I would like to cluster standard errors at subject level in a nonlinear least square regression in R. Is it possible to do this with the nls command?
Asked
Active
Viewed 146 times
2
Andrew
- 678
- 2
- 9
- 19
-
Please make a reproducible example https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example. Show what you've tried and where you are stuck. – emilliman5 Oct 01 '17 at 21:15
-
if you have just one regrssion model, you'd have *one single* standard error for the mean. You cannot cluster that. -- Are you talking about the residuals? Then check the `residuals()` function and the `qqplot()` functions. -- If you have many models, you can use the `broom` package, notably its `tidy()` and `glance()` functions to collect std.errors and other model metrics. – knb Oct 02 '17 at 07:39