I would like to build my own balance table table using datasummary(). The reason is that I would like to get rid of some columns and add a second version of the table with the balanced sample of the table. However I am puzzled with how I can define my own function that calculates the difference in means for each group of a factor so that I can get something like this:
library(modelsummary)
library(estimatr)
data("ToothGrowth")
datasummary_balance(~ supp, data = ToothGrowth)
datasummary(Mean ~ supp, data = ToothGrowth)
| | OJ (N=30) / Mean | OJ (N=30) / Std. Dev. | VC (N=30) / Mean | VC (N=30) / Std. Dev. | Diff. in Means | Std. Error |
|:-----|-----------------:|----------------------:|-----------------:|----------------------:|---------------:|-----------:|
| len | 20.7 | 6.6 | 17.0 | 8.3 | -3.7 | 1.9 |
| dose | 1.2 | 0.6 | 1.2 | 0.6 | 0.0 | 0.2 |
But when using
datasummary(Mean ~ supp, data = ToothGrowth, output = "markdown")
I get
| | OJ | VC |
|:-----|---:|---:|
| Mean | | |