Dear all,

I am trying to use the Heckman correlation test to investigate the possibility of a selection bias in my data. I have panel data - when I run the command "xtheckman" using the code
Code:
 xtheckman wages training_hrs i.high_qual, select (training_hrs i.high_qual i.illness_disability i.sex i.children i.general_health i.region i.age i.sector)
an error message stating "invalid specification of select(); only one selection variable is allowed r(198)" comes up. But when I run the same code using "heckman" command, then I get a nice table with results.

My question is: why can I not include more than one variable under "xtheckman" test - what if I need all of these variables? Also, would it be wrong to run the code under a simple "heckman" command even if my data is panel data?

Thank you very much in advance!