Disclosure and punishment. How social pressure works in lab elections

The Get out the Vote research project (Gerber, Green and Larimer 2008, Panagopoulos 2010, etc.) has documented the effects of social sanctions on turnout. But their experimental interventions (threatening people with the public disclosure of their electoral behaviour) are impossible to replicate outside the US (where voting records are public), are not actual social sanctions but a “threat” of a social sanction that comes from the authority (the researcher) instead of peers, ignore or take for granted important intervening variables such as the duty to vote, and make strong assumptions about the mediators of such interventions (i.e. shame). The present research intends to overcome these limitations by means of a laboratory experiment that put to empirical test the role of social sanctions in enforcing turnout. The design consists of a baseline election, which is used as a control group to benchmark voter turnout, and two treatments in a laboratory setting. High turnout is presented at a public good, with a common pot (distributed at the end of the experience among all participants) being created and enriched every time turnout in the group is at least 65%. But the presence of a voting cost means that individuals are also tempted to free ride and abstain. The 120 individuals (6 groups of 20; two control groups and two groups for each treatment, visibility and sanctions) are asked to take part in a series of 30 elections. In each election they are randomly assigned to party (A or B) and they have to choose between voting for their party or abstaining. While these are the parameters for the control group and the baseline elections (rounds 1 to 10) in all groups, treatment 1 makes public the behaviour of abstainers in rounds 11 to 30. Treatment 2 adds to this the possibility for voters, after each election, to sanction abstainers by taking points away from them. A diff-in-diff analysis show that both treatments have a positive effect on turnout, and that sanctions have the strongest and more lasting one. Further fixed-effects analyses yield interesting insights about the emotional and attitudinal mechanisms through which sanctions work.