Show simple item record

Profiling heteroscedasticity in linear regression models

dc.contributor.authorZhou, Qian M.en_US
dc.contributor.authorSong, Peter X.‐k.en_US
dc.contributor.authorThompson, Mary E.en_US
dc.date.accessioned2015-09-01T19:30:42Z
dc.date.available2016-10-10T14:50:23Zen
dc.date.issued2015-09en_US
dc.identifier.citationZhou, Qian M.; Song, Peter X.‐k. ; Thompson, Mary E. (2015). "Profiling heteroscedasticity in linear regression models." Canadian Journal of Statistics 43(3): 358-377.en_US
dc.identifier.issn0319-5724en_US
dc.identifier.issn1708-945Xen_US
dc.identifier.urihttps://hdl.handle.net/2027.42/113149
dc.description.abstractDiagnostics 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 Statistics 43: 358–377; 2015 © 2015 Statistical Society of CanadaRésuméLes outils de diagnostic pour l'hétéroscédasticité dans les modèles de régression linéaire sont largement étudiés dans la littérature. Toutefois, l'identification des covariables associées aux variances hétéroscédastiques n'a suscité que peu d'intérêt. Ce problème joue pourtant un rôle clé pour l'estimation par les moindres carrés pondérés, puisque la modélisation correcte de la structure de variance accroî t l'efficacité de l'estimation. Les auteurs proposent des statistiques spécifiques aux covariables fondées sur un ratio d'information comparant l'estimateur de la variance basé sur le modèle à l'estimateur sandwich de la variance. Ils développent une procédure diagnostique en deux étapes, détectant d'abord l'hétéroscédasticité et identifiant ensuite les covariables dont peut dépendre la structure de variance. Ils généralisent la méthode proposée afin d'accommoder des complications pratiques telles que l'absence de lien entre la structure de la moyenne et une covariable associée avec l'hétéroscédasticité, ou la forte corrélation des covariables. Les auteurs évaluent la performance de la méthode proposée à l'aide d'une étude de simulation et l'illustrent en analysant un jeu de données montrant l'importance d'identifier correctement les covariables associées avec la structure de variance pour l'estimation et l'inférence. La revue canadienne de statistique xx: 1–20; 2015 © 2015 Société statistique du Canadaen_US
dc.publisherWiley Periodicals, Inc.en_US
dc.subject.otherscreeningen_US
dc.subject.otherweighted least squaresen_US
dc.subject.otherMSC 2010: Primary 62J05en_US
dc.subject.othersecondary 62J20en_US
dc.subject.otherHeteroscedasticityen_US
dc.subject.otherhybrid testen_US
dc.subject.otherinformation ratioen_US
dc.subject.otherlinear regression modelsen_US
dc.subject.othermodel‐based estimatorsen_US
dc.subject.othersandwich estimatorsen_US
dc.titleProfiling heteroscedasticity in linear regression modelsen_US
dc.typeArticleen_US
dc.rights.robotsIndexNoFollowen_US
dc.subject.hlbsecondlevelStatistics and Numeric Dataen_US
dc.subject.hlbtoplevelScienceen_US
dc.description.peerreviewedPeer Revieweden_US
dc.description.bitstreamurlhttp://deepblue.lib.umich.edu/bitstream/2027.42/113149/1/cjs11252.pdf
dc.identifier.doi10.1002/cjs.11252en_US
dc.identifier.sourceCanadian Journal of Statisticsen_US
dc.identifier.citedreferenceMüller, H.‐G. & Zhao, P.‐L. ( 1995 ). On a semiparametric variance function model and a test for heteroscedasticity. The Annals of Statistics, 23, 946 – 967.en_US
dc.identifier.citedreferenceDiblasi, A. & Bowman, A. ( 1997 ). Testing for constant variance in a linear model. Statistics and Probability Letters, 33, 95 – 103.en_US
dc.identifier.citedreferenceGlejser, H. ( 1969 ). A new test for heteroskedasticity. Journal of the American Statistical Association, 64, 316 – 323.en_US
dc.identifier.citedreferenceGoldfeld, S. & Quandt, R. ( 1965 ). Some tests for homoscedasticity. Journal of the American Statistical Association, 60, 539 – 547.en_US
dc.identifier.citedreferenceGolub, G. H. & Van Loan, C. F. ( 1996 ). Matrix Computations, 3rd ed., Johns Hopkins University Press, USA.en_US
dc.identifier.citedreferenceGujarati, D. & Porter, D. ( 2009 ). Basic Econometrics, 5th ed., McGraw‐Hill Irwin, Boston.en_US
dc.identifier.citedreferenceHansen, L. ( 1982 ). Large sample properties of generalized method of moments estimators. Econometrica: Journal of the Econometric Society, 50, 1029 – 1054.en_US
dc.identifier.citedreferenceHarvey, A. ( 1976 ). Estimating regression models with multiplicative heteroscedasticity. Econometrica, 44, 461 – 465.en_US
dc.identifier.citedreferenceHeung, M., Wolfgram, D., Kommareddi, M., Hu, Y., Song, & Ojo, A. ( 2012 ). Fluid overload at initiation of renal replacement therapy is associated with lack of renal recovery in patients with acute kidney injury. Nephrology Dialysis Transplantation, 27, 956 – 961.en_US
