Deep Canvassing to Reduce Abortion Stigma

During the fall of 2015, volunteers from the Los Angeles LGBT Center Leadership LAB "deep canvassed" conservative neighborhoods of Los Angeles in an attempt to decrease stigma towards women who have had abortions and increase support for safe and legal access to abortion.

We conducted an independent experimental evaluation to determine the effectiveness of the LAB’s conversations. The experiment had four steps:

1. We recruited voters to join an online survey panel, which we called the 2015 Los Angeles Opinion Study sponsored by the University of California, Berkeley. The voters were recruited via mail sent to the address at which they were registered to vote. This mail directed them to an online survey.
2. 1,982 voters completed the baseline survey. We randomly assigned these voters to receive the abortion canvass (treatment) or to a recycling conversation (placebo) to serve as a comparison group.
3. The LAB volunteers conducted the canvass, delivering either an abortion conversation or a placebo conversation depending on the voter’s random assignment. Voters do not realize the canvassing is connected to the survey in any way.
4. We invited voters who were successfully reached to take a follow-up survey one week later.
5. We then invited voters to participate in a second follow-up survey, five weeks after canvassing took place.

Based on the data from the one-week and five-week follow-up surveys, we found that the abortion canvassing had no detectable effects on the attitudes we measured in the survey. Despite finding some suggestive evidence that experienced canvassers may have had some success in the one-week survey, we do not find this this effect persists through five-weeks.

Following the successful results on using deep canvassing to reduce transphobia, these results were certainly disappointing. Unlike transgender discrimination, which at the time of our study was a relatively new issue to American politics, abortion is an issue that is deeply ingrained in our politics and often structures voters' "gut responses." Because abortion attitudes are so entwined with partisanship and religion, it is likely an incredibly hard issue on which to move voters. Nevertheless, we hope to conduct more research to discover if and how deep canvassing can meaningfully change attitudes around abortion.

To read more about this experiment, see our academic paper "The Design of Field Experiments With Survey Outcomes: A Framework for Selecting More Efficient, Robust, and Ethical Designs."