Cohort Analysis

Period Effects

As a way to motivate the discussion, let us imagine a population of interest, voting for one of the two parties, A and B, which compete in elections that are held every 4 years. Assuming that the only source of change in the composition of the population is that of succession, that birth and death rates are approximately equal, and that the structure of the party system remains stable, the question is how can political change occur in this context? It turns out that change regarding a social attribute of interest under these circumstances is essentially produced through two different mechanisms, both of which can be traced in Table 1, which distinguishes only between two subgroups of interest, namely, young and old voters (the reason for using age as a forcing variable will be clear in the next paragraph). The first mechanism involves a coming event or a sequence of events that might induce people to change their minds. This is referred to as a period effect. For instance, people might vote predominantly for Party A during Time 1. An event takes place, and they shift to Party B. If this shift occurs at approximately equal rates between different age-groups, one should observe a difference between the two columns of Table 1, whereas no difference should be found between the two rows of each column. This is shown in the upper panel of Table 1.

Cohort Effects

Political events, however, are unlikely to affect all people in the same way. Most of the time researchers [Page 290]observe (or assume) particular cohorts, that is, groups of individuals who pass some crucial stage of their lives at approximately the same time, to be considerably more affected by political events than other cohorts. Although different cohorts of interest have been examined, the most common approach is to distinguish between new cohorts and older ones with the first deemed more susceptible to political stimuli. That this assumption is implicitly made in most cohort analyses is evident from the definitions given to different cohorts so as to distinguish one from the other. In most cases, each cohort is characterized by the salient historical and/or political features that marked its early adulthood (the “war generation,” the “1968 generation,” etc.). If such shaping socialization processes dominate over inertial forces of intergenerational transmission of political attitudes, young adults can differ in terms of their value orientations or any other attitudinal aspect of interest from one generation to another. This second mechanism of change, known as a cohort effect, is presented in the second panel of Table 1. Whereas the older age-group remains stable over time, young cohorts shift from Party A to Party B. According to this scenario, change is driven only through the entrance of new generations.

...