Clustered Ordinal Responses: Random Effects Models
Alan Agresti
Abstract
Alan Agresti
Abstract
This chapter presents the alternative model type for clustered ordinal data, which contains a random effect term in the linear predictor for each cluster. Such models permit heterogeneity in response probabilities for the various clusters that have a particular setting of explanatory variables. The chapter provides examples of models for which the random effect plays the role of an intercept term that varies among clusters. It then discusses multilevel (hierarchical) models in which random effects enter at various levels, such as occurs in educational applications that include random effects for students as well as for schools. The chapter also talks about relevant issues in choosing between a random effects model, a marginal model, and some other type of model, such as a transitional model. Controlled Vocabulary Terms hierarchical linear model
OpenAlex reports 7 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
This chapter presents the alternative model type for clustered ordinal data, which contains a random effect term in the linear predictor for each cluster. Such models permit heterogeneity in response probabilities for the various clusters that have a particular setting of explanatory variables. The chapter provides examples of models for which the random effect plays the role of an intercept term that varies among clusters. It then discusses multilevel (hierarchical) models in which random effects enter at various levels, such as occurs in educational applications that include random effects for students as well as for schools. The chapter also talks about relevant issues in choosing between a random effects model, a marginal model, and some other type of model, such as a transitional model. Controlled Vocabulary Terms hierarchical linear model
Key concepts: Random effects model, Multilevel model, Marginal model, Hierarchical database model, Term (time), Econometrics, Linear model, Ordinal regression