You are here

Does the Media Matter? A Field Experiment Measuring the Effect of Newspapers on Voting Behavior and Political Opinions

April 2009

Alan S. Gerber, Dean Karlan, Daniel Bergan


This article is centered around the first natural field experiment aimed to see the impact of political news content on political behavior and attitudes. To conduct this test, Prince William, a Virginia county, was used to discover which households do not receive the newspapers, the Post or the Times. Individuals were sampled from a list of registered voters and a consumer database list. They were then randomly assigned subscriptions to these papers for 10 weeks, or to neither for the control group. The Post is known to lean left, while the Times leans right. However, receiving either paper led to more Democratic support, which indicated that media slant did not matter as much as simply exposure. The Post estimated increasing probability of choosing the Democrat by 11.2 percentage points among voters, and by 7.2 across all respondents. This experiment also found an increased turnout of citizens in 2006 that was barely significant, but no significant increase in self- reported or administratively measured turnout in 2005. This study found three problems with noncompliance in terms of treatment administration that are worth noting. The first is that six percent of people chose not to receive the free newspaper subscriptions. The second is some of the home addresses were said to be undeliverable. The final issue is that a number of the households were already subscribing to either of the papers before the study was conducted. It also found a few limitations that are important. One is the relatively small sample size, and another is that although houses receive papers, it cannot be proven if the individuals read them.

Read Now

Policy Implications

This article found that the effects of increased news exposure has a much larger consequence than the political slant of the news received. From this one experiment, it was confirmed that even short periods of exposure to newspaper influences a person’s behaviors in terms of voting, and may also influence turnout in elections. To further investigate, this same type of experiment should be done using different forms of media, different people, and different political contexts. This increase in news should be enforced at the state, local, and national level of government for all sectors of society

Find Similar General/Not Specific Research
Find Similar Government Operations Research