Hierarchical Linear Models and Cell Data

Abstract

Hierarchical linear models are a generalization of Bayesian linear regression. They differ from Bayesian regression in that they determine the width of the prior distribution for the regression coefficients automatically from the data. They are particularly appropriate when we want to answer questions about the typical weights in a regression instead of just the typical examples. This paper provides a tutorial on hierarchical linear models, then demonstrates their application to some biochemical data.