Opinion, Berkeley Blogs

In initiative campaigns, can interest groups change minds?

By Joel Middleton

Ballot initiatives legislate on important issues including taxes, spending, law enforcement, education, health care and civil rights. Given the stakes, it is not surprising that vast sums are spent trying to pass or defeat initiatives — amounts that rival the spending on U.S. Presidential campaigns.

However, political scientists have been uncertain about whether all that spending amounts to much. Do campaigns for or against specific initiatives actually sway voters?

My collaborator, Todd Rogers at the Harvard Kennedy School, and I aimed to find out.  In our research, we evaluated the effectiveness of a nonprofit group that wanted to sway voters on 10 ballot initiatives in the 2008 election in Oregon (out of 12 total on the ballot).

polling place

The nonprofit Our Oregon sent persuasive ballot guides to nearly 90 percent of households in the state. To examine the impact, our study used a randomized controlled experiment, whereby some precincts were randomly assigned to a control group that did not receive the ballot guides, while other precincts were assigned to receive the guides.

Our results suggest that these guides can have surprisingly large effects. The ballot guides moved the vote margin by about 4 percentage points in the intended direction in each of 10 ballot measures. This suggests that, on each of the measures, perhaps 1 in 50 people switched their vote specifically because they received the treatment.

In fact, the ballot guides appear to have caused two initiatives to be rejected by voters. Measure 61 would have established mandatory minimum sentences for certain crimes. It lost by a 2.1 percentage point margin while the ballot guide moved the vote margin by over 5 percentage points. Measure 64 would have prohibited money collected with the use of public resources from being used for political purposes. It failed by a margin of 1 percentage point, while the ballot guide moved the vote margin by over 5 percentage points.

This study shows that not only can ballot-initiative campaigns influence an individual citizen's vote on a specific ballot initiative, but that these campaigns can actually affect whether a ballot initiative becomes law.