dc.identifier.citedreferenceKauermann, G. & Carroll, R. J. ( 2001 ). A note on the efficiency ofsandwich covariance matrix estimation. Journal of the American Statistical Association, 96, 1387 – 1396.en_US
dc.identifier.citedreferenceKuk, A. ( 1999 ). Nonparametrically weighted least squares estimation in heteroscedastic linear regression. Biometrical Journal, 41, 401 – 410.en_US
dc.identifier.citedreferenceLin, X. ( 1997 ). Variance component testing in generalised linear models with random effects. Biometrika, 84, 309 – 326.en_US
dc.identifier.citedreferenceLindsay, B. G. ( 1982 ). Conditional score functions: Some optimality results. Biometrika, 69, 503 – 512.en_US
dc.identifier.citedreferenceMak, T. ( 1992 ). Estimation of parameters in heteroscedastic linear models. Journal of the Royal Statistical Society, Series B, 54, 649 – 655.en_US
dc.identifier.citedreferenceMammen, E. ( 1992 ). Bootstrap, wild bootstrap, and asymptotic normality. Probability Theory and Related Fields, 93, 439 – 455.en_US
dc.identifier.citedreferenceMüller, H. & Stadtmuller, U. ( 1987 ). Estimation of heteroscedasticity in regression analysis. The Annals of Statistics, 15, 610 – 625.en_US
dc.identifier.citedreferencePark, R. ( 1996 ). Estimation with heteroscedastic error terms. Econometrica, 34, 888 – 888.en_US
dc.identifier.citedreferencePark, Y. & Wei, L. ( 2003 ). Estimating subject‐specific survival functions under the accelerated failure time model. Biometrika, 9, 717 – 723.en_US
dc.identifier.citedreferencePresnell, B. & Boos, D. D. ( 2004 ). The IOS test for model misspecification. Journal of the American Statistical Association, 99, 216 – 227.en_US
dc.identifier.citedreferenceRuppert, D., Wand, M., Holst, U., & Hössjer, O. ( 1997 ). Local polynomial variance‐function estimation. Technometrics, 39, 262 – 273.en_US
dc.identifier.citedreferenceVerbyla, A. ( 1993 ). Modelling variance heterogeneity: residual maximum likelihood and diagnostics. Journal of the Royal Statistical Society, Series B, 55, 493 – 508.en_US
dc.identifier.citedreferenceWand, M. P. & Jones, M. C. ( 1995 ). Kernel Smoothing, Chapman and Hall, London.en_US
dc.identifier.citedreferenceWhite, H. ( 1980 ). A heteroskedasticity‐consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica, 48, 817 – 838.en_US
dc.identifier.citedreferenceWhite, H. ( 1982 ). Maximum likelihood estimation of misspecified models. Econometrica, 50, 1 – 25.en_US
dc.identifier.citedreferenceWu, C. ( 1986 ). Jackknife, bootstrap and other resampling methods in statistics. The Annals of Statistics, 14, 1261 – 1350.en_US
dc.identifier.citedreferenceZalesny, M. D. & Farace, R. V. ( 1986 ). A field study of social information processing. Human Communication Research, 13, 268 – 290.en_US
dc.identifier.citedreferenceZhou, Q., Song, P. X. K., & Thompson, M. E. ( 2012 ). Information ratio test for model misspecification in quasi‐likelihood inference. Journal of the American Statistical Association, 107, 205 – 213.en_US
dc.identifier.citedreferenceAnscombe, F. ( 1961 ). Examination of residuals. Proceedings of theBerkeley Symposium on Mathematical Statistics and Probability, 1, 1 – 3.en_US
dc.identifier.citedreferenceBickel, P. ( 1978 ). Using residuals robustly I: Tests for heteroscedasticity, nonlinearity. The Annals of Statistics, 6, 266 – 291.en_US
dc.identifier.citedreferenceBreusch, T. & Pagan, A. ( 1979 ). A simple test for heteroscedasticity and random coefficient variation. Econometrica, 47, 1287 – 1294.en_US
dc.identifier.citedreferenceCai, T., Tian, L., & Wei, L. ( 2005 ). Semiparametric Box–Cox power transformation models forcensored survival observations. Biometrika, 92, 619 – 632.en_US
dc.identifier.citedreferenceCarroll, R. & Ruppert, D. ( 1982 ). Robust estimation in heteroscedastic linear models. The Annals of Statistics, 10, 429 – 441.en_US
dc.identifier.citedreferenceCarroll, R. & Ruppert, D. ( 1988 ). Transformation and Weighting in Regression, Chapman & Hall, New York.en_US
dc.identifier.citedreferenceCook, R. & Weisberg, S. ( 1982 ), Residuals and Influence in Regression, Vol. 5, Chapman & Hall, New York.en_US
dc.identifier.citedreferenceCook, R. & Weisberg, S. ( 1983 ). Diagnostics for heteroscedasticity in regression. Biometrika, 70, 1 – 10.en_US
dc.owningcollnameInterdisciplinary and Peer-Reviewed


Files in this item

Show simple item record

Remediation of Harmful Language

The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.

Accessibility

If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.