ps.covariate.residualplot: Covariate Residuals Plot

In OHDSI/Centaur: Centaur Propensity Score Balancing Workflow and Toolkit

Description

A graphical balance diagnostic examining the ratio of residuals orthogonal to the propensity scores

Usage

ps.covariate.residualplot(data, covariates, weights = NULL,
  dichotomous = FALSE)

Arguments

data	Data Frame, containing the dataset. The data frame must contain a treatment indicator variable called 'treat'.
covariates	Vector, containing the list of covariates to include in the plot
weights	Vector, containing the weights to use in assessment of balanced population. Defaults to unweighted
dichotomous	Boolean value indicating if dichotomous covariates should be included in the calculation or not

Details

This function creates a plot of the residuals orthogonal to the propensity scores. Each covariate is regressed on the propensity scores for the control and treatment populations. The variance of the residuals of this regression are then compared. For a balanced population, the ratio of variances should be close to 1. An imbalance in the variances indicates that there is additional variance potentially attributed to the covariate in one of the populations that has not been captured in the propensity scores.

Examples

## Not run: 
ps.covariate.residualplot(myData, covariates)

## End(Not run)

OHDSI/Centaur documentation built on May 7, 2019, 8:22 p.m.