Multiply robust matching estimators of average and quantile treatment effects

Propensity score matching has been a long-standing tradition for handling confounding in causal inference, however, requiring stringent model assumptions. In this article, we propose novel double score matching (DSM) utilizing both the propensity score and prognostic score. To gain the protection of possible model misspecification, we posit multiple candidate models for each score. We show that the debiasing DSM estimator achieves the multiple robustness property in that it is consistent if any one of the score models is correctly specified. We characterize the asymptotic distribution for the DSM estimator requiring only one correct model specification based on the martingale representations of the matching estimators and theory for local normal experiments. We also provide a two-stage replication method for variance estimation and extend DSM for quantile estimation. Simulation demonstrates DSM outperforms single-score matching and prevailing multiply robust weighting estimators in the presence of extreme propensity scores.     

Performance evaluation of propensity score methods for estimating average treatment effects with multi-level treatments*

The propensity score (PS) method is widely used to estimate the average treatment effect (TE) in observational studies. However, it is generally confined to the binary treatment assignment. In an extension to the settings of a multi-level treatment, Imbens proposed a generalized propensity score which is the conditional probability of receiving a particular level of the treatment given pre-treatment variables. The average TE can then be estimated by conditioning solely on the generalized PS under the assumption of weak unconfoundedness. In the present work, we adopted this approach and conducted extensive simulations to evaluate the performance of several methods using the generalized PS, including subclassification, matching, inverse probability of treatment weighting (IPTW), and covariate adjustment. Compared with other methods, IPTW had the preferred overall performance. We then applied these methods to a retrospective cohort study of 228,876 pregnant women. The impact of the exposure to different types of the antidepressant medications (no exposure, selective serotonin reuptake inhibitor (SSRI) only, non-SSRI only, and both) during pregnancy on several important infant outcomes (birth weight, gestation age, preterm labor, and respiratory distress) were assessed.     

A comparison of doubly robust estimators of the mean with missing data

We consider data with a continuous outcome that is missing at random and a fully observed set of covariates. We compare by simulation a variety of doubly-robust (DR) estimators for estimating the mean of the outcome. An estimator is DR if it is consistent when either the regression model for the mean function or the propensity to respond is correctly specified. Performance of different methods is compared in terms of root mean squared error of the estimates and width and coverage of confidence intervals or posterior credibility intervals in repeated samples. Overall, the DR methods tended to yield better inference than the incorrect model when either the propensity or mean model is correctly specified, but were less successful for small sample sizes, where the asymptotic DR property is less consequential. Two methods tended to outperform the other DR methods: penalized spline of propensity prediction [Little RJA, An H. Robust likelihood-based analysis of multivariate data with missing values. Statist Sinica. 2004;14:949–968] and the robust method proposed in [Cao W, Tsiatis AA, Davidian M. Improving efficiency and robustness of the doubly robust estimator for a population mean with incomplete data. Biometrika. 2009;96:723–734].     

A comparison of mixed and minimax estimators of linear models

In this paper the stochastic properties of two estimators of linear models, mixed and minimax, based on different types of prior information, are compared using quadratic risk as the criterion for superiority. A necessary and sufficient condition for the minimax estimator to be superior to the comparable mixed estimator is derived as well as a simpler necessary but not sufficient condition.     

A comparison of micro versus macro point estimators for markov-process models

We compare, by means of factorially designed Monte Carlo simulation experiments, the performance of (macro-data) restricted least-squares point estimators with that of (micro-data) maximum likelihood estimators for Markov-process models. We find, by various measures of estimator accuracy, that micro data are approximately ten times more valuable than macro data. We also find evidence that a small state space, a long time series, and a large number of entities observed enhance performance.     

A comparison of some robust, adaptive, and partially adaptive estimators of regression models

Numerous estimation techniques for regression models have been proposed. These procedures differ in how sample information is used in the estimation procedure. The efficiency of least squares (OLS) estimators implicity assumes normally distributed residuals and is very sensitive to departures from normality, particularly to "outliers" and thick-tailed distributions. Lead absolute deviation (LAD) estimators are less sensitive to outliers and are optimal for laplace random disturbances, but not for normal errors. This paper reports monte carlo comparisons of OLS,LAD, two robust estimators discussed by huber, three partially adaptiveestimators, newey's generalized method of moments estimator, and an adaptive maximum likelihood estimator based on a normal kernal studied by manski. This paper is the first to compare the relative performance of some adaptive robust estimators (partially adaptive and adaptive procedures) with some common nonadaptive robust estimators. The partially adaptive estimators are based on three flxible parametric distributions for the errors. These include the power exponential (Box-Tiao) and generalized t distributions, as well as a distribution for the errors, which is not necessarily symmetric. The adaptive procedures are "fully iterative" rather than one step estimators. The adaptive estimators have desirable large sample properties, but these properties do not necessarily carry over to the small sample case.

