A default conjugate prior for variance components in generalized linear mixed models (comment on article by Browne and Draper)

By Robert E. Kass, R. Natarajan. 2006

81 citations recorded by the source corpus.

. For a scalar random-eﬁect variance, Browne and Draper (2005) have found that the uniform prior works well. It would be valuable to know more about the vector case, in which a second-stage prior on the random eﬁects variance matrix D is needed. We suggest consideration of an inverse Wishart prior for D where the scale matrix is determined from the ﬂrst-stage variance.

Subfields: Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/c6fd818ac1211b1ecd1693554071cfb9e9b23b27

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

brms: An R Package for Bayesian Multilevel Models Using Stan (2017, background): “However, there are good arguments against the Inverse-Wishart prior (Natarajan and Kass 2000; Kass and Natarajan 2006).”
A Stochastic Variational Framework for Fitting and Diagnosing Generalized Linear Mixed Models (2012, methodology): “Kass and Natarajan (2006) contain a justiﬁcation of ˆ R being a reasonable guess for D in the absence of any other prior knowledge.”
Sensitivity analysis in Bayesian generalized linear mixed models for binary data (2011, background): “Sensitivity to prior assumptions for the random effect precision was discussed in Browne and Draper (2006) and discussion articles following it (Gelman 2006; Kass and Natarajan 2006; Lambert 2006).”
Variational Inference for Generalized Linear Mixed Models Using Partially Noncentered Parametrizations (2012, methodology): “Following the suggestion by Kass and Natarajan (2006), we set ν = r and let the scale matrix S be determined from ﬁrst-stage data variability.”