News Release 

Televised political campaign ads with different features have similarly small effects on voters

The small effects of political advertising are small regardless of context, message, sender, or receiver: Evidence from 59 real-time randomized experiments

American Association for the Advancement of Science

Research News

New research suggests that the effect of any individual televised political ad on how much a subject favors a candidate and who they plan to vote for is relatively minor. This is regardless of message, context sender, and receiver. The study's findings are based on a comprehensive analysis of 49 professionally produced political campaign ads presented to 34,000 nationally representative subjects. Despite the marginal effect observed for each individual ad, Alexander Coppock and colleagues note that advertising campaigns as a whole may still play an important role in election outcomes, particularly during a close election. Social scientists have believed that persuasive messages tend to result in small effects because persuasion only works well under particular sets of circumstances. However, it has been difficult to synthesize the results of previous experiments into one coherent theory since different experiments hold variables constant at inconsistent levels, and contain divergent research designs, instruments, and sampling methods. To understand the relative persuasive power of varying televised campaign ads between March and November 2016, Coppock et al. conducted 59 experiments covering both the primaries and the period leading up to the general election. Subjects were shown videos of either one of two campaign ads released that week, both campaign ads (in some trials), or a car insurance ad (as a control). After watching the videos, the subjects rated each candidate on a five-point scale and indicated who they planned to vote for in the general election. The ads swayed candidate favorability by a measure of about 0.05 points on a 5-point scale and influenced vote choice by an average of 0.7%, with little variation from person to person or from one ad to another. Each ad's impact was about the same no matter which campaign produced it, in what context it was presented (including when and where during the election cycle), or whether it belonged to an attack or promotional campaign. Furthermore, subjects showed about the same degree of response to an ad even when they lived in different states or favored different parties. Based on these findings, Coppock et al. suggest election campaigns carefully consider whether tailoring ads for specific audiences is worth the cost.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.