The monte carlo comparisons of the alternative estimators are based on four different specifications for the error distribution: a normal, a mixture of normals (or variance-contaminated normal), a bimodal mixture of normals, and a lognormal. Five hundred samples of 50 are used. The adaptive and partially adaptive estimators perform very well relative to the other estimation procedures considered, and preliminary results suggest that in some important cases they can perform much better than OLS with 50 to 80% reductions in standard errors.     

A comparison of some robust,adaptive, and partially adaptive estimators of regression models

Numerous estimation techniques for regression models have been proposed. These procedures differ in how sample information is used in the estimation procedure. The efficiency of least squares (OLS) estimators implicity assumes normally distributed residuals and is very sensitive to departures from normality, particularly to "outliers" and thick-tailed distributions. Lead absolute deviation (LAD) estimators are less sensitive to outliers and are optimal for laplace random disturbances, but not for normal errors. This paper reports monte carlo comparisons of OLS,LAD, two robust estimators discussed by huber, three partially adaptiveestimators, newey's generalized method of moments estimator, and an adaptive maximum likelihood estimator based on a normal kernal studied by manski. This paper is the first to compare the relative performance of some adaptive robust estimators (partially adaptive and adaptive procedures) with some common nonadaptive robust estimators. The partially adaptive estimators are based on three flxible parametric distributions for the errors. These include the power exponential (Box-Tiao) and generalized t distributions, as well as a distribution for the errors, which is not necessarily symmetric. The adaptive procedures are "fully iterative" rather than one step estimators. The adaptive estimators have desirable large sample properties, but these properties do not necessarily carry over to the small sample case.

The monte carlo comparisons of the alternative estimators are based on four different specifications for the error distribution: a normal, a mixture of normals (or variance-contaminated normal), a bimodal mixture of normals, and a lognormal. Five hundred samples of 50 are used. The adaptive and partially adaptive estimators perform very well relative to the other estimation procedures considered, and preliminary results suggest that in some important cases they can perform much better than OLS with 50 to 80% reductions in standard errors.

     

Using generalized doubly robust estimator to estimate average treatment effects of multiple treatments in observational studies

The generalized doubly robust estimator is proposed for estimating the average treatment effect (ATE) of multiple treatments based on the generalized propensity score (GPS). In medical researches where observational studies are conducted, estimations of ATEs are usually biased since the covariate distributions could be unbalanced among treatments. To overcome this problem, Imbens [The role of the propensity score in estimating dose-response functions, Biometrika 87 (2000), pp. 706–710] and Feng et al. [Generalized propensity score for estimating the average treatment effect of multiple treatments, Stat. Med. (2011), in press. Available at: http://onlinelibrary.wiley.com/doi/10.1002/sim.4168/abstract] proposed weighted estimators that are extensions of a ratio estimator based on GPS to estimate ATEs with multiple treatments. However, the ratio estimator always produces a larger empirical sample variance than the doubly robust estimator, which estimates an ATE between two treatments based on the estimated propensity score (PS). We conduct a simulation study to compare the performance of our proposed estimator with Imbens’ and Feng et al.’s estimators, and simulation results show that our proposed estimator outperforms their estimators in terms of bias, empirical sample variance and mean-squared error of the estimated ATEs.     

A comparison of parametric conditional error-rate estimators for the two-group linear discriminant function

Using mean absolute deviation, we compare the efficay of two new parametric conditional error rate estimators with six others, four of which are well known.The performance of both new estimators is found to be superior to the six competing estimators examined in this paper, especially when the ratio of the training sample size to the feature dimensionality is small.     

A Monte Carlo comparison of GMM and QMLE estimators for short dynamic panel data models with spatial errors

We suggest a generalized spatial system GMM (SGMM) estimation for short dynamic panel data models with spatial errors and fixed effects when n is large and T is fixed (usually small). Monte Carlo studies are conducted to evaluate the finite sample properties with the quasi-maximum likelihood estimation (QMLE). The results show that, QMLE, with a proper approximation for initial observation, performs better than SGMM in general cases. However, it performs poorly when spatial dependence is large. QMLE and SGMM perform better for different parameters when there is unknown heteroscedasticity in the disturbances and the data are highly persistent. Both estimates are not sensitive to the treatment of initial values. Estimation of the spatial autoregressive parameter is generally biased when either the data are highly persistent or spatial dependence is large. Choices of spatial weights matrices and the sign of spatial dependence do affect the performance of the estimates, especially in the case of the heteroscedastic disturbance. We also give empirical guidelines for the model.     

