There are a lot of political campaigns every year.  And there are a lot of political scientists working in academia desperate for real-world experiments to publish results so they can stay in academia.  These two facts result in an enormous amount of theory-led, testing and experimenting in politics.

I’ve often wondered how much of this testing and the findings make their way into mainstream campaigns.  I don’t know if Youngkin, the newly elected Governor of Virginia, has a behavioral scientist or political scientist on staff or maybe a political consultant capable of reading and not afraid to use the skill but somebody got some science infused into their get-out-the-vote (GOTV) effort.

There have been dozens if not hundreds of GOTV tests done over the last 10 plus years.  By way of example,  one, large scale test with over 200,000 households was done in the Michigan 2006 primary.  They sent 3 different, GOTV mailers to randomly assigned households plus a control group that got nothing.   The least nudgy version underscored voting as the social norm with a line saying, “Do Your Civic Duty, Vote!”.  Thirty-two percent of these folks voted compared to the control group at 30%.   That is a pretty solid lift.

However, the most extreme test version had:

• Social norming – Do Your Civic Duty, Vote
• A layer of social pressure in the form of stated observation – “You are being studied”.  It noted researchers are studying voter turnout by examining public records.
• Another layer of social pressure showing the household’s voting record and a message saying they planned to send another postcard report after the election to note whether the person in the household had in fact voted
• The kicker that dials it up to 11 – showing neighbors’ voting status and promising to share an updated report after the election with all the neighbors seeing who did/did not vote.

This version jacked up voter participation by over 8 percentage points.  That is off the charts, massive.  Set aside, for the moment, the manipulative aspects of this and how it might backfire and how the effect may (likely will) dissipate quickly over time.  Politicians are nothing if not myopic and expedient.  8 pts is a crazy high bump.

Canvassing door to door can generate this kind of effect size/lift but at a much higher cost per vote.   Canvassing is about $20 a vote, this mailer was just under $2.

Here is the Youngkin GOTV flyer

The Michigan study did an advanced analysis that answered the main question I had with this tactic used by the Youngkin campaign: does the lift work mostly or only with those having lower, internal motivation to vote?  The researchers used past voting  history as a proxy for internal motivation and found folks who consistently voted the last 5 or more years (high, internal motivation) saw a bump in turnout that was commensurate with those having little or no prior vote history (low, internal motivation).

In other words this social pressure was additive for everybody.  Pressuring tactics often don’t work in the long-term and maybe these are one-trick pony tactics.  On the flip side, identifying as a Voter and seeing oneself as a “good” Voter is almost universally positive and it’s the overwhelming default as decades of survey research showing almost everyone identifies as a Voter even though many of those same people vote infrequently or not at all.

It’s possible that all this social pressure is pushing people into an Identity they already mentally own and want to keep owning.

Kevin