Add functionality to compare coefficiants

[mattansb]

See https://rdrr.io/cran/psych/man/r.test.html

Would be nice to have an easystats version of this.

Here are some mockups:

Independent correlations

Interface 1: pass a grouped data frames

mtcars |> 
  dplyr::group_by(am) |> 
  compare_correlations(select = c("mpg", "hp", "drat"))
#> # Differences between independant Pearson correlations
#> 
#> Group | Parameter1 | Parameter2 |    r1 |    r2 |  diff |         95% CI |     z |       p
#> -------------------------------------------------------------------------------------------
#> 0 - 1 |        mpg |         hp | -0.83 | -0.13 | -0.71 | [-0.93, -0.61] | -6.17 | < .001***
#> 0 - 1 |        mpg |       drat |  0.47 |  0.50 | -0.03 | [ 0.02,  0.76] |  2.19 | 0.086    
#> 0 - 1 |         hp |       drat | -0.34 | -0.01 | -0.34 | [-0.69,  0.13] | -1.50 | 0.151    
#> 

Interface 2: name variables

compare_correlations(mtcars, select = c("mpg", "hp", "drat"), group = "am")

Dependent correlations

Not sure...

Effect size

Differences between correlations are already pretty standardized... but there is also the difference between Fished z-transformed correlations, which is called Cohen's q.

[DominiqueMakowski]

that'd be very nice indeed!

in terms of API, perhaps we could include an argument by_group to correlation(), which would take a column or a vector?