Political Science

Hubbard Finds Proximity to Terrorist Attack Increases Voter Turnout

Posted Jun 19, 2017

                 Great Britain has been struck by four terrorist attacks the past 4 months. In March of this year, five people were killed by a knifeman outside Parliament while others were hurt by the man’s van careening across a bridge trying to hit people. On May 22, twenty-two people were killed in a terrorist attack at an Ariana Grande concert in Manchester. Then on June 3, seven people were killed by a van trying to hit people on London Bridge and then stabbed near the Borough Market. Then today (June 19), another van attack killed one and injured 10 others. If terrorism is going to become more common (and we hope it will not), what impact will this have on elections? Will it increase voter turnout? Will it lead to a loss of support to the governing party because they will be held responsible for failing to prevent it? Will it lead to support for the party perceived to be stronger against terror?

                While we cannot answer every question, Union sophomore Clark Hubbard provides some preliminary insights into whether terrorist attacks will increase voter turnout. Using data on 12 terrorist attacks in the United States, Hubbard finds preliminary evidence that turnout will increase if the attack is close to the election. However, the small dataset makes it difficult to draw definite conclusions and this concept will require further study. In this blog post, we find a brief summary of the literature regarding turnout, some possible explanations for how terrorism will influence voter turnout, a discussion of his methods and data, his results, and a discussion of the implications for his results for our democracy.

Explanations for the Decline in Voting

                The American public has seen a significant decline in voter turnout for several decades (Abramson and Aldrich, 1982; Cassel and Luskin, 1988; Jackman, 1987). This lower rate of voter turnout is especially striking when compared to other democratic, developed nations. When explaining the decline in voter turnout, political scientists turn to several theories to explain this phenomena. First, modernization theory states that a decrease in voter turnout is directly connected with modern conveniences which distract us from politics as the rise of television has distracted the public from politics, replacing what was once a sort of form of entertainment with another (Gentzkow, 2006) and as technology replaces a more personal and face-to-face politics (Gerber and Green 2000).

                Second, rational choice theories argues that voters do not participate because the costs of voting outweigh the benefits (Filer & Kenny, 1980). Essentially, voters do not want to seek out the information and pay the costs of being a citizen unless an issue important to a voter is raised (Hahn 1970) or a candidate directly represents one’s interest (Matsusaka 1995).

                Third, the institutional approach suggests that the rules of voting make it more or less difficult to vote. Scholars have found that the complex registration system (Powell 1986), similar candidate on major issues (Jackman 1987), gerrymandering or redrawing districts to favor certain parties (Merrifield 1993), and voting rules in general (Engstrom 2012) discourage turnout.  

                Fourth, economics influences voter turnout in various ways. Some argue that increased wealth and quality of life results in lower turnout, outside of presidential elections, because voters are largely content (Charles and Stephens 2013) while others argue that less education and difficulties placed in front of poor people discourage voting (Rosenstone 1982).

                Finally, the non-systematic approach examines those variables that are unpredictable that influence voting. For example, Gomez, Hansford, & Krause (2007) stating that both rain and snow result in a lower voter turnout, especially for Democrats while Binder, Binder & Childers (2012) find that citizen initiative campaigns can be used to bolster turnout. From this theory also comes the idea that terrorist activity will increase voter turnout (Robbins, Hunter, Murray, 2013; McAllister, 2004). Hubbard builds upon these latter studies to see if terrorist attacks improve voter turnout.

The Impact of Terrorism of Voting

                Hubbard suggests several hypotheses on the relationship between terrorism and voter turnout. First, he posits that terrorist attacks will generally increase voter turnout. The logic is that a terrorist attack will generate a lot of media coverage because terrorist attacks in the United States are so rare. With greater media coverage, voters will be more aware of the attack and potentially more scared. Americans will respond to a terrorist attack with a surge in patriotism, or at least discontent with the system, which leads more people to vote as higher turnout expresses a belief that the Americans will stand up to terrorist.

                However, it is possible that terrorist attacks matter in different circumstances. First, the closer to an election, the more likely turnout will be higher. The logic of this hypothesis is simple. The more recent the attack, the more likely the attack is on the voters’ minds and the more mobilized voters will be. If the attack occurred less recently, it is possible that other issues have risen in the voters’ minds that will influence their decision to vote.

