Missing Data, Imputation, and the Bootstrap

By Bradley Efron. 1994

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Multiple Imputation After 18+ Years (1996, methodology): “Replicates-Another Alternative One way to try to get around these inefficiencies is to try to do first-moment proper (multiple) imputation in K non-independent samples, i.e., jackknife or bootstrap replicates (e.g., Burns 1990; Efron 1994).”
Assessing Mediational Models: Testing and Interval Estimation for Indirect Effects (2010, background): “Bootstrapping approaches (2) are relatively straightforward to extend to missing data if computationally intensive (Efron, 1994).”
Maximum Likelihood Multiple Imputation: Faster Imputations and Consistent Standard Errors Without Posterior Draws (2012, background): “An old knock against bootstrapped MI was that it seemed to require a large number of imputations D for each bootstrapped sample [10].”
Improving the Synthetic Data Generation Process in Spatial Microsimulation Models (2009, methodology): “We note that alternative methods for the creation of population microdatasets include a wide variety of statistical data-imputation methods (see Efron, 1994).”
Inference for imputation estimators (2000, methodology): “It is of some interest to compare this with the nonparametric bootstrap estimate of the standard error of β̂sex (Efron & Tibshirani, 1993, Ch. 6; Efron, 1994), since, like our standard error estimator, the bootstrap estimator is consistent for the asymptotic standard error of n1/2(β̂sex − β∗sex)…”
Bootstrap Inference When Using Multiple Imputation (2016, methodology): “The use of the bootstrap in the context of missing data has often been viewed as a frequentist alternative to multiple imputation [12], or an option to obtain conﬁdence intervals after single imputation [13].”
Local multiple imputation (2002, methodology): “The advantage of the method is that it generates samples with exactly the same range of X values as in the original sample, avoiding samples which might only poorly re(cid:176)ect the regression structure, the latter which is essential in Step 2.”
A Review of Hot Deck Imputation for Survey Non‐response (2010, methodology): “A resampling method that allows for estimation of smooth or non-smooth statistics is the bootstrap (Efron, 1994), and its application to the hot deck was discussed by Shao & Sitter (1996) and Saigo et al. (2001).”
The Dynamic Effect of Innovation on Market Structure (2004, background): “First, nonstationarity is an important issue.4 The usual approach to addressing nonstationarity is to filter the data in the hope of making the series mean stationary and covariance stationary.”
International handbook of survey methodology. (2008, methodology): “A more broadly applicable but computationally-intensive approach is to use a replication technique such as balanced repeated replication, the jackknife, or the bootstrap for variance estimation, with the imputation procedure repeated separately for each replicate; see for example, Efron (1994) and S”
Impact of the Bootstrap on Sample Surveys (2003, methodology): “For simple random sampling, Efron (1994) showed that the standard bootstrap procedure together with reimputing bootstrap datasets produces a valid approximation to the distribution of θ̂I .”
Amelia II: A Program for Missing Data (2011, methodology): “While creative applications of bootstrapping have been developed for several application-specific missing data problems (Rubin and Schenker 1986; Rubin 1994; Efron 1994; Shao and Sitter 1996; Lahlrl 2003), but to our knowledge the technique has not been used to develop and implement a general purpos”
A Repeated HalfSample Bootstrap and Balanced Repeated Replications for Randomly Imputed Data (2001, methodology): “There are two methods available for handling randomly imputed data for both smooth and nonsmooth estimators: the adjusted balanced repeated replication (BRR) methods proposed by Shao, Chen and Chen (1998); and the bootstrap method proposed by Shao and Sitter (1996) (see also Efron 1994)”