A Clumped Beta-Binomial Model for the Analysis of Clustered Attribute Data

Publisher Website
Google Scholar
Abstract
Clusters of subjects [mice] form experimental units to which different treatments are applied, and the response of each subject is binary. The clusters are categorized by frequency of positive response, but information on some categories is clumped. For analysis of data in such experiments, a beta-binomial model is considered formulated and tested for the hypothesis of no interaction between treatments and cluster sizes. The model fits data presented by Kastenbaum and Lamphiear (1959) and the analysis in this case confirms the absence of interaction.

This publication has 1 reference indexed in Scilit: