Skip to content. | Skip to navigation

Personal tools

Navigation

You are here: Home / Coming Up / Seminar Series: Constrained Estimators and the Cross-Classified Models for the Age-Period-Cohort Analysis

Seminar Series: Constrained Estimators and the Cross-Classified Models for the Age-Period-Cohort Analysis

Liying Luo, Research Assistant, University of Minnesota
When Oct 01, 2012
from 12:00 PM to 01:00 PM
Where 0124B Cole Student Activities Building
Contact Name
Contact Phone 301-405-6403
Add event to calendar vCal
iCal

About the Talk

In many different fields, social scientists desire to understand temporal variation associated with age, time period, and cohort membership. Among methods proposed to address the identification problem in Age-Period-Cohort analysis, the cross classified approach, including Cross-Classified Fixed Effects Models (CCFEM) and Cross-Classified Random Effects Models (CCREM), appear to solve the identification problem and to yield good estimates of the independent effects of age, period, and cohort groups. This paper assesses the validity and application scope of CCFEM and CCREM theoretically and illustrates their properties with simulations. It shows that the cross-classified methods do not break the identification problem; rather, they implicitly assume multiple constraints on the age, period, and cohort effects. These constraints not only depend on the widths of the age, period, and cohort intervals but also have non-trivial implications for estimation.  Because these assumptions are extremely difficult, if not impossible, to verify in empirical research, the author concludes that CCFEM and CCREM cannot and should not be used to recover the underlying age, period, and cohort trends.

About the Speaker

Liying Luo

Liying Luo is a Ph.D. candidate in Sociology at the University of Minnesota. She obtained her M.S. degree in Biostatistics, School of Public Health at the University of Minnesota. Ms. Luo's research interest includes demography, social stratification, and health inequalities. Her dissertation critically examines the validity of existing Age-Period-Cohort (APC) methods and proposes a new APC model that is more theoretically informed and methodologically valid. She is also collaborating with Dr. John Rob Warren, Dr. Alberto Palloni, and Dr. Jim Raymo to examine whether and under what circumstances different methods for modeling longitudinal trajectories yield the same results. After graduation, Ms. Luo will be looking for a faculty position in a research institute in the areas of demography, quantitative methods, and health disparities.

October 2020 »
October
SuMoTuWeThFrSa
123
45678910
11121314151617
18192021222324
25262728293031