[Statistical issue; proposal] Changing the SEs to CIs or SDs about ggcorset HOT 2 CLOSED

Dear @Generalized ,

Thank you for your kind comment and suggestion. I appreciate you taking the time to write in-depth reasoning for the use of SD or CIs instead of SE. I agree that SE is often used incorrectly to infer significant differences between means, and this is not the intent of use in the 'ggcorset' package (nor do I recommend its use in this manner). Rather, the analyst should explore these group differences themselves, applying the appropriate test (if desired).

With that said, I think there is merit in keeping SE to better understand the "precision" of the sample mean relative to sample size (se = sd / $\sqrt{n}$ ). I also think that offering an alternative 'eyelet type' (one of SD) would be best for statistical integrity. Specifically, I think by offering both SD and SE (in a clear manner), will help to combat against uses as you mentioned in points 1 and 2.

I appreciate your suggestion to provide CIs by using a flexible approach (and I appreciate your creative solution to providing this flexibility). At this time, it is beyond the scope of the package. I will, however, keep it in mind for future versions if the demand/ need is there. In the meantime, since the 'eyelet' feature is optional, user-defined CIs could be added to the visualization as both 'ggcorset' functions return a 'ggplot2' object.

Again, thank you for bringing attention to this. I will work to add the SDs to the GitHub version of the package, and will close this comment once this is completed.

from ggcorset.

This enhancement has been made by adding an e_type argument to both gg_corset() and gg_corset_elongated() functions. To visualize 1 standard deviation above and below the mean instead of standard error means, the argument e_type = "SD" needs to be specified. Further documentation has been added to the README of this repository.

from ggcorset.