[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: some notes on the usage of -estout-
Offlist, Clive Nicholas pointed out to me that -estout-
> does not appear to properly present two clean columns of
> regression output, as it should!
My reply was:
short answer: add "modelwidth(10)" to the command
long answer: My first intention was to program a utility
to save estimates to a spreadsheet (i.e. a tab-delimited
text file) which can be easily imported to a program like
excel. (...) However, such tab-delimited files are not
displayed nicely by text editors/viewers like the stata
results window or stata viewer because the columns have
no fixed width (this is called "free format").
I understand that one is temted to use estout as a substitute
to -estimates table- because it can do much more. The problem
is, that estout has no real display routine - it just replays
the produced tab-delimited file.
Because I wanted -estout- to be a general program I added
support for other kinds of output files like e.g. TeX, which
is basically just a matter of specifying different delimiters.
Now, raw TeX tabulars are really messy and unpleasant to work
with if they are in free format. Thus, I added the possibility
to fix the variable stub and the model colums to a certain
As a side effect, these options (-varwidth-, -modelwidth-, and
-delimiter-) open the backdoor to use -estout- as a substitute
of -estimates table-. If modelwidth and varwidth are specified,
a fixed format file is produced (you have to make sure, though,
that modelwidth is large enough; also it is advisable to get
rid of the tab-character expansion by either specifying
-showtabs- or changing the delimiter to, e.g., a blank:
-delimiter(" ")-). Such files look all right in the stata
* For searches and help try: