Results from 2016 Illinois Text Message Experiment
During the 2016 general election, NextGen Climate conducted an experiment to assess whether peer-to-peer texting is effective at increasing voter turnout. NextGen Climate randomly assigned 94,257 individuals to receive a text message encouraging them to vote and 94,229 to a control group that received no outreach from NextGen Climate through Relay, the program they used for peer-to-peer texting. Individuals targeted were under 40 years old (average of 25) and had an average partisanship score of 81 (SE=15). The text messages were sent on November 5th, three days before Election Day. Overall, we find that 32.50% of the treatment group voted while 32.26% of the control group did. After adjusting for pre-treatment covariates, we find that the peer-to-peer text message increased turnout by 0.25 percentage points (p=0.20; SE=0.20).