Acceso usuarios
Thomas L. Brunell, Bernard Grofman, Samuel Merrill III
The two hallmarks of a critical election and, hence, of a critical realignment are the magnitude of the observed change and the durability of that change. In addition to offering a new approach to measuring durable change in national party dominance, and providing a non-parametric criterion to identify unusual changes in seat/vote shares, we provide fresh insights via a unifying statistical approach that reflects both of these factors simultaneously. Furthermore, we assess the robustness of critical election determinations in two ways. First, we compare the magnitude of inter-election shifts with both average volatility over the entire time period and volatility relative to a particular time period. Second, as an alternative to the usual perspective, we consider critical elections not as a one-time cataclysm, but rather as a pair (or perhaps even triple) of consecutive substantial shifts, generated by the same underlying factors. Overall, we distinguish six elections that marginally or provisionally meet our criteria to be critical elections. But focusing on pairs of elections, 1858–60 and 1930–32 stand out as critical among all elections since the 1850s.
