Missingness in explanatory variables requires a model for the covariates even if the interest lies only in a model for the outcomes given the covariates. An incorrect specification of the models for the covariates or for the missingness mechanism may lead to biased inferences for the parameters of interest. Previously published articles either use semi-/non-parametric flexible distributions for the covariates and identify the model via a missing at random assumption, or employ parametric distributions for the covariates and allow a more general non-random missingness mechanism. We consider the analysis of binary responses, combining a missing not at random mechanism with a non-parametric model based on a Dirichlet process mixture for the continuous covariates. We illustrate the proposal with simulations and the analysis of a dataset.
AntoniakCE (1974) Mixtures of Dirichlet processes with applications to Bayesian nonparametric problems. The Annals of Statistics, 2, 1152–74.
2.
BlackwellDMacQueenJB (1973) Ferguson distributions via polya urn schemes. The Annals of Statistics, 1, 353–55.
3.
ChenHY (2002) Double-semiparametric method for missing covariates in Cox regression models. Journal of the American Statistical Association, 97, 565–75.
4.
ChenHY (2004) Nonparametric and semi-parametric models for missing covariates in parametric regression. Journal of the American Statistical Association, 99, 1176–89.
5.
ChenHY (2009) Estimation and inference based on neumann series approximation to locally efficient score in missing data problems. Scandinavian Journal of Statistics, 36, 713–34.
6.
ChenHYLittleRJA (1999) Proportional hazards regression with missing covariates. Journal of the American Statistical Association, 94, 896–908.
7.
CongdonP (2006) Bayesian Statistical Modelling, 2nd ed. New York: Wiley.
8.
EscobarMDWestM (1995) Bayesian density estimation and inference using mixtures. Journal of the American Statistical Association, 90, 577–88.
9.
EscobarMDWestM (1998) Computing nonparametric hierarchical models. In DeyDMüllerPSinhaD eds. Practical nonparametric and semiparametric Bayesian statistics (Lecture notes in statististics 133). New York: Springer, 1–22.
10.
FergusonTS (1973) A Bayesian analysis of some nonparametric problems. The Annals of Statistics, 1, 209–30.
11.
GelmanARubinDB (1992) A single series from the gibbs sampler provides a false sense of security. In BernardoJMBergerJODawidAPSmithAFM eds. Bayesian Statistics 4. Oxford: Oxford University Press, 625–31.
12.
GewekeJ (1992) Evaluating the accuracy of sampling-based approaches to the calculation of posterior moments (with discussion). In BernardoJMBergerJODawidAPSmithAFM eds. Bayesian Statistics 4. Oxford: Oxford University Press, 169–93.
13.
GibbonsLEHosmerDW (1991) Conditional logistic regression with missing data. Communications in Statistics—Simulation and Computation, 20, 109–20.
14.
GriffinJHolmesC (2010) Computational issues arising in Bayesian nonparametric hierarchical models. In HjortNLHolmesCMüllerPWalkerSG eds. Bayesian Nonparametrics. Cambridge: Cambridge University Press, 208–22.
15.
HeidelbergerPWelchPD (1983) Simulation run length control in the presence of an initial transient. Operations Research, 31, 1109–44.
16.
HortonNJLairdNM (1999) Maximum likelihood analysis of generalized linear models with missing covariates. Statistical Methods in Medical Research, 8, 37–50.
17.
HortonNJLairdNM (2001) Maximum likelihood analysis of logistic regression models with incomplete covariate data and auxiliary information. Biometrics, 57, 34–42.
18.
HuangLChenM-HIbrahimJG (2005) Bayesian analysis for generalized linear models with nonignorably missing covariates. Biometrics, 61, 767–80.
19.
IbrahimJG (1990) Incomplete data in generalized linear models. Journal of the American Statistical Association, 85, 765–69.
20.
IbrahimJGChenM-HLipsitzSRHerringAH (2005) Missing-data methods for generalized linear models: a comparative review. Journal of the American Statistical Association, 100, 332–46.
21.
IbrahimJGLipsitzSRChenM-H (1999) Missing covariates in generalized linear models when the missing data mechanism is non-ignorable. Journal of the Royal Statistical Society. Series B: Statistical Methodology, 61, 173–90.
22.
IshwaranHJamesLF (2001) Gibbs sampling methods for stick-breaking priors. Journal of the American Statistical Association, 96, 161–73.
23.
IshwaranHJamesLF (2002) Approximate Dirichlet process computing finite normal mixtures: smoothing and prior information. Journal of Computational and Graphical Statistics, 11, 508–32.
24.
LeonardT (1996) On exchangeable sampling distributions for uncontrolled data. Statistics & Probability Letters, 26, 1–6.
25.
LipsitzSRIbrahimJG (1996) A conditional model for incomplete covariates in parametric regression models. Biometrika, 83, 916–22.
26.
LipsitzSRIbrahimJGChenM-HPetersonH (1999) Non-ignorable missing covariates in generalized linear models. Statistics in Medicine, 18, 2435–48.
27.
LipsitzSRParzenMEwellM (1998) Inference using conditional logistic regression with missing covariates. Biometrics, 54, 295–303.
28.
LittleRJARubinDB (2002) Statistical analysis with missing data, 2nd ed. New York: John Wiley & Sons.
29.
