Ver artículo

involved in the form of cohort analysis that concerns us can be elucidated with special clarity in the discrete data context. The version of cohort analysis we consider involves measurements on some dependent quantity conditional on the ages of the units of observation and the period at which the measurements were taken. We assume that replicated cross-sections are available and, in general, that repeated measurements on the same units of observation across replications are not available. The dates of data collection define the periods; and for each cross-section, measurements on age, on the dependent quantity, and possibly on other variables are available. Cohort membership is then defined by knowledge of age and the date or period at which that age was attained. The questions of interest for the analyst are whether age, period, and cohort simultaneously determine the endogenous quantity-and if so, how …