How to get controls selected for the variables of interest using lasso for inference

[Charlotte Doidge]

I run a lasso for inference model using double selection:

dsregress depvar i.variable_of_interest1 i.variable_of_interest2, controls($cont i.($cat)) rseed(12345)

And store the estimates:
estimates store ds_plugin

I can look at the controls selected for the dependent variable:
lassocoef(ds_plugin, for (dapvar))

What I would like to do is look at the controls selected for the variables of interest - so at the moment I am writing:
lassocoef(ds_plugin, for (variable_of_interest1))

However I get an error with this:

can not find result for variable_of_interest1
r(111)

I am inputting the same code as in the lasso manual so am unsure why it is not working for me. If anyone could explain that would be really helpful

[Andrew Musau]

You use factor variable notation to include the full set of categories less one. So with k levels of var

i.var

results in k-1 variables. So your syntax

lassocoef(ds_plugin, for (variable_of_interest1))

makes no sense as lassocoef expects one variable. In any case, you can view how each level is named using

 di e(varsofinterest)

[Charlotte Doidge]

Thank you for explaining