A class of shrinkage estimators in linear regression

We consider the problem of using shrinkage estimators that shrink towards subspaces in linear regression, in particular subspaces spanned by principal components. This is especially important when multicollinearity is present and the number of predictors is not small compared to the sample size. New theoretical results about Stein estimation are used to get estimators with lower theoretical risk than standard Stein estimators used by Oman (1991). Application of the techniques to real data is largely successful.     

A comparison of alternative rate estimators in some frequently used growth models

This paper deals with the linear, exponential and Gompertz models and compares the bias and variance of a number of estimators of the growth-rate parameters of these models. Some of these estimators have been commonly recommended in the literature and others are most frequently used in empirical practice. The present analysis points to the strengths and weaknesses of the estimators in their usage.     

A note on the convergence rates for empirical bayes estimators of parameters in multiple-parameter exponential families

Under suitable conditions upon prior distribution, the convergence rates for empirical Bayes estimators of parameters in multi-parameter exponential families (M-PEF) are obtained. It is shown that the assumptions Tong (1996) imposed on the marginal density can be reduced. The above result can also be extended to more general forms of M-PEF. Finally, some examples which satisfy the conditions of the theorems are given.     

A family of estimators of population variance in two-occasion rotation patterns

Abstract

In this article, we have considered the problem of estimation of population variance on current (second) occasion in two occasion successive (rotation) sampling. A class of estimators of population variance has been proposed and its asymptotic properties have been discussed. The proposed class of estimators is compared with the sample variance estimator when there is no matching from the previous occasion and the Singh et al. (2013) estimator. Optimum replacement policy is discussed. It has been shown that the suggested estimator is more efficient than the Singh et al. (2013) estimator and a usual unbiased estimator when there is no matching. An empirical study is carried out in support of the present study.     

A note on non-negative mean square error estimation of regression estimators in randomized response surveys

Generalized regression estimators are considered for the survey population total of a quantitative sensitive variable based on randomized responses. Formulae are presented for ‘non-negative’ estimators of approximate mean square errors of these biased estimators when population and sample sizes are large.     

A variant on the argument for the invariance of estimators in a singular system of equations

Allecation models, such as consumer demand systems, typically imply a degenerate error structure. The usual approach in estimation is to delete one equation, and to appeal to the results of Barten (1969), for example, that parameter estimates are invariant to the equation deleted. However such proofs of invariance are not straightforward. This paper demonstrates that such systems are observationally equivalent to structures common in the simultaneous equations literature, for which invariance is obvious, and hence provides a more transparent demonstration of conditions for invariance.     

A comparison of estimators for the mean in the inverse gaussian distribution with a known coefficient of variation

In this paper, we discuss an estimation problem of the mean in the inverse Gaussian distribution with a known coefficient of variation. Two types of linear estimators for the mean, the linear minimum variance unbiased estimator and the linear minimum mean squared error estimator, are constructed by using the squared error loss function and their properties are examined. It is observed that, for small samples the performance of the proposed estimators is better than that of the maximum likelihood estimator, when the coefficient of variation is large.     

Developing a restricted two-parameter Liu-type estimator: A comparison of restricted estimators in the binary logistic regression model

In the context of estimating regression coefficients of an ill-conditioned binary logistic regression model, we develop a new biased estimator having two parameters for estimating the regression vector parameter β when it is subjected to lie in the linear subspace restriction Hβ = h. The matrix mean squared error and mean squared error (MSE) functions of these newly defined estimators are derived. Moreover, a method to choose the two parameters is proposed. Then, the performance of the proposed estimator is compared to that of the restricted maximum likelihood estimator and some other existing estimators in the sense of MSE via a Monte Carlo simulation study. According to the simulation results, the performance of the estimators depends on the sample size, number of explanatory variables, and degree of correlation. The superiority region of our proposed estimator is identified based on the biasing parameters, numerically. It is concluded that the new estimator is superior to the others in most of the situations considered and it is recommended to the researchers.     

A power comparison of three tests for design effects in a random effects covariance model

A preliminary testing procedure for design ettecta in a ran-dom effects covariance model is Compared with the usual procedure to see if the power of the latter can be improved. A procedure which ignores the random covariate effects is included for comparison and for study of misspecification effects. Methodology is based on Roebruck's (1982) results for regular linear models.