RWAP: Alan Yan and Gabe Lenz
Conducting representative surveys has become increasingly hard and expensive. In this presentation, we report on an inexpensive approach: emailing registered voters. Almost 40% of registered voters in California listed an email address when registering to vote and the law allows us to contact them for surveys. With a small incentive, we achieved a response rates of about 1.35%. We compare our sample to various benchmarks, including an IGS Poll from earlier this year, which used a probability sample of the California population. Respondents to the email survey are disproportionately young and unrepresentative on age-related traits, such as ideology. When we weight on age, education, race, gender, and partisan identification, the poll comes reasonably close to the IGS Poll estimates of public opinion across a range of items (mean squared error of about five and half points). Additionally, we replicate two canonical experiments using our email sample. Although emailing so many individual seems problematic, about 70% of respondents who took the survey agreed to let us contact them for future surveys.