, door to door conversations had a durable effect in reducing transphobia—even several months later, shifts in voter attitudes towards transgender individuals endured. The advent of “deep” canvassing is promising, though testing shows that the effects of deep and standard canvassing are about equal. We welcome more research into what feature of the 2016 transgender canvasses made them so effective.
Can’t reach every house. Not everyone lives in housing that is easily accessible for canvassers; apartment buildings and rural areas could be particularly difficult and time consuming to reach. About 26% of Americans live in multi-unit homes, and although apartment buildings offer great density, they often have more access restrictions. Likewise, canvassing rural areas is also difficult since the distance between each door to knock is greater. Phone calls are a useful supplement in these cases.
Can't reach every voter. Even in a suburban district where it is possible to knock on every door, it’s quite unlikely that canvasses will reach each voter. First, not every voter will be home. Other voters might pretend they aren’t home, or, even if they open the door, refuse to engage with the canvasser. Finally, some canvassers might not finish their list for the shift. Due to these many possibilities, the average conversion rate for canvasses is often between 15% and 25% (for one pass).
Paid canvasses can be expensive. As progressives, we want all workers to earn a living wage with decent benefits, including paid canvassers. Thus, independent expenditures who engage in paid canvassing need to be prepared to pay over $10 per conversation with voters.
Quality control is hard. Unlike other forms of persuasion (e.g., television, digital, mail), the message is out of your hands; canvassers have the final say about what gets communicated to voters. Activist volunteers and professional organizers engaged in canvassing generally have stronger views than the average voter. Thus, they may misunderstand voters’ worldviews. A few canvassing studies have demonstrated that it is possible to lose votes through canvasser-voter interactions. Message research, training, and quality control are crucial for a successful canvass.
On the message side, voter backlash can be mitigated through in-cycle testing, aka “experiment informed programs” (EIPs). Recent innovations have made canvass effects easier to measure, with fewer doors needed to be knocked. Focusing on voters who are likely to be responsive—who are best found via persuasion models developed from real-world canvass tests—will maximize program impact.
Effects have more time to decay. Since we have yet to find a persuasion tactic whose effect persists in noisy electoral environment, our usual advice (while keeping in mind that every campaign’s situation is unique) is to save resources for the end of the campaign season. However, such stockpiling is not possible with canvass, as training and scale require a slower build up. Thus, the effect of the early canvasses has more time to decay. Learning how to mitigate this decay is one of Analyst Institute’s top research priorities, and some of our partners are actively exploring opportunities to improve.
Canvassing is one of the few times (if not the only time) during a campaign when voters will speak face-to-face with a candidate or representative of her campaign. Organizations should take advantage of this opportunity by both listening to voters and trying to earn their support. Campaign volunteerism is likely to reach new heights this year, providing a fantastic opportunity to engage with the electorate. Campaigns would be smart to think about how testing, training, and monitoring can maximize the electoral impact of this groundswell of enthusiasm.
A caveat we’ll add at the end of every digest in this series: it’s important to tailor your spending to your particular district, candidate, electorate and campaign. Different voters will be accessible through different modes and it’s important to strike a balance. We hope this series helps you fine-tune your strategy. And remember to test and create excellent content!
Special thanks to Jesse Ferguson for inspiring this series. We also appreciate Dave Fleischer, Matt Morrison, Clay Schroers, David Broockman, and Joshua Kalla for their comments. And these “Cases” couldn’t be made without progressive groups running experiments and sharing the results. Thank you!