The small effects of political advertising are small regardless of context, message, sender, or receiver: Evidence from 59 real-time randomized experiments
Decades of political communication research has pointed to political persuasion (or changing someone’s mind) being one of the most difficult things to achieve, with campaigns having small effects on average. This would suggest that attempts at political persuasion, such as political ads in an electoral campaign, mostly do not move the needle much (if at all) with audiences. However, one response has been that these small average effects may be masking larger individual effects from sources of heterogeneity, or contextual pieces of a political ad such as message (content, tone, sponsor), audience (partisanship), and setting (battleground status, primary vs. general election, timing).
Using an astonishing real-time design consisting of real ads from US presidential candidates, 59 unique experiments and 34,000 respondents over the course of the 2016 presidential election, the authors of this study find little evidence that any of the above conditions matter much when it comes to the persuasive effect of political ads on candidate favorability and vote choice. Indeed, owing to the large number of respondents (which makes finding even tiny effects more likely), the authors found a small but significant effect that ads can influence candidate favorability but no significant effect on ads actually influencing vote choice. In other words, the effectiveness of presidential campaign ads – even targeted ones – on persuading people to vote for one candidate or another may be greatly overstated.