bbolker/broom.mixed

robust argument in tidyMCMC

Opened this issue · 1 comments

Hello,

In tidyMCMC function, help says :

robust : use mean and standard deviation (if FALSE) or median and mean absolute deviation (if TRUE) to compute point estimates and uncertainty?

While function's code does the opposite :

m <- if (robust)
colMeans(ss)
else apply(ss, 2, median)
stdfun <- if (robust)
stats::mad
else stats::sd_

Thank you

oops! (it's just m that's backward, stdfun() is OK - right?)