Roger, thank you so much! This is really great. I've been able to add
asterisks for significance and put regression coefficients & SEs, R2, N, and
lincom results for multiple regressions into a single data file.
One final question:
Is there an equivalent to -lincomest- for F-tests, i.e. for -test-?
There is no equivalent to -lincomest- for F-tests, because an F-test
creates no estimates and standard errors (only F-statistics and P-values),
and therefore cannot be wrapped up in an estimation-class command like
-lincomest-. However, -test- (like -lrtest-) is a rclass command, and saves
the P-value in a scalar called r(p) (see -[R] test- and -[R] lrtest-). You
can easily access this value, and put it into a data set with only the
P-value variable -p-, and -dsconcat- that data set with the -parmest-
outputs. Or you might add r(p) to a -parmest- output data set as an extra
variable. There are many possibilities. For further help on handling
returned results, see on-line help for -return-, or the hardcopy manuals
under -[R] saved results- or -[P] return-. A lot of helpful advice about
saving results in Stata data sets can be found in the hardcopy programming
manual under -[P] postfile-.
