Stata Technical Bulletin
15
Summary information of the independent variables is obtained via the mean option.
. diest, mean fb(%9.3f) fm(%8.3f) fsd(%8.3f)
|
rep78 Repair Record 1978 |
I |
Coef. |
p>∣t∣ |
Mean |
Std Dev |
|
— |
— | ||||
|
price Price |
I |
0.000 |
0.544 |
6165.257 |
2949.496 |
|
length Length (in.) |
I |
0.013 |
0.136 |
187.932 |
22.266 |
|
mpg Mileage (mpg) |
I |
0.067 |
0.018 |
21.297 |
5.786 |
|
foreign Car type |
I |
1.257 |
0.000 |
0.297 |
0.460 |
|
_cons |
I |
-0.930 |
0.645 |
Finally, we consider a more complicated example of a regression in which interaction effects are generated with xi.
. xi: regress rep78 price i.foreign*length i.foreign*mpg
!.foreign Iforei-0-1 (naturally coded; Iforei-0 omitted)
i.foreign*length IfXlen.# (coded as above)
i.foreign*mpg IfXmpg.# (coded as above)
Source I SS df MS Number of obs = 69
|
---------+- |
— |
— |
F( 6, 62) |
= 8.25 | ||
|
Model I |
29.5749189 |
6 4.92915315 |
Prob > F |
= 0.0000 | ||
|
Residual I |
37.0627623 |
62 .597786488 |
R-Squared |
= 0.4438 | ||
|
Adj R-Squared |
= 0.3900 | |||||
|
--——--———+— |
— | |||||
|
Total I |
66.6376812 |
68 .979965899 |
Root MSE |
= .77317 | ||
|
— rep78 I |
Coef. |
Std. Err. |
t |
p>∣t∣ |
[957. Conf. |
— Interval] |
|
———————+— |
— | |||||
|
price I |
.0000155 |
.0000396 |
0.393 |
0.696 |
-.0000636 |
.0000947 |
|
Iforei-I I |
-2.841789 |
4.520248 |
-0.629 |
0.532 |
-11.87763 |
6.194057 |
|
length I |
.0125723 |
.0109983 |
1.143 |
0.257 |
-.0094131 |
.0345576 |
|
IfXlen.! I |
.0263108 |
.0201156 |
1.308 |
0.196 |
-.0138997 |
.0665213 |
|
Iforei.l I |
(dropped) | |||||
|
≡Pg I |
.0857285 |
.0465508 |
1.842 |
0.070 |
-.0073252 |
.1787822 |
|
IfXmpg-I I |
-.0163025 |
.0573871 |
-0.284 |
0.777 |
-.1310177 |
.0984126 |
|
-Cons I |
-1.232495 |
2.959517 |
-0.416 |
0.679 |
-7.148485 |
4.683496 |
After issuing this command, one can easily obtain more readable output via the command
. diest, fb(%9.3f) fse(%9.3f)
|
rep78 Repair Record 1978 |
I |
Coef. |
Std. Err. |
t |
p>∣t∣ |
|
— |
— | ||||
|
price Price |
I |
0.000 |
0.000 |
0.393 |
0.696 |
|
Iforei-I foreign==l |
I |
-2.842 |
4.520 |
-0.629 |
0.532 |
|
length Length (in.) |
I |
0.013 |
0.011 |
1.143 |
0.257 |
|
IfXlen-I (foreign==l)*length |
I |
0.026 |
0.020 |
1.308 |
0.196 |
|
Iforei-I foreign==l |
I |
(dropped) | |||
|
mpg Mileage (mpg) |
I |
0.086 |
0.047 |
1.842 |
0.070 |
|
IfXmpg-I (foreign==l)*mpg |
I |
-0.016 |
0.057 |
-0.284 |
0.777 |
|
-Cons |
I |
-1.232 |
2.960 |
-0.416 |
0.679 |
sg61 Bivariate probit models
James W. Hardin, Stata Corp., FAX 1-409-696-4601, [email protected]
In this article, we discuss 3 different two-equation probit models that researchers may wish to estimate. They include
Bivariate probit regression for models where the two dependent variables depend on the same list of independent variables and
are correlated.
Seemingly unrelated two-equation probit regression for models where the two dependent variables may not depend on the
same list of independent variables, but are still correlated.
Nested probit regression for models where the outcome of one equation depends on the outcome of the other equation.
Interested readers may also find more information on these models in Greene (1993). Note also that although it is not
discussed in this article, these two commands could be used to extend Heckman-type models to consider two participation
equations.