Premium
Heteroscedastic Models and an Application to Block Designs
Author(s) -
Stirling W. Douglas
Publication year - 1985
Publication title -
journal of the royal statistical society: series c (applied statistics)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.205
H-Index - 72
eISSN - 1467-9876
pISSN - 0035-9254
DOI - 10.2307/2347882
Subject(s) - heteroscedasticity , block (permutation group theory) , computer science , econometrics , mathematics , statistics , combinatorics
SUMMARY Models are considered that explain the heteroscedasticity of a response variable as an explicit function either of the response mean or of other explanatory variables. Maximum likelihood parameter estimation assuming normality of the response is shown to be possible with a sequence of weighted least squares calculations and this is illustrated with the analysis of a standard data set. Finally all normal linear models with a random block effect are shown to be transformable into fixed effect heteroscedastic models and this is used to find maximum likelihood estimates in an unbalanced block design.