You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Most tidyverse packages do not really support weights and there is probably a good reason for that. Still, weights can be important when using survey data.
I am not an expert at package writing, but since no standard errors are needed this could be done by multiplying the observations with an integer weight internally before calculating the correlations. Granted, this is quite easily done in the tidyverse (e.g. tidyr::uncount), but I still think it would improve usability if this option was built in?
All the best!
The text was updated successfully, but these errors were encountered:
Would it be possible to add a "weights" option?
Most tidyverse packages do not really support weights and there is probably a good reason for that. Still, weights can be important when using survey data.
I am not an expert at package writing, but since no standard errors are needed this could be done by multiplying the observations with an integer weight internally before calculating the correlations. Granted, this is quite easily done in the tidyverse (e.g. tidyr::uncount), but I still think it would improve usability if this option was built in?
All the best!
The text was updated successfully, but these errors were encountered: