Setting function to Cohen's d or Hedges' g

[Santihago]

Hi,

In a paired comparison, is it possible to set the function (func =) to obtain the Cohen's d or Hedges g effect size, as it is on the website version?

On the website version, I can choose between mean, median, Cohen's d and Hedges' g. On R, I tried with func = mean and func = median, and both work. But what argument should I input to have the Cohen's d effect size, as in the website?

Thank you very much.

paired_mean_diff <- data %>% dabest(., x, y, idx = c("Control", "Group"), paired = TRUE, id.column = id, ci = 95, func = mean)

[josesho]

Hi @Santihago ,

Thanks for using our package!

Unfortunately standardized effect sizes have not been implemented in dabestr yet. We do plan on doing so but we don't have a deadline yet!

An aside: the func argument in dabest() applies that function to both the control and test groups, and then takes the difference. So even if we wrote a custom Hedges' g function, for example, it can't be passed along like this. (I have ideas on how to tweak the current API, stay tuned!)

The Python version of dabestr features Cohen's d and Hedges' g; if you are comfortable with Python you could have a look at that.

Regards,
Joses

[Santihago]

Thank you @josesho, I will have a look at the Python version. Good luck with the development and thank you for the great package!