Profiling heteroscedasticity in linear regression models
Authors:
- 15 July 2015
- journal article
- research article
- Published by Wiley in Canadian Journal of Statistics
- Vol. 43 (3) , 358-377
Abstract
Diagnostics for heteroscedasticity in linear regression models have been intensively investigated in the literature. However, limited attention has been paid on how to identify covariates associated with heteroscedastic error variances. This problem is critical in correctly modelling the variance structure in weighted least squares estimation, which leads to improved estimation efficiency. We propose covariate‐specific statistics based on information ratios formed as comparisons between the model‐based and sandwich variance estimators. A two‐step diagnostic procedure is established, first to detect heteroscedasticity in error variances, and then to identify covariates the error variance structure might depend on. This proposed method is generalized to accommodate practical complications, such as when covariates associated with the heteroscedastic variances might not be associated with the mean structure of the response variable, or when strong correlation is present amongst covariates. The performance of the proposed method is assessed via a simulation study and is illustrated through a data analysis in which we show the importance of correct identification of covariates associated with the variance structure in estimation and inference.The Canadian Journal of Statistics43: 358–377; 2015 © 2015 Statistical Society of CanadaKeywords
Funding Information
- Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada
- National Science Foundation
This publication has 31 references indexed in Scilit:
- Semiparametric Box–Cox power transformation models for censored survival observationsBiometrika, 2005
- Testing for constant variance in a linear modelStatistics & Probability Letters, 1997
- Transformation and Weighting in Regression1988
- Diagnostics for heteroscedasticity in regressionBiometrika, 1983
- Large Sample Properties of Generalized Method of Moments EstimatorsEconometrica, 1982
- Robust Estimation in Heteroscedastic Linear ModelsThe Annals of Statistics, 1982
- A Simple Test for Heteroscedasticity and Random Coefficient VariationEconometrica, 1979
- Using Residuals Robustly I: Tests for Heteroscedasticity, NonlinearityThe Annals of Statistics, 1978
- A New Test for HeteroskedasticityJournal of the American Statistical Association, 1969
- Some Tests for HomoscedasticityJournal of the American Statistical Association, 1965