stdDiffDotPlot: Plot the absolute standardised differences of variables used...

In jgodet/utilitR: Lots of useful functions

Plot the absolute standardised differences of variables used to generate the propensity score

Description

Plot (dotchart) the absolute standardised differences of variables used to generate the propensity score. Standardised differences are calculated using stddiff package. Standardized differences are used to compare the balance in measured baseline covariates. Standardized difference in excess of 10

Usage

stdDiffDotPlot(
  dataSMD,
  data,
  treat = "treat",
  showRes = TRUE,
  plotGraph = TRUE,
  lateX = FALSE,
  inset = 0.02,
  xlim = c(0, 0.5),
  col,
  ...
)

Arguments

dataSMD	dataframe of SMD results
data	dataframe
treat	name of the treatment variable
showRes	boolean if true print result table
plotGraph	boolean if true show graph in the output
lateX	boolean if true print result table in lateX format
inset	inset for legend position relative to the bottom right
xlim	xlim see ?par
col	colors

Value

data.frame

Author(s)

JuG

Examples

#synthetic data
set.seed(2020)
treat<-round(abs(rnorm(100)+1)*10,0) %% 2
numeric1<-round((rnorm(100)+1)*10,0)
numeric2 <- round(abs(rnorm(100)+1)*10,0)
binary<-factor(round(abs(rnorm(100)+1)*10,0) %% 2)
binary2<-factor(round(abs(rnorm(100)+1)*10,0) %% 2)
category<-factor(round(abs(rnorm(100)+1)*10,0) %% 3)
category2<-factor(round(abs(rnorm(100)+1)*10,0) %% 5)
data<-data.frame(treat,numeric1, numeric2,binary, binary2,category, category2)


stdDiffDotPlot(data = data, treat = "treat", showRes = TRUE, plotGraph = FALSE)
stdDiffDotPlot(data = data, treat = "treat", showRes = FALSE, plotGraph = TRUE, xlim=c(-.5,1), pch=21, bg="lightgrey", col=c("blue","black"))
#res <- stdDiffDotPlot(data = data, treat = "treat", showRes = TRUE, plotGraph = FALSE)
#segments(x0=res$stddiff.l, x1=res$stddiff.u, y0=1:6)


data2 <- data.frame(age=c(rnorm(2929, 69.6, 13.5), rnorm(6178, 65, 13.3)), ttt=c(rep("NobetaBlock",2929), rep("betaBlock", 6178)))
stdDiffDotPlot(data = data2, treat = "ttt", showRes = TRUE, plotGraph = FALSE)

example 2 : with dataSMD
dataSMD = data.frame(unmatched = c(.4,.2,.25,.3),
   matched = c(.1,.1,.05,.02),
   weighted = c(.02,.05,.08,.03))
rownames(dataSMD) <- c("Age", "Height", "Weight", "ParamX")

stdDiffDotPlot(dataSMD=dataSMD,col=c("black", "blue", "purple") )

jgodet/utilitR documentation built on April 13, 2025, 11:16 p.m.