Generations of Presidential Voting

19 July 2014, 1158 EDT

As we all know, the social sciences are a messy business. People change their minds, don’t always follow law-like rules, and often have the guts to defy our theories by reflecting on their past behavior. For these reasons, it is always nice to see when our work receives support from other scholars, especially when these scholars operate within another paradigm, sub-field, or use a different methodological approach. In the case of my own work, which focuses on the explanatory value of the concept of generations and its applications in Foreign Policy and International Relations, this just happened. On July 7, Yahir Ghitza and Andrew Gelman at Columbia University published “The Great Society, Reagan’s Revolution, and Generations of Presidential Voting,” a working paper, which was featured prominently, with some very pretty graphics, in the New York Times. Ghitza and Gelman develop a generational model of presidential voting, which is then tested against a data set that combines four different sources of presidential polling data. The model is an adaption of a weighted “running tally,” which allows the authors to determine the age at which retrospective evaluations of presidential performance are most formative for determining future presidential voting choices. Ghitza and Gelman show convincingly that political events which are experienced between the ages of 18 and 24 years are most significant in shaping the presidential voting preferences of individuals throughout the rest of their life cycles. This does not necessarily constitute big news for generation scholars, who have more or less agreed on those rough age brackets for decades now, as the authors point out themselves. However, given the large number of observations, 306,011 to be precise, and the strong support that the model receives in the statistical analysis, the results are clearly impressive and provide strong support for the argument that the age of youth is crucial for forming the political worldviews of voters, at least amongst non-Southern whites in the United States. At the same time, the model by Ghitza and Gelman does have a number of limitations and their general approach is not really applicable to studying generational phenomena in foreign policy or international politics. Why this is the case, and how we can study generations in the context of FP and IR will be discussed in the next couple posts. Stay tuned.