手法を比較

	Cohort-Sequential Design	Age-Period-Cohort Analysis
分野	Social Epidemiology	Social Epidemiology
系統	Process / pipeline	Regression model
提唱年	1953	1983
提唱者	Richard Q. Bell (convergence/accelerated approach); Yasuo Miyazaki & Stephen Raudenbush (cohort-linkage tests)	Theodore R. Holford; Yang Yang & Kenneth C. Land (intrinsic estimator)
種類	Observational longitudinal design linking overlapping age cohorts	Generalized linear model for rates indexed by age, period, and cohort
原典	Bell, R. Q. (1953). Convergence: an accelerated longitudinal approach. Child Development, 24(2), 145-152.	Holford, T. R. (1983). The Estimation of Age, Period and Cohort Effects for Vital Rates. Biometrics, 39(2), 311-324.
別名	Accelerated Longitudinal Design, Convergence Design, Cohort-Sequential Accelerated Design, Overlapping-Cohort Longitudinal Design	APC Analysis, Age-Period-Cohort Models, Cohort Analysis of Rates, Intrinsic Estimator APC
関連	3	4
概要	The cohort-sequential design — also called the accelerated longitudinal design — spans a long age range quickly by following several overlapping age cohorts for a short time each and then statistically linking their trajectory segments into one long developmental curve. Richard Bell introduced the idea in 1953 as 'convergence,' a way to study development over many years without waiting many years. Instead of following one cohort from, say, age 10 to age 20 for a full decade, the design enrolls cohorts aged 10, 12, 14, 16, and 18 and follows each for two or three years, with adjacent cohorts overlapping in age so their pieces can be joined. Yasuo Miyazaki and Stephen Raudenbush later supplied the formal multilevel tests for whether the cohorts can legitimately be linked. The design trades a single continuous within-person record for speed, while using overlap to check that the assembled curve is coherent.	Age-period-cohort (APC) analysis decomposes variation in disease or mortality rates into three temporal components: the effect of age (biological and accumulated risk), the effect of period (influences hitting everyone alive at a given calendar time, such as a new treatment or a recession), and the effect of cohort (lasting imprints of the conditions into which a birth generation was born). Theodore Holford's 1983 Biometrics paper gave the canonical generalized-linear-model formulation and exposed the method's defining obstacle: because cohort equals period minus age, the three predictors are exactly linearly dependent, so their individual linear slopes cannot be separately identified. A large methodological literature has since proposed constraints, reparameterizations, and estimators to extract whatever the data can legitimately support. Yang, Schulhofer-Wohl, Fu, and Land's 2008 work popularized the intrinsic estimator, a principled choice among the infinitely many fitting solutions. APC analysis is a workhorse of descriptive epidemiology and demography, used to read the temporal fingerprints left on rates of cancer, suicide, obesity, and mortality. Done carefully it separates signal from artifact; done carelessly it manufactures trends that the identification problem makes unknowable.