Is There Backlash to Social Pressure? A Large-scale Field Experiment on Voter Mobilization

This is a review of Is There Backlash to Social Pressure? A Large-scale Field Experiment on Voter Mobilization (2010) by Christopher B. Mann. Political Behavior 32 (September): 387-407. You can find the original in Google Scholar.

Mike Lee sent Republican voters their neighbors’ turnout records, hoping that Republicans would use this information to mobilize one another to vote. Instead, he was accused of unethically violating voter privacy

In 2008, Gerber et al. published a pioneering study of mobilization. Using heavy-handed tactics, they found that they could shame people into voting (read more). Using heavy-handed tactics might be fine for academics, but real-world campaigns can’t use those tactics without alienating their supporters.

For example, Utah’s Republican candidate for U.S. Senate, Mike Lee, sent out mobilization postcards in October 2010 that were apparently inspired by Gerber et al.’s heavy-handed tactics. Mike Lee sent Republican voters their neighbors’ turnout records, hoping that Republicans would use this information to mobilize one another to vote. Instead, he was accused of unethically violating voter privacy (see news coverage from a Utah paper).

Gerber et al.’s original study reported that tactics like these generated complaints, but when academics do an experiment these complaints don’t target a candidate. By contrast, Mike Lee’s partisan use of the same tactics generated blowback .

All this leads to Christopher Mann’s article in the current issue of Political Behavior, part of a special issue on social pressure and turnout (read some background). Mann replicates two of the heavy-handed tactics from Gerber et al.’s article. Then, in two additional treatments, he softens the tone of each mailer to see whether it will still have the same effect on turnout. So here’s the four resulting treatments:

  • “Self”: Like Gerber et al, send a postcard informing voters that turnout is a public record, and list that voter’s turnout history for the past few elections.
  • “Self” plus help: After listing the voter’s turnout history, invite the voter to correct any errors, then provide website and phone numbers for voter information.
  • “Hawthorne”: Inform voters that their turnout is being studied (to control for Hawthorne effects). Unlike Gerber et al, this mailer also included the information from the “self” mailer.
  • “Hawthorne” plus survey: Invite recipients to help the researchers understand turnout by filling out a brief survey about voter turnout.

The “help” and “survey” additions to Gerber et al.’s treatments were designed to reduce complaints and soften the “big brother” feel of the mailers. Perusing the sample mailers in Mann’s appendix (well, in the article’s appendix), the revised mailers do sound far gentler. And here’s what campaign consultants will love: The gentler mailers had the same strong effect on turnout. Here’s a summary:

Treatment Effect
(Gerber et al)
Effect
(Mann)
“Self” +4.9 +2.1
“Self” plus help NA +1.9
“Hawthorne” +2.5 +2.1
“Hawthorne” plus survey NA +2.2

Punchline: Gentler mailers can have the same effect as heavy-handed ones. There is no statistically meaningful difference between Mann’s harsh treatments and the gentler ones. Perhaps candidates like Mike Lee should soften their approach if they want to harness the power of social pressure.

Mann’s research is well done. I see no methodological concerns with it. He included 78,179 voters in his study, which looked at the 2007 Kentucky gubernatorial race. Mann’s effects are slightly smaller than Gerber et al.’s (roughly +2.0 for Mann vs +4.9 for Gerber et al.), but that probably reflects the different contexts that each study worked in; I discussed this issue yesterday while reviewing Panogopoulos’s work.

Possibly related posts:

One Trackback

  1. By Why did Mike Lee send those emails? | Utah Data Points on November 17, 2010 at 10:00 am

    [...] More recent research has tried to find other ways of exploiting this “peer pressure” effect without provoking negative blowback. If future candidates want to try something like what Lee did, they might want to read up on this more recent research. I have reviewed this more recent research elsewhere. [...]