Reconciling Individual and Aggregate Evidence Concerning Partisan Stability: Applying Time-Series Models to Panel Survey Data

Party identification has been studied extensively using bothindividual- and aggregate-level data. This paper attempts toformulate a statistical model that can account for the rangeof empirical generalizations that have emerged from aggregatetime series and panel surveys. Using Monte Carlo simulation,we show that only certain types of data generation processescan account for these empirical regularities. Deciding whichof the remaining types best explains the data means investigatingthe ways in which individual-level partisanship behaves overtime. Partisanship at the aggregate level tends to be highlyautocorrelated, reequilibrating slowly in the wake of each perturbation.Working downward from the analysis of aggregate data, previousresearchers argued that aggregate partisanship is fractionallyintegrated and contended that dynamics at the individual levelare therefore heterogeneous. Using data from three panel surveys,we present the first direct assessment of individual-level dynamics.We also investigate the hypothesis that these dynamics varyamong individuals, a claim that motivates much recent work onfractionally integrated time series. The model that best explainsthe observed characteristics of party identification is onein which individuals respond in similar ways to external shocks,reequilibrate rapidly thereafter, and seldom change their equilibriumlevel of partisan attachment.