The effect of internet usage on voter turnout in the European Union

Abstract

This master thesis investigates the effect of internet usage on voter turnout in the European Union over the sample period from 1990 to 2016. The methodology applies is both ordinary least squares (OLS) estimation and the fixed effect model approach, in which the dependent variable is voter turnout and the independent variable is the internet usage. Both socioeconomic variables such as population, gender (female) and age, in addition to macroeconomic variables such as GDP per capita and the unemployment rate are used as a control variable in the regressions. The main findings suggest a positive and statistically significant effect of internet usage on voter turnout in the European Union. Moreover, the findings from the OLS estimation and the fixed effects models only differ slightly, which makes the simultaneity problem less likely in the empirical analysis. The sensitivity analysis conducted in this thesis examine the robustness of the main findings by firstly excluding the female variable as a control variable and secondly by excluding Belgium and Luxembourg from the data set due to compulsory voting in these countries. In both cases, the estimated effect of internet usage on the voter turnout remains positive and huge in magnitude and statistically significant at conventional levels. That said, all findings reported in this thesis should be considered with some caution, as more comprehensive sensitivity analysis with respect to control variables not used in the empirical analysis may be conducted. Such comprehensive sensitivity analysis has been beyond the scope of this thesis and is left for future research.