Add function for chi square.

[elinw]

This function follows the pattern in the documentation, and I found it useful to create it for my students. It might be good to let it make a caption the way that the means_by_groups() has the ANOVA results.

[strengejacke]

You may take a look at the effectsize package where you can calculate lots of effect sizes from different R objects.

[strengejacke]

Ok, I see the intention of this function. Maybe it's better to implement a chisq.test.dwtable() method instead? @etiennebacher wdyt?

[etiennebacher]

Personally I don't really see the need for this function, it seems to simply replace a call to as.table() and lapply()? I get that it makes it easier for students but we also can't implement a function for all possible computations one could do on a table.

library(datawizard)

### Without data_chisq()
suppressWarnings({
  orig <- head(mtcars) |> 
    data_tabulate() |> 
    as.table() |> 
    lapply(stats::chisq.test)
})

### With data_chisq()
suppressWarnings({
  new <- head(mtcars) |> 
    data_tabulate() |> 
    data_chisq()
})

identical(orig, new)
#> [1] TRUE

Maybe it's better to implement a chisq.test.dwtable() method instead?

I would be in favor of adding this as a method, but chisq.test() isn't generic:

sloop::is_s3_generic("chisq.test")
#> [1] FALSE