# st: RE: Any ways to specify the regression output to 2 decimal places?

 From "Nick Cox" To Subject st: RE: Any ways to specify the regression output to 2 decimal places? Date Mon, 18 May 2009 10:41:45 +0100

```You can't. This is one reason why so many user-written commands exist in
this territory.

It would be possible in principle to add a large menagerie of options to
commands like -regress-, but there would need to be many such options to
cope with the gamut of possible user desires. StataCorp decided to solve
this problem by ignoring it, or more precisely by providing other tools
for you to pick up and process output.

Note that there is not much reason to expect the results of the -format-
command to apply here, as that concerns how values of a variable are to
be displayed. But (e.g.) the units of the coefficient of X1 in a
regression are the units of Y divided by the units of X1. So which
format should apply? And t and P are dimensionless, so no variable's
format is pertinent.

Nick
n.j.cox@durham.ac.uk

Charles Man@CCR

Is there any way to specify the regression results to 2 decimal places?
I don't mean the figures after the -esttab- or -estout-command. I mean
the regression result table after -regress- command?

It seems fmt% doesn't work in -reg- command. How can I specify the
regression table to show coeff, t, se, p>t and CI with only 2 decimal
places rather than the default 7 decimal places?

Y          |      Coef.    Std. Err.      t        P>|t|
[95% Conf. Interval]
-------------+----------------------------------------------------------
------
X1        |   .0000975   .0068564     0.01   0.989    -.0133751
.0135701
X2         |   .0058169   .0043916     1.32   0.186    -.0028125
.0144463
_cons |     .2223101   .0274083     8.11   0.000       .1684536
.2761665

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```