The following is a guest post from Princeton University political scientist Victoria Shineman.
*****
As the 2012 US elections come to a close, there will inevitably be much discussion of varying voter turnout. Low voter turnout introduces concerns because voters are not always a representative sample of the population. The number of non-voters is often greater than the margin of victory between the top candidates, suggesting that increasing mobilization could change electoral outcomes.
There are a number of electoral policies that could increase voter turnout. In addition to reducing the cost of voting (e.g. adding early voting or same-day voter registration), another alternative would be to offset the cost of voting by adding incentives for participation. For example, a 2006 ballot referendum proposed to enter every Arizona voter into a state-wide voter turnout lottery, with one random winner selected to receive $1 million after each general election. The proposal created quite a stir in the media, but ultimately received only 33.4% approval, and did not pass.
Participation incentives could also consist of direct handouts given only to voters. In recent elections, these types of promotions have been proposed more often than one might think. Many private companies have initiated promotions to reward voters with free or discounted items or services, including free Starbucks coffee, free Ben and Jerry’s ice cream, free red, white, and blue donuts from Krispy Kreme, free chicken sandwiches from Chick-fil-A, a 15% discount from the French Connection, free tacos and hamburgers, free beer, and many others (see here and here). Alternative promotions have even offered voters free body piercings, free tattoo removal services, free sex toys, free medical marijuana, and the chance to win a free rifle or pistol. However, in most cases, companies have been required to alter these promotions in order to comply with state and federal election laws.
When is it Legal to Incentivize Participation?
Simeon Nichter clarifies an important distinction between vote buying (compensation for the content of one’s vote choice) and turnout buying (compensation for the act of participation alone). Vote buying is strictly illegal in the United States, although it does occur on occasion. Turnout buying is illegal in all federal elections, and is also forbidden in 48 states. However, it is legal to offer valuable incentives in exchange for participation in local and state-level contests in California and Alaska, as long as there is no federal contest on the ballot.
There is an ongoing debate as to what constitutes payments for voting. Are campaigns allowed to offer voters a ride to the polling station? Can a candidate offer to pay the postage on absentee ballots? Accusations of illegally using free food to promote voter turnout have recently been made against Harry Reid, Mitt Romney, and Obama supporters.
In the subset of elections where participation incentives are legal, candidates have seized upon the opportunity. A notorious example occurred when Oakland Mayor Elihu Harris offered free chicken dinners to citizens who could produce their ballot receipt. Harris was accused of running a racist promotion because his flyers were allegedly only sent to predominantly African-American communities. He lost the election to Audie Bock, and Bock became the first Green Party candidate to serve in the California Legislature. Another example occurred when Mike Gipson initiated a raffle for $250 Target gift cards, and offered raffle tickets to any citizens who brought their ballot receipt to his campaign office.
Should this be legal? California’s chief elections officer, Secretary of State Debra Bowen, said “it appears that it is not illegal under California law, though it probably should be.”
What are the Effects of Incentivized Participation?
Despite the growing interest in such promotions, there have been few comparative studies estimating the effects of incentivized participation (see previous exceptions here and here). What types of incentives would be required to substantially increase voter turnout? And what other effects would incentivized participation have on individuals, the mass population, and electoral outcomes?
I recently conducted a field experiment to try to answer these questions. I surveyed 349 people before and after the 2011 San Francisco Municipal Election. Half the people in the survey were randomly chosen to receive a financial incentive to vote, in the form of a prepaid $25 Visa card that would only be activated if the subject cast a ballot. The mobilization treatment increased voter turnout from 45% to over 80%.
Continue after the break to see a summary of additional findings.
The experiment integrated an intensive mobilization treatment into a panel survey completed by 349 subjects before and after the 2011 San Francisco Municipal Election. Half the subjects received the mobilization treatment, and half did not. In order to test the effects of mobilization in varying information environments, half the subjects in each group were also randomly selected to receive an additional information treatment, producing a 2×2 experimental treatment design.
2×2 Experimental Treatment Design
Subjects receiving the mobilization treatment were given a voter registration card, a package of information about all the different ways that citizens can cast ballots in San Francisco, and were sent two reminder e-mails about the upcoming election. The mobilization treatment also offered subjects a financial incentive to vote, in the form of a prepaid $25 Visa card. I told subjects that I would activate the Visa card after the election, but if the subject failed to cast a ballot for any reason, I would cancel the card, and take the money back.
Voter Turnout Effects
Actual voter turnout was validated through the official San Francisco voter history file. The turnout records from previous elections demonstrate that the subjects in each treatment group were comparable before the experimental intervention. Voter turnout records from the 2011 election suggest that the mobilization treatment increased participation by more than 35 percentage points: turnout increased from 44.6% in the baseline group to over 80% in both groups receiving the mobilization treatment.
Political Information Effects
Incentivizing participation clearly increased voter turnout. But at what cost?
A common criticism of incentivized participation argues that increasing participation through incentives would hurt the quality of electoral outcomes. Comparative studies have consistently found that non-voters tend to be less politically informed than voters. Motivating uninformed citizens to cast random votes wouldn’t add representation – it would add noise. Therefore, many have argued, increasing voter turnout through a lottery or other material incentives would make electoral outcomes worse.
However, this critique is misguided for two reasons. First of all, most democratic elections use secret ballots. If an uninformed citizen was compelled to cast a ballot in order to receive some incentive, the citizen could still cast a blank or invalid vote, instead of voting at random.
Moreover, the full turnout critique relies on the assumption that information levels are fixed, and that an uninformed non-voter would remain uninformed if he or she were mobilized to vote. But information is not fixed. It is dynamic. People decide whether to invest in becoming informed about politics within the context of whether or not they expect to participate in the election. If someone knows they aren’t going to vote, they are less likely to pay attention to the campaign and the issues. However, if that same person were mobilized to cast a ballot in the election, they would be more likely to seek out information about the election.
In previous projects, I found support for this argument through a laboratory experiment and an observational comparison. The results from the field experiment in San Francisco also support my hypotheses. Subjects who received the mobilization treatment became more politically informed over the course of the election, in comparison to subjects in the baseline group. The graphs below depict the average information scores recorded in the post-election survey, comparing the baseline group to the group who received the mobilization treatment.
Subjects who received the mobilization treatment were better able to identify the ideological positions of the candidates competing in the three electoral contests, were more likely to watch the debates between the candidates, were better able to describe how ranked-choice voting (the voting system used in San Francisco) works, and were more likely to express preferences on the eight ballot referenda. Political engagement beyond the electoral campaign was unchanged.
Providing subjects with a financial incentive to participate not only caused those subjects to be more likely to cast a ballot – mobilized subjects were also more likely to acquire the types of information that were necessary for making good vote choices. Incentivizing participation increased both voter turnout and political information.
The results suggest that promotions like free coffee and donuts for voters would also cause people to become more informed about politics! Similarly, electoral policies that reduce the cost of voting should also lead to greater incentives to invest in political information. The inverse would also be true: electoral reforms which make voting more costly would decrease incentives for people to invest in political information.