Identification and Estimation of Generalized Additive Partial Linear Models with Nonignorable Missing Response

Abstract

The generalized additive partial linear models (GAPLM) have been widely used for flexible modeling of various types of response. In practice, missing data usually occurs in studies of economics, medicine, and public health. We address the problem of identifying and estimating GAPLM when the response variable is nonignorably missing. Three types of monotone missing data mechanism are assumed, including logistic model, probit model and complementary log-log model. In this situation, likelihood based on observed data may not be identifiable. In this article, we show that the parameters of interest are identifiable under very mild conditions, and then construct the estimators of the unknown parameters and unknown functions based on a likelihood-based approach by expanding the unknown functions as a linear combination of polynomial spline functions. We establish asymptotic normality for the estimators of the parametric components. Simulation studies demonstrate that the proposed inference procedure performs well in many settings. We apply the proposed method to the household income dataset from the Chinese Household Income Project Survey 2013.

Keywords

Generalized additive partial linear models / Nonignorable missingness / Identifiability / Observed likelihood

References

[1.] Baccini M, Biggeri A, Lagazio C, Lertxundi A, Saez M. Parametric and semi-parametric approaches in the analysis of short-term effects of air pollution on health. Comput. Stat. Data Anal., 2007, 51: 4324-4336,

[2.] Brehm JO. . The Phantom Respondents: Opinion Surveys and Political Representation, 2009 Ann Arbor University of Michigan Press

[3.] Cameron AC, Trivedi PK. . Microeconometrics: Methods and Applications, 2005 Cambridge Cambridge University Press,

[4.] Carroll R, Fan J, Gijbels I, Wand MP. Generalized partially linear single-index models. J. Am. Stat. Assoc., 1997, 438: 477-489,

[5.] Chen J, Shao J, Fang F. Instrument search in pseudo-likelihood approach for nonignorable nonresponse. Ann. Inst. Stat. Math., 2021, 73: 519-533,

[6.] Cui X, Guo J, Yang G. On the identifiability and estimation of generalized linear models with parametric nonignorable missing data mechanism. Comput. Stat. Data Anal., 2017, 107: 64-80,

[7.] De Boor, C.: A Practical Guide to Splines, revised ed. Applied Mathematical Sciences, vol. 27. Springer, New York (2001)

[8.] DeVore, R.A., Lorentz, G.G.: Constructive Approximation: Polynomials and Splines Approximation (1993)

[9.] Fan, J., Gijbels, I., Hu, T.-C., Huang, L.-S.: A study of variable bandwidth selection for local polynomial regression. Stat. Sin. 113–127 (1996)

[10.] Fang C. Growth and structural changes in employment in transitional China. Econ. Res. J., 2007, 7: 4-14

[11.] Fang, F., Shao, J.: Model selection with nonignorable nonresponse. Biometrika asw039 (2016)

[12.] Gao W, Smyth R. Education expansion and returns to schooling in urban china, 2001–2010: evidence from three waves of the china urban labor survey. J. Asia Pac. Econ., 2015, 20: 178-201,

[13.] Greenlees JS, Reece WS, Zieschang KD. Imputation of missing values when the probability of response depends on the variable being imputed. J. Am. Stat. Assoc., 1982, 77: 251-261,

[14.] Härdle W, Sperlich S, Spokoiny V. Structural tests in additive regression. J. Am. Stat. Assoc., 2001, 96: 1333-1347,

[15.] Härdle WK, Müller M, Sperlich S, Werwatz A. . Nonparametric and Semiparametric Models, 2004 Berlin Springer,

[16.] He X, Fung WK, Zhu Z. Robust estimation in generalized partial linear models for clustered data. J. Am. Stat. Assoc., 2005, 100: 1176-1184,

[17.] Ibrahim JG, Chen M-H, Lipsitz SR. Missing responses in generalised linear mixed models when the missing data mechanism is nonignorable. Biometrika, 2001, 88: 551-564,

[18.] Kang L, Peng F. Real wage cyclicality in urban China. Econ. Lett., 2012, 115: 141-143,

[19.] Kim JK, Yu CL. A semiparametric estimation of mean functionals with nonignorable missing data. J. Am. Stat. Assoc., 2011, 106: 157-165,

[20.] Krosnick, J.A.: The causes of no-opinion responses to attitude measures in surveys: they are rarely what they appear to be. Surv. Nonresponse 87–100 (2002)

[21.] Li W, Yang L. Spline-backfitted kernel smoothing of nonlinear additive autoregression model. Ann. Stat., 2007, 35: 2474-2503

[22.] McCullagh P, Nelder JA. . Generalized Linear Models, 1989 Boca Raton CRC Press,

[23.] Miao W, Ding P, Geng Z. Identifiability of normal and normal mixture models with nonignorable missing data. J. Am. Stat. Assoc., 2016, 111: 1673-1683,

[24.] Nelder JA, Wedderburn RW. Generalized linear models. J. R. Stat. Soc. Ser. A (Gen.), 1972, 135: 370-384,

[25.] Pollard D. Asymptotics for least absolute deviation regression estimators. Economet. Theor., 1991, 7: 186-199,

[26.] Qin J, Leung D, Shao J. Estimation with survey data under nonignorable nonresponse or informative sampling. J. Am. Stat. Assoc., 2002, 97: 193-200,

[27.] Rubin DB. Inference and missing data. Biometrika, 1976, 63: 581-592,

[28.] Sasieni, P.: Generalized additive models. T. J. Hastie and R. J. Tibshirani, Chapman and Hall, London, 1990. no. of pages: xv + 335. price: £25. ISBN: 0-412-34390-8. Stat. Med. 11, 981–982 (1992)

[29.] Sicular, T., Li, S., Yue, X., Sato, H.: Changing Trends in China’s Inequality: Evidence, Analysis, and Prospects (2020)

[30.] Stone CJ. Additive regression and other nonparametric models. Ann. Stat., 1985, 13: 689-705,

[31.] Tang G, Little RJ, Raghunathan TE. Analysis of multivariate missing data with nonignorable nonresponse. Biometrika, 2003, 90: 747-764,

[32.] Tang N, Ju Y. Statistical inference for nonignorable missing-data problems: a selective review. Stat. Theory Relat. Fields, 2018, 2: 105-133,

[33.] Tang N, Zhao P, Zhu H. Empirical likelihood for estimating equations with nonignorably missing data. Stat. Sin., 2014, 24: 723, pmcid: 4071774

[34.] Wang L, Shao J, Fang F. Propensity model selection with nonignorable nonresponse and instrument variable. Stat. Sin., 2021, 31: 647-672

[35.] Wang, L., Yang, L.: Spline single-index prediction model. arXiv preprint arXiv:0704.0302 (2007)

[36.] Wang, S., Shao, J., Kim, J.K.: An instrumental variable approach for identification and estimation with nonignorable nonresponse. Stat. Sin. 1097–1116 (2014)

[37.] Wood SN. On confidence intervals for generalized additive models based on penalized regression splines. Aust. N. Z. J. Stat., 2006, 48: 445-464,

[38.] Xue L, Yang L. Additive coefficient modeling via polynomial spline. Stat. Sin., 2006, 16: 1423-1446

[39.] Zhao J, Shao J. Semiparametric pseudo-likelihoods in generalized linear models with nonignorable missing data. J. Am. Stat. Assoc., 2015, 110: 1577-1590,

[40.] Zhao P, Wang L, Shao J. Sufficient dimension reduction and instrument search for data with nonignorable nonresponse. Bernoulli, 2021, 27: 930-945,

Funding

national natural science foundation of china(11731015)