Adjusting for Nonignorable Drop-Out Using Semiparametric Nonresponse Models

Scharfstein, Daniel O.; Rotnitzky, Andrea; Robins, James M.

doi:10.2307/2669923

Cited by 377 publications

(447 citation statements)

References 0 publications

Supporting

Mentioning

438

Contrasting

Unclassified

Order By: Relevance

“…A remarkable advantage is that the method is asymptotically efficient, when either the parametric regression model m (x, β) or the propensity score model p(x, α) is correctly specified. This is termed the double-robustness (DR) property by Scharfstein et al (1999), and it has been extensively used with semiparametric inference. Thereafter, the DR property has attracted much discussion, for example, Robins and Rotnitzky (2001), Carpenter et al (2006), Kang and Schafer (2007), and Qin et al (2008).…”

Section: Introductionmentioning

confidence: 99%

A comparison study of nonparametric imputation methods

Ning

Cheng

2010

Stat Comput

View full text Add to dashboard Cite

Consider estimation of a population mean of a response variable when the observations are missing at random with respect to the covariate. Two common approaches to imputing the missing values are the nonparametric regression weighting method and the Horvitz-Thompson (HT) inverse weighting approach. The regression approach includes the kernel regression imputation and the nearest neighbor imputation. The HT approach, employing inverse kernelestimated weights, includes the basic estimator, the ratio estimator and the estimator using inverse kernel-weighted residuals. Asymptotic normality of the nearest neighbor imputation estimators is derived and compared to kernel regression imputation estimator under standard regularity conditions of the regression function and the missing pattern function. A comprehensive simulation study shows that the basic HT estimator is most sensitive to discontinuity in the missing data patterns, and the nearest neighbors estimators can be insensitive to missing data patterns unbalanced with respect to the distribution of the covariate. Empirical studies show that the nearest neighbor imputation method is most effective among these imputation methods for estimating a finite population mean and for classifying the species of the iris flower data.

show abstract

Section: Introductionmentioning

confidence: 99%

A comparison study of nonparametric imputation methods

Ning

Cheng

2010

Stat Comput

View full text Add to dashboard Cite

show abstract

“…The magnitude of the bias caused by the misspecification depends on various factors, such as the number of missing covariates in z i , their strength in explaining the dependence between C i andT i , the true hazard rate model, and the censoring rate. Generally, the model misspecification problem in inverse probability weighting can be weakened through the doubly robust estimation procedure (Scharfstein et al 1999;Robins and Rotnitzky 2005). The doubly robust estimation combines weighting and regression, and this estimation still achieves consistency when the regression part is specified correctly even if the weighting part is misspecified.…”

Section: Implementation Issuesmentioning

confidence: 99%

Gini index estimation for lifetime data

Zhang

Guang-yu

2016

Lifetime Data Anal

View full text Add to dashboard Cite

Lifetime data is often right-censored. Recent literature on the Gini index estimation with censored data focuses on independent censoring. However, the censoring mechanism is likely to be dependent censoring in practice. This paper proposes two estimators of the Gini index under independent censoring and covariate-dependent censoring, respectively. The proposed estimators are consistent and asymptotically normal. We also evaluate the performance of our estimators in finite samples through Monte Carlo simulations. Finally, the proposed methods are applied to real data.

show abstract

“…Their method is known as the doubly robust estimation method. This is because their estimation method requires the correct specification of either the missing mechanism or the distribution of the missing covariate, but not both (see also Scharfstein, Rotnitzky, and Robins 1999). More concretely, the purpose of their method is to obtain a consistent estimator of the parameter of distribution of the dependent variables y, given the fully observed covariate x and the partially observed covariate r, p(y|x, r).…”

Section: Introductionmentioning

confidence: 97%