multilevelGPSStratification.Rd
Stratification on GPS with multilevel treatments
multilevelGPSStratification(Y, W, X, NS, GPSM = "multinomiallogisticReg", linearp = 0, nboot)
Y | A continuous response vector (1 x n) |
---|---|
W | A treatment vector (1 x n) with numerical values indicating treatment groups |
X | A covariate matrix (p x n) with no intercept |
NS | The number of strata: (only required in the function
|
GPSM | An indicator of the methods used for estimating GPS, options include "multinomiallogisticReg", "ordinallogisticReg", and "existing" |
linearp | An indicator of subclassification on GPS (=0) or linear
predictor of GPS (=1): (only required in the function
|
nboot | The number of boot replicates for variance estimation: (only
required in the function |
A list with two elements,
tauestimate
, varestimate
, where tauestimate
is a
vector of estimates for pairwise treatment effects, and varestimate
is a vector of variance estimates, using bootstrapping method.
Yang, S., Imbens G. W., Cui, Z., Faries, D. E., & Kadziola, Z. (2016) Propensity Score Matching and Subclassification in Observational Studies with Multi-Level Treatments. Biometrics, 72, 1055-1065. https://doi.org/10.1111/biom.12505
Abadie, A., & Imbens, G. W. (2006). Large sample properties of matching estimators for average treatment effects. Econometrica, 74(1), 235-267. https://doi.org/10.1111/j.1468-0262.2006.00655.x
Abadie, A., & Imbens, G. W. (2016). Matching on the estimated propensity score. Econometrica, 84(2), 781-807. https://doi.org/10.3982/ECTA11293
simulated_data <- multilevelMatching::simulated_data set.seed(123) multilevelGPSStratification( Y = simulated_data$outcome , W = simulated_data$treatment, X = simulated_data[ ,names(simulated_data) %in% paste0("covar", 1:6)], GPSM = "multinomiallogisticReg", NS = 5, linearp = TRUE, nboot = 10 )#> $tauestimate #> EY(2)-EY(1) EY(3)-EY(1) EY(3)-EY(2) #> -0.3038424 0.4201045 0.7239469 #> #> $varestimate #> EY(2)-EY(1) EY(3)-EY(1) EY(3)-EY(2) #> 0.1658798 0.2881434 0.6232490 #>