Estimating treatment effect using movestay command

[Laurine Mthawali]

• I would like to estimate the treatment effect on crop yields using endogenous switching regression for farmers who participated in Farmer
  Field Schools and those who did not. However, whenever I execute the movestay-command, I get a notification that some variables are
  naturally coded and omitted. Some variables are also not found yet they are in the data set.
  My command is:
  global Yields Yeild_per_acre i.Sexhhd agehhd Agehhd2 i.eduhhd i.edusps i.District landsize hhsize Improved_seed Hired_labor
  global select_eqn FFS_membership i.Sexhhd agehh Agehhd2 i.eduhhd i.edusps hhsize i.Credit_org i.District crop_bags_bf
  xi:movestay ($Yields), select($select_eqn)
  
  Please help me overcome this problem.
  How do I also use the heckam syntax to estimate the effect please.
  
  Thanks
  
  Laurine

[Stephen Jenkins]

Laurine: welcome to the Forum. You do not provide sufficient information for readers to provide helpful answers. Please read the FAQ (hit the black bar at the top of the page), and note well all the tips there. In particular, please do (I) show us exactly what you typed into Stata and exactly what you got back, while (ii) posting this information to the list using CODE delimiters to facilitate legibility (see example below); and (iii) tell us where you got user-written programs (like movestay) from. And, of course, (iv) provide more informative detail about the nature of your substantive problem.

I presume you are referring to the program accompanying a Stata Journal article, with updated code thus:

SJ-5-3  st0071_2  . . . . . . . . . . . . . . . . Software update for movestay
        (help movestay if installed)  . . . . . . . . M. Lokshin and Z. Sajaia
        Q3/05   SJ 5(3):471
        minor bug fix and new option added for movestay;
        modifications made to mspredict

I found this with search movestay

Does movestay accept factor variable notation? If so, your xi: prefix is redundant.
You also say:

How do I also use the heckam syntax to estimate the effect please.

I guess you mean heckman, It's unclear why you're asking about this because that fits a differently specified selection model.