About the Society
Papers, Posters, Syllabi
Submit an Item
Polmeth Mailing List
Below results based on the criteria 'random effects model'
Total number of records returned: 2
Voter Turnout and the Life Cycle: A Latent Growth Curve Analysis
random effects model
latent growth models
The distinctive relationship between age and voter turnout has intrigued students of electoral behavior since at least the early 1960s. Nevertheless, political scientists actually know little about how individuals acquire the habit of voting during young adulthood. Moreover, previous speculations and explanations are all questionable because they are based on data and models that are inappropriate for what is essentially a developmental process. Problems include confounding age with generational effects, assumptions of reversibility of gains in participation from key life events, and a failure account for the fact that an individual's probability of turnout at any particular age is a function of two distinct latent variables: their turnout rate in the very first elections, and their subsequent rate of increase. Theory construction is muddled because these two variables are negatively correlated and have different predictors. This study uses longitudinal data covering young voters over their first four presidential elections and uses latent growth curve models -- a special case of multi-level or Hierarchical Linear Models which are finding wide applicability in the social sciences. Given appropriate data, this approach permits statistical models that better correspond to life-cycle hypotheses. The findings clarify the role of parental influence, marriage and parenthood, while raising questions about the costs of mobility.
Estimation in Dirichlet Random Effects Models
generalized linear mixed model
Dirichlet process random effects model
precision parameter likelihood
probit mixed Dirichlet random effects model
We develop a new Gibbs sampler for a linear mixed model with a Dirichlet process random effect term, which is easily extended to a generalized linear mixed model with a probit link function. Our Gibbs sampler exploits the properties of the multinomial and Dirichlet distribution, and is shown to be an improvement, in terms of operator norm and efficiency, over other commonly used MCMC algorithms. We also investigate methods for the estimation of the precision parameter of the Dirichlet process, finding that maximum likelihood may not be desirable, but a posterior mode is a reasonable approach. Examples are given to show how these models perform on real data. Our results complement both the theoretical basis of the Dirichlet process nonparametric prior and the computational work that has been done to date. Forthcoming: Annals of Statistics.