                Second, the more people killed in an attack, the more likely turnout will increase. If more people are killed, the attack is more newsworthy because the results are more traumatic. The greater media coverage raises the importance of terrorism in the minds of voters and leads them to vote to show solidarity with the nation and its ideals. Unfortunately, terrorist attacks with smaller casualty counts may slide from public memory faster.

Third, Islamic terrorism will lead to higher turnout. Unfortunately, the amount of gun violence in this country makes mass shootings feel common today. However if ISIS or Islamic radicalism is involved, people will pay more attention because these kinds of attacks are less frequent and more newsworthy. Moreover, the national government is directly responsible for protecting Americans from foreign attack which raises the stakes of the terrorist attack and increases the incentives for Americans to vote.

Data and Methods

To test this hypothesis, Hubbard collected voter turnout data from 12 separate terrorist attacks both before and after the attacks. This data collection would allow Hubbard to conduct an intervention analysis for each county in which the terrorist attack occurred to determine its impact. For this paper terrorist attacks are not just attacks from foreign operatives attempting to disrupt the order of American society. In this paper, terrorist attacks can be domestic terrorism, and can include not just terrorist attacks, but also school shootings, bombings, and similar incidents. The dictionary definition of a “terrorist attack” is an attack by non-governmental entities for religious or political gains. For brevity’s sake, only terrorist attacks with 5 or more deaths have been examined.  Using Wikipedia, he identified terrorist attacks with the basic data needed for his variables (the number of deaths, date, number killed, and whether or not it was Islamic). The information Wikipedia was then reconfirmed using local papers where the attack occurred.

The number of deaths is measured as a continuous variable that reflects the number of those killed. The data proximity variable is how many months have passed since the terrorist attack. Islamic terrorism is identified by a dummy variable where 1 equals an attack motivated by radical Islam and 0 otherwise. There is also a dummy variable for a terrorist attacks occurred during this election season. His dependent variable is county voter turnout and is the percentage one gets when dividing the number of voters in a specific election by the number of total registered voters. Most of this data came from state Secretary of State websites.

Results

In four of the twelve attacks for which significant data could be found (the September 11th attacks, the Orlando Nightclub shooting, the University of Iowa shooting, and the Westroads Mall shooting) there was an increase in voter turnout. In the other eight instances, no bump was seen. This does not disprove the overall hypothesis per se, but demonstrates that the number of testable instances in the United States is not sufficient, and is too small of a sample size. However if you look at counties that had terrorist attacks occur in them, there was generally a 2-2.1% increase in voter turnout. To demonstrate how terrorism may influence voter turnout, Hubbard provides four graphs that demonstrate the effect (not shown). These graphs show that there was an initial surge after the attack but that turnout usually shows a general decline after the attack. The upsurge in these four attacks demonstrates the potential that terrorist attacks have on turnout.

         Hubbard then uses linear regression on the twelve instances to determine the impact of his three other hypotheses on voter turnout. As Table 1 demonstrates, the only variable that is statistically significant is data proximity which has a positive relationship. In other words, the closer the terrorist attack to the election, the higher turnout will increase. This result is not surprising since the more recent the attack, the more voters will recall it which may affect their decision to vote or not. The other three variables are not significant which suggests that larger attacks, Islamic radical inspired terrorism, and terrorist attacks, in general, do not influence turnout. The proxy for terrorist attacks is negative and barely misses the test for significance. Thus, this variable would suggest that terrorist attacks might decrease voter turnout, unless near the election. This result would be rather surprising but could probably be explained by other issues rising in importance and replacing terrorism as a cause of voter turnout.

 ** significant at the .05 level

Conclusion

While there is insufficient data to conclude that terrorist attack actually increase voter turnout, there is preliminary evidence that suggests the closer the attack occurs, the more likely turnout will increase. This result suggests that terrorists may not win by changing how we act. If terrorism occurs near an election, the American people respond by voting to send a message that we will not be intimidated and the values that the world so highly values are the ones that we will stand up for when under attack.

However due to the small number of terrorist attacks in the United States, there is insufficient data to test. Hopefully, the United States will not see more terrorist attacks for which to find additional data but general trends suggest, unfortunately, that more terrorist attacks may come to the United States. With the rise of terrorist attacks around the world, there may begin to be more data to more fully test this hypothesis.