st: Re: xtmixed with log-transfered dependent variable: back to non-log on margins and marginsplot

["Joseph Coveney" <stajc2@gmail.com>]

Wensheng Sun wrote:

Hi, Please help. How to do post multilevel regression estimation after gllamm? I
tried "margins" command, it seems not work. Please fix it for me. Thanks.

[snip]

--------------------------------------------------------------------------------

You could try something like that below.  

If a variable is not a factor variable (that is, if it is not included in the
model with the i.variable notation*), then it is considered by Stata to be a
continuous variable.  So, despite its being coded 0/1, _I_girl_1 is considered
by Stata to be a continuous variable.  

-margins- requires its first arguments to be factor variables--variables that
were in the model as i.variable.  Because _I_girl_1 is a continuous variable and
not a factor variable, you must treat it as a continuous variable when you refer
to it with -margins-.

Joseph Coveney

*-anova- and -manova- are notable exceptions.


[Model fitting omitted for brevity.  To see it, look at my most recent post in
this thread.]

. margins , at(_Igirl_1 = (0 1) age = (0(0.5)2.5))
Warning: cannot perform check for estimable functions.

Predictive margins                                Number of obs   =        198
Model VCE    : OIM

Expression   : predict()

1._at        : age             =           0
               _Igirl_1        =           0

2._at        : age             =           0
               _Igirl_1        =           1

3._at        : age             =          .5
               _Igirl_1        =           0

4._at        : age             =          .5
               _Igirl_1        =           1

5._at        : age             =           1
               _Igirl_1        =           0

6._at        : age             =           1
               _Igirl_1        =           1

7._at        : age             =         1.5
               _Igirl_1        =           0

8._at        : age             =         1.5
               _Igirl_1        =           1

9._at        : age             =           2
               _Igirl_1        =           0

10._at       : age             =           2
               _Igirl_1        =           1

11._at       : age             =         2.5
               _Igirl_1        =           0

12._at       : age             =         2.5
               _Igirl_1        =           1

------------------------------------------------------------------------------
             |            Delta-method
             |     Margin   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         _at |
          1  |   1.828415    .042767    42.75   0.000     1.744593    1.912237
          2  |    1.74101    .029594    58.83   0.000     1.683007    1.799013
          3  |    1.99406   .0351737    56.69   0.000     1.925121    2.062999
          4  |   1.906655   .0277658    68.67   0.000     1.852235    1.961075
          5  |   2.159704   .0286555    75.37   0.000     2.103541    2.215868
          6  |   2.072299   .0290146    71.42   0.000     2.015432    2.129167
          7  |   2.325349   .0241011    96.48   0.000     2.278112    2.372586
          8  |   2.237944   .0329929    67.83   0.000     2.173279    2.302609
          9  |   2.490994   .0227236   109.62   0.000     2.446456    2.535531
         10  |   2.403589   .0388715    61.83   0.000     2.327402    2.479775
         11  |   2.656638   .0250527   106.04   0.000     2.607536    2.705741
         12  |   2.569233   .0459263    55.94   0.000     2.479219    2.659247
------------------------------------------------------------------------------

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