Split Samples and Design Sensitivity in Observational Studies

By Ruth Heller, Paul R. Rosenbaum, Dylan S. Small. 2009

88 citations recorded by the source corpus.

Subfields: Management Science and Operations Research, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/4c0b7ee1eab142403c5e18e2a321bef81babdeac

How researchers see this paper

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

In Management Science and Operations Research

In Statistics and Probability

Citations in the corpus

On Sensitivity Value of Pair-Matched Observational Studies (2017, methodology): “Based on this observation, Heller et al. (2009) proposed a sample splitting procedure to screen out uninteresting outcomes and gain power; see also Zhao et al.”
Evaluation of subset matching methods and forms of covariate balance (2016, methodology): “For instance, this knowledge could be obtained by splitting the sample into a small planification sample to learn features of the outcome model (for instance, by using LASSO regression with the original covariates and transformations of them) and a larger analysis sample to conduct the actual matchi”
Does data splitting improve prediction? (2013, methodology): “Some have used data splitting for quite different reasons — see Heller et al. (2009).”
Modeling heterogeneous treatment effects in large-scale experiments using Bayesian Additive Regression Trees (2010, background): “Heller, Rosenbaum, and Small (2009) randomly split the data from an observational study into a smaller planning sample and a larger analysis sample, using the planning sample to determine tuning parameters and the analysis sample for actual estimation.”
Matching methods for causal inference: A review and a look forward (2010, background): “Third, matching methods have straightforward diagnostics by which their performance can be assessed.”
Effect Modification and Design Sensitivity in Observational Studies (2013, methodology): “When the sample size I is large, an alternative to the treeprocedure in Section 4 is sample splitting, as discussed by Heller, Rosenbaum, and Small (2009).”
Valid post-selection inference (2013, methodology): “We ordered the covariates by priority, year of birth being first because of the structure of the study, followed by gestational age in weeks (0, 33], (33, 36], (36, 38], (38, 40] and (40,∞), categories based on an estimated propensity score for the propensity to be born in Philadelphia, mother’s age”