College students, political engagement and Snapchat in the 2020 general election

Prof Laurie L. Rice

Professor of Political Science at Southern Illinois University Edwardsville. She co-authored Web 2.0 and the Political Mobilization of College Students and her work on social media use and civic engagement also appears in Social Science Computer Reviewand Journal of Information Technology & Politics.

Email: larice@siue.edu

Prof Kenneth W. Moffett

Professor of Political Science at Southern Illinois University Edwardsville. His research interests lie in American politics and policy. He is coauthor of Web 2.0 and the Political Mobilization of College Students, and specializes in American politics and policy.

Email: kmoffet@siue.edu

Section 5: Social media

Snapchat has grown among young adults, with Snapchat now claiming over 90% of Americans between the ages of 13 and 24 as users. Usually, we associate Snapchat with entertaining filters and disappearing content. However, Snapchat provides a home for political activity among young adults, too.

In 2016, we found that sending political pictures and political videos increased civic engagement among college students. Among other activities, civic engagement consists of persuading others why they should vote for or against a party or candidate, attending political meetings, and participating in political activities like marches or protests.

The 2020 election has seen record voter turnout and political participation among young people, with records being shattered in Texas, California, and Wisconsin. Does Snapchat deserve part of the credit for enhancing civic engagement? Did these politically-oriented activities on Snapchat increase civic activity among college students in 2020 as they did in 2016?

To examine those questions, we performed a survey of students at a large midwestern public university (Southern Illinois University Edwardsville) in October 2020. Students at this university are reasonably representative of college students as a whole on several dimensions including their voting rates.

Our results show that sending pictures and videos about candidates, political parties, or interest groups increased civic engagement levels, as shown in Figures 1 and 2. In both figures, the horizontal axis shows the frequency of political Snapchat activity and the vertical axis is the average treatment effect on the treated at the varying frequencies of actions on Snapchat. In addition, the solid line is the average treatment effect on the treated at each Snapchat activity level, and the dashed lines are 95% confidence intervals around these estimates.

These figures indicate that even rarely sending political pictures or videos is associated with two to three-point increases in civic engagement. Meanwhile, sending political pictures or political videos regularly on Snapchat both result in slightly over four-point increases in civic engagement. For context, a two-point increase in civic engagement is equivalent to participating rarely in two activities that one otherwise would not have engaged in or to moving from participating never in an activity to sometimes in an activity.

Meanwhile, a four-point increase is equivalent to engaging in one additional activity very often that one would otherwise not have participated in or participating sometimes in two additional activities that one otherwise would not have done. Yet, we did not find any evidence that monitoring what candidates for office, political parties, or interest groups post on Snapchat is connected with increased civic engagement.

We employed a statistical technique called matching to obtain our results. Matching allows us to isolate the impact of political uses of Snapchat among those otherwise similarly inclined to civic engagement. To do so, we considered many other factors connected with civic engagement, with a complete listing available in the online appendix.

To ensure confidence in our results, we need to address two potential limitations to our analyses. First, matching requires that we satisfy a series of stringent assumptions or else the results are not valid. Our dataset and analyses meet these assumptions.

Second, if the groups that engaged in Snapchat by sharing pictures or videos at varying levels differ from those who did not perform such activities, then we cannot determine a causal effect of these Snapchat activities. This means that we would not be able to differentiate between the effects of other, preexisting differences (like interest levels in politics) and activities on Snapchat on civic engagement. We checked to see if this occurred. After matching, there are no statistically significant preexisting differences between those who engaged in sharing pictures or videos at varying levels and those who did not. Thus, we can rule out the effects of preexisting differences.

Yet, what if we have specified our matching routines such that they conveniently produce the results that we report here? That’s known as p-hacking, and it’s a large, well-known problem in scientific research. To guard against this practice, we removed one of the factors that we discuss in the previous paragraph and re-ran our statistical model. We repeat that procedure for each factor and each level of monitoring what others say on Snapchat, sending pictures, and sending videos on Snapchat. Our results are robust to these alternate specifications, as 82.41% (or, 89/108) of the alternate specifications confirm our results.

In sum, political expression on Snapchat is partly responsible for higher levels of civic activity among college students in 2020. While Snapchat may not be as frequently used as social media apps like Facebook and Instagram, it provided an important vehicle for increasing civic and political activity among college students in 2020. Thus, online political engagement increases civic activity away from the internet, too. We anticipate that this trend will continue.