Let outcome.name take a null value so that create.formula() can create a one-sided formula

[BorgeJorge]

Title says it all!

[dachosen1]

@BorgeJorge one sided formulas are typically not used in functions like lm or glm, which expects both a predictor and and a outcome variables. How would this be useful.

[BorgeJorge]

One-sided formulas are used in xtabs(), cor.test(), nlme(), reStruct(), pairs(), plot.lme(), model.matrix(), snm(), matchit(), weightit(), and several other functions. It's common enough that the stats package provides a function to create a one-sided formula (asOneSidedFormula()), cobalt::f.build() will also create one, and is.one.sided() will test for one.

[dachosen1]

@BorgeJorge this seems reasonable we'll work on designing a solution. Thanks