LunnDJSpiegelhalterDThomasABestN (2009) The BUGS project: evolution, critique and future directions (with discussion). Statistics in Medicine, 28, 3049–82.
30.
LunnDJThomasABestNSpiegelhalterD (2000) WinBUGS—a Bayesian modelling framework: concepts, structure, and extensibility. Statistics and Computing, 10, 325–37.
31.
MirandaARabe-HeskethS (2010) Missing ordinal covariates with informative selection. Tech. rep., NCRM Working Paper. Department of Quantitative Social Science, Institute of Education, University of London.
32.
MolenberghsGKenwardMG (2007) Missing data in clinical studies. New York: Wiley.
33.
MüllerPQuintanaFA (2004) Nonparametric Bayesian data analysis. Statistical Science, 19, 95–110.
34.
PaddockSM (2002) Bayesian nonparametric multiple imputation of partially observed data with ignorable nonresponse. Biometrika, 89, 529–38.
35.
PaikMC (2004) Nonignorable missingness in matched case-control data analyses. Biometrics, 60, 306–14.
36.
PlummerM (2003) JAGS: A program for analysis of Bayesian graphical models using Gibbs sampling. In Proceedings of the 3rd International Workshop on Distributed Statistical Computing (DSC 2003), 20–22.
37.
PoletoFZPaulinoCDMolenberghsGSingerJM (2011) Inferential implications of over-parameterization: a case study in incomplete categorical data. International Statistical Review, 79, 92–113.
38.
RafteryAELewisSM (1992) How many iterations in the Gibbs sampler? In BernardoJMBergerJODawidAPSmithAFM eds, Bayesian Statistics 4. Oxford: Oxford University Press, 763–73.
39.
RaghunathanTELepkowskiJMVan HoewykJSolenbergerP (2001) A multivariate technique for multiply imputing missing values using a sequence of regression models. Survey Methodology, 27, 85–95.
40.
RobinsJMRotnitzkyAZhaoLP (1994) Estimation of regression coefficients when some regressors are not always observed. Journal of the American Statistical Association, 89, 846–66.
41.
RubinDB (1987) Multiple imputation for nonresponse in surveys. New York: Wiley.
42.
SattenGACarrollRJ (2000) Conditional and unconditional categorical regression models with missing covariates. Biometrics, 56, 384–88.
43.
ScharfsteinDODanielsMJRobinsJM (2003) Incorporating prior beliefs about selection bias into the analysis of randomized trials with missing outcomes. Biostatistics, 4, 495–512.
44.
ScharfsteinDOIrizarryRA (2003) Generalized additive selection models for the analysis of studies with potentially nonignorable missing outcome data. Biometrics, 59, 601–13.
45.
ScottDW (1992) Multivariate density estimation: theory, practice, and visualization. New York: Wiley.
46.
SethuramanJ (1994) A constructive definition of Dirichlet priors. Statistica Sinica, 4, 639–50.
47.
StubbendickALIbrahimJG (2003) Maximum likelihood methods for nonignorable missing responses and covariates in random effects models. Biometrics, 59, 1140–50.
48.
StubbendickALIbrahimJG (2006) Likelihood-based inference with nonignorable missing responses and covariates in models for discrete longitudinal data. Statistica Sinica, 16, 1143–67.
49.
VachW (1997) Some issues in estimating the effect of prognostic factors from incomplete covariate data. Statistics in Medicine, 16, 57–72.
50.
VachWBlettnerM (1995) Logistic regression with incompletely observed categorical covariates—investigating the sensitivity against violation of the missing at random assumption. Statistics in Medicine, 14, 1315–29.
51.
VachWSchumacherM (1993) Logistic regression with incompletely observed categorical covariates: a comparison of three approaches. Biometrika, 80, 353–62.
52.
VansteelandtSGoetghebeurEKenwardMGMolenberghsG (2006) Ignorance and uncertainty regions as inferential tools in a sensitivity analysis. Statistica Sinica, 16, 953–79.
53.
WalkerSGDamienPLaudPWSmithAFM (1999) Bayesian nonparametric inference for random distributions and related functions (with discussion). Journal of the Royal Statistical Society. Series B: Statistical Methodology, 61, 485–27.
54.
WestM (1992) Modelling with mixtures (with discussion). In BernardoJMBergerJODawidAPSmithAFM eds, Bayesian Statistics 4. Oxford: Oxford University Press, 503–24.
55.
WickiJPernegerTVJunodAFBounameauxHPerrierA (2001) Assessing clinical probability of pulmonary embolism in the emergency ward. Archives of Internal Medicine, 161, 92–97.
56.
ZhangZRocketteHE (2005) On maximum likelihood estimation in parametric regression with missing covariates. Journal of Statistical Planning and Inference, 134, 206–23.
57.
ZhangZRocketteHE (2006) Semiparametric maximum likelihood for missing covariates in parametric regression. Annals of the Institute of Statistical Mathematics, 58, 687–706.
58.
ZhangZRocketteHE (2007) An EM algorithm for regression analysis with incomplete covariate information. Journal of Statistical Computation and Simulation, 77, 163–73.
59.
ZhaoY (2009) Regression analysis with covariates missing at random: a piece-wise nonparametric model for missing covariates. Communications in Statistics—Theory and Methods, 38, 3736–44.
