The prognostic analogue of the propensity score

By Ben B. Hansen. 2008

400 citations recorded by the source corpus.

Subfields: Economics and Econometrics, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/6249e633fa1038728d6a363120baaa5420e4fbba

How researchers see this paper

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

In Economics and Econometrics

In Statistics and Probability

Citations in the corpus

The Augmented Synthetic Control Method (2018, methodology): “Using the Basque and California calibrated simulations, Figure 5 shows the empirical distribution for the p-values from uniform permutation tests, punif as defined in Appendix A.5, with a placebo gap test statistic, ∣∣∣Wi(Y obs i − Ỹi)∣∣∣ , where Ỹi is the leave-one-out synthetic control for Yi, ∑ j”
MALTS: Matching After Learning to Stretch (2018, methodology): “Further, matching on prognostic score in Hansen (2008), on L2 distances or Mahalanobis norm to adjust for confounding produces consistent estimates of average treatment effects when these are smooth distance metrics.”
Adaptive Hyper-box Matching for Interpretable Individualized Treatment Effect Estimation (2020, methodology): “Other matching methods, such as propensity score matching (Rosenbaum and Rubin 1983) or prognostic score matching (Hansen 2008) are more automated in that they only require the user to select a model class, and may yield better estimates of average treatment effects.”
A Bayesian view of doubly robust causal inference (2016, background): “If the selected set of features Si captures all relevant prognostic information, then Y0i ⊥ Xi |Si (Hansen, 2008).”
Bayesian regression tree models for causal inference: regularization, confounding, and heterogeneous effects (2017, positive): “2) Under this model, μ(x) = E(Y | Z = 0, X = x) is a prognostic score in the sense of Hansen (2008), another interpretable quantity, to which we apply a prior distribution independent of τ (as detailed below).”
Matching methods for causal inference: A review and a look forward (2010, background): “A more recently developed distance measure is the “prognosis score” (Hansen, 2008).”
Overlap in observational studies with high-dimensional covariates (2017, background): “There exists some function r(X1:p) satisfying the following two conditions (Y (0), Y (1)) ⊥ X1:p | r(X1:p), (17) η ≤ er(X1:p) := P (T = 1 | r(X1:p)) ≤ 1− η. (18) Modifying Hansen (2008)’s nomenclature slightly, we call r(X1:p) a prognostic score.”
A Survey on Causal Inference (2020, background): “distance metrics are learned by utilizing both the covariates and the outcome information so that the transformed space can preserve more information. One representative metric is the prognosis score [49], which is the estimated control outcome. The transformation function is represented as: f (x)= ”
Adjusting for Confounding with Text Matching (2020, methodology): “4 We match censored bloggers to uncensored bloggers who write almost identical blog posts and have very similar censorship histories and find that censored bloggers are much more likely to be censored after censorship, (3)Various solutions to these three problems have been proposed: penalized regres”
Causal Inference without Balance Checking: Coarsened Exact Matching (2012, background): “Although (8) uses the outcome variable, it only does so for control units (as in Hansen, 2008), and so inducing selection bias is not a risk.”
Estimating Treatment Effects using Multiple Surrogates: The Role of the Surrogate Score and the Surrogate Index (2016, background): “These conditional means are related to what Hansen (2008) calls the prognostic score, although in the setting Hansen considers there is no surrogate variable, and”
Assessing lack of common support in causal inference using bayesian nonparametrics: Implications for evaluating the effect of breastfeeding on children's cognitive outcomes (2013, methodology): “Another option that is close in spirit to the propensity score techniques but makes use of outcome data (at least in the control group) would be a prognostic score approach [Hansen (2008)].”
Balance diagnostics after propensity score matching. (2019, methodology): “Then that fitted model is used to predict outcome for all subjects (13).”