weights not allowed in rwolf

[Enrico Miglino]

Hi there,

I am trying to use the command rwolf to adjust the p-values of my regressions for multiple hypothesis testing.
My regressions are weighted with inverse probability weights to account for attrition. The help section of the command rwolf says that it is possible to include weights: the sintax is rwolf depvars [if] [in] [weight], indepvar(varname) [options].

I tried to include the weights, typing the following command: rwolf Y1 Y2 Y3 [pweight=myweights], indepvar(treatment) method(regress) seed(1234) controls(X1 X2) cluster(cluster_var)

When I run the command, Stata comes up with the following:
"weights not allowed
r(101)."

When I delete [pweight=myweights], everything runs well.

Can somebody explain me why I cannot use the weights?

Thank you in advance.

[Nick Cox]

rwolf is from SSC, as you are asked to explain (FAQ Advice #12).

I'd

set traced 1 
set trace on

to see where the error occurs.

[daniel klein]

rwolf calls bootstrap, which does not allow weights for command. See

help bootstrap

This is a "technical" explanation not a theoretical/conceptual one. I have no idea what rwolf is supposed to do and how it does it, so I cannot further advice.

Best
Daniel

[Enrico Miglino]

Thank you all,

yes the problem was that bootstrap does not allow weights.

For whoever may be interested, I found an experimental version of rwolf that allows for pweight at https://github.com/damiancclarke/rwo...cda8/rwolf.ado

Best,
Enrico