Age, Period and Cohort Effects最新文献

{"title":"Age–period–cohort analysis","authors":"Herbert L. Smith","doi":"10.4324/9780429056819-9","DOIUrl":"https://doi.org/10.4324/9780429056819-9","url":null,"abstract":"","PeriodicalId":210706,"journal":{"name":"Age, Period and Cohort Effects","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121069594","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bayesian age–period–cohort models","authors":"E. Fosse","doi":"10.4324/9780429056819-8","DOIUrl":"https://doi.org/10.4324/9780429056819-8","url":null,"abstract":"Researchers in a wide range of fields have long sought to understand social and cultural change by identifying the unique contributions of age, period, and cohort (APC) processes on various outcomes (Ryder 1965). The basic idea is that any temporal change can be attributed to three kinds of processes (Glenn 2005: 11): (1) changes over the life course of individuals, or age effects;1 (2) changes due to the events in particular years, or period effects; (3) changes due to the replacement of older cohorts of individuals with younger ones with different characteristics, or cohort effects. However, in what has been called the APC identification problem (Mason and Fienberg 1985b: 68), the linear effects of an APC model cannot be uniquely estimated due to the perfect linear dependency among the age, period, and cohort variables. Intuitively, once we known a person’s age and the year of measurement (or period), then we also know that person’s birth year (or cohort). A variety of approaches have been proposed to deal with the APC identification problem, but the great majority of studies have taken a frequentist rather than Bayesian perspective. As I discuss in this chapter, the main advantage of the Bayesian framework for APC analysis is that the analyst can, in principle, explicitly incorporate theoretical considerations, qualitative judgments, and additional data into the inferential process. In the frequentist tradition, researchers have long noted the importance of integrating what the political scientist Philip Converse (1976) dubbed “side information” into an APC analysis, typically in the form of constraints on the parameters. However, the Bayesian paradigm, which places specification of the prior distribution front and center, underscores that it is more appropriate to talk about the importance of incorporating primary information into any APC analysis. The task ahead, then, is to develop a general Bayesian APC model that allows researchers to easily and explicitly incorporate such primary information into their analyses. Doing so will require specifying what the statistician Paul Gustafson (2015:15-18) calls a transparent reparameterization of an APC model, which renders clear the impact of nonidentifiability on one’s conclusions. The rest of this chapter is organized as follows. First, I introduce the basics of the Bayesian approach to inference. Second, I outline the classical APC (or C-APC) regression model in terms of the Bayesian perspective, comparing it with the classical (or frequentist) approach. Third, I discuss the identification problem from a Bayesian perspective using a transparent reparameterization of the C-APC model. Fourth, I delineate a general framework for Bayesian APC modeling. In doing so, I review previous Bayesian approaches, which have focused primarily on developing reliable forecasts and applying mechanical, one-size-fits-all prior distributions. In addition, I discuss the specification of prior distributions in APC an","PeriodicalId":210706,"journal":{"name":"Age, Period and Cohort Effects","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124020907","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}