In an era when elections scholars expected American national presidential election turnout to increase, its steep, prolonged post-1960 decline sparked deep concern and generated an avalanche of individual-level analyses searching for explanation. The post-1960 decline, however, no longer dominates turnout’s trajectory; it has been on the upswing since 1996. This complicates our understanding as we have yet to settle on turnout’s description, much less its explanation. Here we introduce the first political science-oriented, multivariate modeling of American national presidential election turnout. Our results offer a mix of important confirmatory and original findings. First, we discover that modeling turnout’s decline as a post-1968 secular disturbance reveals turnout’s expected steady increase across the modern era (1952–2020). Second, we show that turnout’s increase can be traced to increased polarization working its influence indirectly through the direct, positive turnout affects of voter external efficacy and negative presidential campaign advertising (1960–2012).
Color texture image retrieval based on Copula multivariate modeling in the Shearlet domain
