Abstract
We develop a non-standard linear regression analysis by considering that the dependent variable is left censored and also that some of the explanatory variables are measured with additive errors. Our censored measurement error regression model is specified by assuming heavy-tailed distributions for the underlying probabilistic process. Specifically, we focus on assuming a multivariate \(t\) joint distribution for the error terms and the unobserved true covariates. For the model estimation, we consider the maximum likelihood methodology in which we include the estimation of the asymptotic variance of the maximum likelihood estimators. We also develop an EM algorithm to obtain the estimates. The performance of the newly developed methodology is evaluated throughout a simulation study as well as a case study analysis.
Similar content being viewed by others
References
Amemiya T (1985) Advanced econometrics. Harvard University Press, Cambridge
Arellano-Valle RB, Bolfarine H (1995) On some characterizations of the \(t\) distribution. Stat Probab Lett 25:179–185
Arellano-Valle RA, Bolfarine H (1996) Elliptical structural models. Commun Stat Theory Methods 25:2319–2341
Arellano-Valle RB, Ozan S, Bolfarine H, Lachos VH (2005) Skew normal measurement error models. J Multivar Anal 96:265–281
Arellano-Valle RB, Castro LM, González-Farias G, Muñoz-Gajardo KA (2012) Student-\(t\) censored regression model: properties and inference. Stat Methods Appl 21:453–473
Azzalini A (2014) The skew-normal and related families. Cambridge University Press (Monograph), New York
Bolfarine H, Arellano-Valle RB (1998) Weak nondifferential erros models. Stat Probab Lett 40:279–287
Cameron AC, Trivedi PK (2010) Microeconometrics using stata, revised edn. Stata Press, College Station
Dempster AP, Laird NM, Rubin DB (1977) Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc Ser B (Methodological) 39(1):1–38
Fang KT, Kotz S, Ng KW (1990) Symmetric multivariate and related distributions. Chapman and Hall, London
Fuller WA (1987) Measurement error models. Wiley, New York
Gleser LJ (1992) The importance of assessing measurement reliability in multivariate regression. J Am Stat Assoc 87:696–707
Lin TI, Ho HJ, Chen CL (2009) Analysis of multivariate skew normal models with incomplete data. J Multivar Anal 100:2337–2351
Lin TI, Lin TC (2011) Robust statistical modelling using the multivariate skew \(t\) distribution with complete and incomplete data. Stat Model 11:253–277
Lucas A (1997) Robustness of the student \(t\) based \(M\)-estimator. Commun Stat Theory Methods 26(5):1165–1182
Maddala GS (1983) Limited-dependent and qualitative variables in econometrics. Cambridge University Press, New York
Marchenko YV, Genton MG (2012) A Heckman selection-t model. J Am Stat Assoc 107:304–317
Matos LA, Prates MO, Chen MH, Lachos VH (2013) Likelihood-based inference for mixed-effects models with censored response using the multivariate-\(t\) distribution. Stat Sinica 23:1323–1342
Meng XL, Rubin DB (1993) Maximum likelihood estimation via the ECM algorithm: a general framework. Biometrika 80:267–278
Polasek W, Krause A (1993) Bayesian regression model with simple errors in variables structure. J R Stat Soc Ser D (The Statistician) 42:571–580
Tobin J (1958) Estimation of relationships for limited dependent variables. Econometrica 26:24–36
Wang L (1998) Estimation of censored linear error-in-variables models. J Econ 84:383–400
Wang L (2002) A simple adjustement for measurement erros in some limited dependent variable. Stat Probab Lett 58:427–433
Acknowledgments
We would like to express our gratitude to the editors and two anonymous referees. Their constructive criticisms and suggestions contributed definitively to the improvement of the paper. The research of G. H. M. A. Rocha was partially supported by CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) of Brazil. The research of R. B. Arellano-Valle was partially supported by Grants FONDECYT 1120121 from Chilean government. R. H. Loschi would like to thank CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) of the Ministry for Science and Technology of Brazil, grants 301393/2013-3 and 306085/2009-7 for a partial allowance to her researches.
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Rocha, G.H.M., Arellano-Valle, R.B. & Loschi, R.H. Maximum likelihood methods in a robust censored errors-in-variables model. TEST 24, 857–877 (2015). https://doi.org/10.1007/s11749-015-0439-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11749-015-0439-1
Keywords
- Censored regression
- Multivariate \(t\) distribution
- Scale mixtures of normal distributions
- Expectation-maximization algorithm