*
* Example 4.9 from pp 150-151
*
open data bwght.raw
data(format=free,org=columns) 1 1388 faminc cigtax cigprice bwght fatheduc motheduc $
 parity male white cigs lbwght bwghtlbs packs lfaminc
*
* The LINREG will automatically drop the data points which have missing values
* for motheduc and fatheduc.
*
linreg bwght
# constant cigs parity faminc motheduc fatheduc
exclude
# motheduc fatheduc
*
* However, if we want to run the restricted regression, we'll leave out the two
* variables which have the missing variables, so the automatic sample selection
* will give us the full sample, rather than the limited one. The SMPL option can
* be used to select data points based upon some criterion, in this case, the
* points where both motheduc and fatheduc are "valid", that is, not missing.
*
set smpl = %valid(motheduc).and.%valid(fatheduc)
linreg(smpl=smpl) bwght
# constant cigs parity faminc