ABSTRACT

Tang et al. (2003. Analysis of multivariate missing data with nonignorable nonresponse. Biometrika, 90(4), 747–764) and Zhao & Shao (2015. Semiparametric pseudo-likelihoods in generalized linear models with nonignorable missing data. Journal of the American Statistical Association, 110(512), 1577–1590) proposed a pseudo likelihood approach to estimate unknown parameters in a parametric density of a response Y conditioned on a vector of covariate X, where Y is subjected to nonignorable nonersponse, X is always observed, and the propensity of whether or not Y is observed conditioned on Y and X is completely unspecified. To identify parameters, Zhao & Shao (2015. Semiparametric pseudo-likelihoods in generalized linear models with nonignorable missing data. Journal of the American Statistical Association, 110(512), 1577–1590) assumed that X can be decomposed into U and Z, where Z can be excluded from the propensity but is related with Y even conditioned on U. The pseudo likelihood involves the estimation of the joint density of U and Z. When this density is estimated nonparametrically, in this paper we apply sufficient dimension reduction to reduce the dimension of U for efficient estimation. Consistency and asymptotic normality of the proposed estimators are established. Simulation results are presented to study the finite sample performance of the proposed estimators.

References

1. Baker, S. G., & Laird, N. M. (1988). Regression analysis for categorical variables with outcome subject to nonignorable nonresponse. Journal of the American Statistical association, 83(401), 62–69. doi: 10.1080/01621459.1988.10478565 [Taylor & Francis Online], [Google Scholar]
2. Bierens, H. J. (1987). Kernel estimators of regression functions. In Advances in econometrics: Fifth world congress (Vol. 1, pp. 99–144). [Google Scholar]
3. Cook, R. D., & Weisberg, S. (1991). Comment. Journal of the American Statistical Association, 86(414), 328–332. [Taylor & Francis Online], [Google Scholar]
4. Greenlees, J. S., Reece, W. S., & Zieschang, K. D. (1982). Imputation of missing values when the probability of response depends on the variable being imputed. Journal of the American Statistical Association, 77(378), 251–261. doi: 10.1080/01621459.1982.10477793 [Taylor & Francis Online], [Google Scholar]
5. Hansen, B. E. (2008). Uniform convergence rates for kernel estimation with dependent data. Econometric Theory, 24(3), 726–748. doi: 10.1017/S0266466608080304 [Web of Science ®], [Google Scholar]
6. Li, K.-C. (1991). Sliced inverse regression for dimension reduction. Journal of the American Statistical Association, 86(414), 316–327. doi: 10.1080/01621459.1991.10475035 [Taylor & Francis Online], [Google Scholar]
7. Li, B., & Wang, S. (2007). On directional regression for dimension reduction. Journal of the American Statistical Association, 102(479), 997–1008. doi: 10.1198/016214507000000536 [Taylor & Francis Online], [Google Scholar]
8. Little, R. J., & Rubin, D. B. (2014). Statistical analysis with missing data (Vol. 333). New York: John Wiley & Sons. [Google Scholar]
9. Ma, Y., & Zhu, L. (2012). A semiparametric approach to dimension reduction. Journal of the American Statistical Association, 107(497), 168–179. doi: 10.1080/01621459.2011.646925 [Taylor & Francis Online], [Google Scholar]
10. Nadaraya, E. A. (1964). On estimating regression. Theory of Probability & Its Applications, 9(1), 141–142. doi: 10.1137/1109020 [Google Scholar]
11. Newey, W. K., & McFadden, D. (1994). Large sample estimation and hypothesis testing. Handbook of Econometrics, 4, 2111–2245. doi: 10.1016/S1573-4412(05)80005-4 [Google Scholar]
12. Qin, J., Leung, D., & Shao, J. (2002). Estimation with survey data under nonignorable nonresponse or informative sampling. Journal of the American Statistical Association, 97(457), 193–200. doi: 10.1198/016214502753479338 [Taylor & Francis Online], [Google Scholar]
13. Robins, J., & Ritov, Y. (1997). Toward a curse of dimensionality appropriate (coda) asymptotic theory for semi-parametric models. Statistics in Medicine, 16(3), 285–319. doi: 10.1002/(SICI)1097-0258(19970215)16:3<285::AID-SIM535>3.0.CO;2-# [Google Scholar]
14. Silverman, B. W. (1986). Density estimation for statistics and data analysis (Vol. 26). New York: CRC press. [Google Scholar]
15. Tang, G., Little, R. J., & Raghunathan, T. E. (2003). Analysis of multivariate missing data with nonignorable nonresponse. Biometrika, 90(4), 747–764. doi: 10.1093/biomet/90.4.747 [Web of Science ®], [Google Scholar]
16. Wang, S., Shao, J., & Kim, J. K. (2014). An instrumental variable approach for identification and estimation with nonignorable nonresponse. Statistica Sinica, 24(3), 1097–1116. [Google Scholar]
17. Watson, G. S. (1964). Smooth regression analysis. Sankhyā: The Indian Journal of Statistics, Series A, 26(4), 359–372. [Google Scholar]
18. Xia, Y., Tong, H., Li, W., & Zhu, L.-X. (2002). An adaptive estimation of dimension reduction space. Journal of the Royal Statistical Society: Series B (Statistical Methodology), 64(3), 363–410. doi: 10.1111/1467-9868.03411 [Google Scholar]
19. Zhao, J., & Shao, J. (2015). Semiparametric pseudo-likelihoods in generalized linear models with nonignorable missing data. Journal of the American Statistical Association, 110(512), 1577–1590. doi: 10.1080/01621459.2014.983234 [Taylor & Francis Online], [Google Scholar]

Zhiping Qiu, Huijuan Ma, Jianhua Shi. (2020) Reweighting estimators for the transformation models with length-biased sampling data and missing covariates. Communications in Statistics - Theory and Methods 0:0, pages 1-24.