Suchergebnisse

Social indicators and demographic rates are often arrayed over time by age. The patterns of rates by age at one point in time may not reflect the effects associated with aging, which are more properly studied in cohorts. Cohort succession, aging, and period-specific historical events provide accounts of social and demographic change. Because cohort membership can be defined by age at a particular period, the statistical partitioning of age from period and cohort effects focuses attention on identifying restrictions. When applying statistical models to social data, identification issues are ubiquitous, so some of the debates that vexed the formative literature on age–period– cohort models can now be understood in a larger context. Four new articles on age–period–cohort modeling call attention to the multilevel nature of the problem and draw on advances in methods including nonparametric smoothing, fixed and random effects, and identification in structural or causal models.

In a widely influential essay, Ryder argued that to understand social change, researchers should compare cohort careers, contrasting how different cohorts change over the life cycle with respect to some outcome. Ryder, however, provided few technical details on how to actually conduct a cohort analysis. In this article, the authors develop a framework for analyzing temporally structured data grounded in the construction, comparison, and decomposition of cohort careers. The authors begin by illustrating how one can analyze age-period-cohort (APC) data by constructing graphs of cohort careers. Although a useful starting point, the major problem with this approach is that the graphs are typically of sufficient complexity that it can be difficult, if not impossible, to discern the underlying trends and patterns in the data. To provide a more useful foundation for cohort analysis, the authors therefore introduce three distinct improvements over the purely graphical approach. First, they provide a mathematical definition of a cohort career, demonstrating how the underlying parameters of interest can be estimated using a reparameterized version of the conventional APC model. The authors call this the life cycle and social change (LC-SC) model. Second, they contrast the proposed model with two alternative three-factor APC models and all logically possible two-factor models, showing that none of these other models are adequate for fully representing Ryder's ideas. Third, the authors present the article's major accomplishment: using the LC-SC model, they show how a collection of cohort careers can be decomposed into just four basic components: a curve representing an overall intracohort trend (or life cycle change); a curve representing an overall intercohort trend (or social change); a set of common cross-period temporal fluctuations that permit variability across cohort careers; and, finally, a set of terms representing cell-specific heterogeneity (or, equivalently, interactions among age, period, and/or cohort). As the authors demonstrate, these parts can be reassembled into simpler versions of cohort careers, revealing underlying trends and patterns that may not be evident otherwise. The authors illustrate this approach by analyzing trends in political party strength in the General Social Survey.

This paper examines the relationship between family career and age cohort. The family career concept originates in the work of developmental theorists such as Rodgers (1973). It is argued that the concept of the family career represents a more sophisticated approach and understanding of developmental theory than does the use of the present stage of the life cycle. The results of this research suggest that age cohort has little effect on the pattern of stages in the family career but that age cohort is moderately associated with the timing or spacing between stages.

This paper develops a design-based approach to identifying cohort effects in APC analyses. Cohort effects arise when one cohort is treated by a unique set of formative socialization experiences, which causes it to differ from other cohorts in relevant outcomes. APC analyses typically compare treated and untreated cohorts from a single population. Our approach introduces a second group-a control group, in which no unit is treated but that is otherwise similar to the first-and adapts difference-in-differences estimation to the APC framework. The approach yields two identification strategies, each based on transparent and testable assumptions. We illustrate how the method works and what is to be gained through three examples. © 2013 Elsevier Ltd.