Issue with simulations data - non r-class program may not set r()

[Barc]

I have written a program which does not output my values correctly

program Datageneration, rclass
matrix my = (2\2\2)
matrix cov = (3,1.25,0.75\1.25,1,0.6\0.75,0.6,1)
drawnorm x1 x2 x3, n(200) means(my) cov(cov)
gen y=x1
regress y x1
return scalar b1=_b[x]
end

which returns "non r-class program may not set r()".

Where in my programming have I gone wrong?

[Maarten Buis]

Works fine for me:

. clear all

.
. program Datageneration, rclass
  1. matrix my = (2\2\2)
  2. matrix cov = (3,1.25,0.75\1.25,1,0.6\0.75,0.6,1)
  3. drawnorm x1 x2 x3, n(200) means(my) cov(cov)
  4. gen y=x1
  5. regress y x1
  6. return scalar b1=_b[x]
  7. end

.
. Datageneration
(obs 200)

      Source |       SS           df       MS      Number of obs   =       200
-------------+----------------------------------   F(1, 198)       =         .
       Model |   563.58089         1   563.58089   Prob > F        =         .
    Residual |           0       198           0   R-squared       =    1.0000
-------------+----------------------------------   Adj R-squared   =    1.0000
       Total |   563.58089       199  2.83206478   Root MSE        =         0

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          x1 |          1          .        .       .            .           .
       _cons |          0  (omitted)
------------------------------------------------------------------------------

.
end of do-file

. return list

scalars:
                 r(b1) =  1

[Barc]

Hmm. Thats weird. When I run the data generating process then it retuns this:
Be aware that I have set a seed as to easier compare results.

. clear all

. set seed 1

.
. matrix my = (2\2\2)

. matrix cov = (3,1.25,0.75\1.25,1,0.6\0.75,0.6,1)

. drawnorm x1 x2 x3, n(200) means(my) cov(cov)
(obs 200)

. gen y=x1

. regress y x1

      Source |       SS           df       MS      Number of obs   =       200
-------------+----------------------------------   F(1, 198)       =         .
       Model |  576.942035         1  576.942035   Prob > F        =         .
    Residual |           0       198           0   R-squared       =    1.0000
-------------+----------------------------------   Adj R-squared   =    1.0000
       Total |  576.942035       199  2.89920621   Root MSE        =         0

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          x1 |          1          .        .       .            .           .
       _cons |  -4.44e-16          .        .       .            .           .
------------------------------------------------------------------------------

. return scalar b1=_b[x]
non r-class program may not set r()


. return list

scalars:
              r(level) =  95

matrices:
              r(table) :  9 x 2

.



end of do-file

[FernandoRios]

Hi Barc
The difference between what you did and what Marteen did is that he run the code from the dofile editor , all at once.
You instead are trying to do it line by line. So the line, return scalar, is not reunning within an rclass program.
Fernando

[Maarten Buis]

you did not include the lines:

program Datageneration, rclass

and

end

in your example. Now the error message makes sense: If you are not in a program, than that non-existing program obviously cannot return anything.

[Barc]

Hey Fernando and Maarten

I feel very stupid right now, as I have been sitting with this for a couple of hours... Thank you very much both, and have a great day.