st: mi chained (logit) error

["Jacqueline Jodl" <jmj2138@tc.columbia.edu>]

Dear Statalist,

I am attempting mi chained (logit) for a series of variables indicating
whether or not grandmothers and grandfathers live in the household. 

The measures are dichotomous with 1 for yes and 0 for no.  I am receiving an
error message that the outcomes do not vary (see below).

I have recoded my soft missing or invalid skips (-1s, -2s, -3s) =. But I
have left my valid skips (-7s).  I believe it is the -7s that are the
problem. 

Can I have a dichotomous variable in mi chained (logit) with values other
than just 0 and 1?  Is the -7 the issue?

Thanks,
Jackie Jodl
Teachers College Columbia University

Summary of output:


recode * (-1 -2 -3 = .)

. codebook grandmaHH_1988

----------------------------------------------------------------------------
-----------------------
grandmaHH_1988
1988 grandmaHH_
----------------------------------------------------------------------------
-----------------------

                  type:  numeric (float)

                 range:  [-7,1]                       units:  1
         unique values:  3                        missing .:  15/4188

            tabulation:  Freq.  Value
                           178  -7
                          3577  0
                           418  1
                            15  .

. mi impute chained (logit) grandmaHH_1988 grandmaHH_1990 grandmaHH_1992
grandmaHH_1994 grandmaHH_1
> 996 grandmaHH_1998 grandmaHH_2000 ///
> grandmaHH_2002 grandmaHH_2004 grandmaHH_2006 grandmaHH_2008 grandmaHH_2010
///
> grandpaHH_1988 grandpaHH_1990 grandpaHH_1992 grandpaHH_1994 grandpaHH_1996
grandpaHH_1998 grandpa
> HH_2000 ///
> grandpaHH_2002 grandpaHH_2004 grandpaHH_2006 grandpaHH_2008 grandpaHH_2010
///
> = highgrade highgrademom race male mombirthage birthorder, add(5)
outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandpaHH_2006 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandmaHH_2008 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandpaHH_2008 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandpaHH_1990 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandmaHH_2000 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandpaHH_2000 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandmaHH_1990 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandmaHH_2004 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandpaHH_2004 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandmaHH_1988 = highgrade
highgrademom race male
    mombirthage birthorder

outcome does not vary; remember:
                                  0 = negative outcome,
        all other nonmissing values = positive outcome
 -- above applies to specification (logit ) grandpaHH_1988 = highgrade
highgrademom race male
    mombirthage birthorder

r(2000);

end of do-file

r(2000);

.

*
*   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/