Clive,
This is kind of late in the game, but was it necessary to alter -
mlbeta- to
obtain estimated coefficients for a set of mutually exclusive dummy
variables?
Could you not use the -lincom- with the constant and dummy variables
(with the
omitted dummy) to obtain the full set of coefficients?
For example:
. sysuse auto,clear
(1978 Automobile Data)
. qui sum mpg
. gen m2 = mpg/(r(max)+1)
. qui tab rep, gen(D)
. mlbeta m2 trunk D*, nocons
-----------------------------------------------------------------------
-------
m2 | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+---------------------------------------------------------
-------
eq1 |
trunk | -.0593707 .0143788 -4.13 0.000 -.0875525 -
.0311888
D1 | .5046506 .3548657 1.42 0.155 -.1908734
1.200175
D2 | .694107 .2678713 2.59 0.010 .1690888
1.219125
D3 | .7618373 .2348673 3.24 0.001 .301506
1.222169
D4 | .8648664 .2248695 3.85 0.000 .4241302
1.305603
D5 | 1.480025 .2255771 6.56 0.000 1.037902
1.922148
-------------+---------------------------------------------------------
-------
eq2 |
_cons | 2.851085 .1657322 17.20 0.000 2.526256
3.175914
-----------------------------------------------------------------------
-------
. mlbeta m2 trunk D*
note: D1 dropped due to collinearity
-----------------------------------------------------------------------
-------
m2 | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+---------------------------------------------------------
-------
eq1 |
trunk | -.0593707 .0143788 -4.13 0.000 -.0875525 -
.0311888
D2 | .1894564 .3828169 0.49 0.621 -
.5608509 .9397637
D3 | .2571867 .3572618 0.72 0.472 -
.4430335 .957407
D4 | .3602158 .3588014 1.00 0.315 -.3430221
1.063454
D5 | .9753741 .368718 2.65 0.008 .2527001
1.698048
_cons | .5046506 .3548657 1.42 0.155 -.1908734
1.200175
-------------+---------------------------------------------------------
-------
eq2 |
_cons | 2.851085 .1657322 17.20 0.000 2.526256
3.175914
-----------------------------------------------------------------------
-------
. lincom D2 + _cons
( 1) [eq1]D2 + [eq1]_cons = 0
-----------------------------------------------------------------------
-------
m2 | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+---------------------------------------------------------
-------
(1) | .694107 .2678714 2.59 0.010 .1690888
1.219125
-----------------------------------------------------------------------
-------
. lincom D3 + _cons
( 1) [eq1]D3 + [eq1]_cons = 0
-----------------------------------------------------------------------
-------
m2 | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+---------------------------------------------------------
-------
(1) | .7618373 .2348673 3.24 0.001 .301506
1.222169
-----------------------------------------------------------------------
-------
.
. lincom D4 + _cons
( 1) [eq1]D4 + [eq1]_cons = 0
-----------------------------------------------------------------------
-------
m2 | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+---------------------------------------------------------
-------
(1) | .8648664 .2248695 3.85 0.000 .4241302
1.305603
-----------------------------------------------------------------------
-------
.
. lincom D5 + _cons
( 1) [eq1]D5 + [eq1]_cons = 0
-----------------------------------------------------------------------
-------
m2 | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+---------------------------------------------------------
-------
(1) | 1.480025 .2255771 6.56 0.000 1.037902
1.922148
-----------------------------------------------------------------------
-------
Scott
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
