st: Problem with setting the right reference category when using mlogit

[Amal Khanolkar <Amal.Khanolkar@ki.se>]

Hi,

I've been running some multinomial logistic regression models using mlogit.

My dependent variable has three categories as follows:

tab moth_wtgainbybmi

mother's weight gain by |
 pre-pregnancy BMI, IOM |
             guidelines |      Freq.     Percent        Cum.
------------------------+-----------------------------------
        Weight gain low |     51,716       22.37       22.37
Weight gain appropriate |     66,311       28.68       51.05
       Weight gain high |    113,164       48.95      100.00
------------------------+-----------------------------------
                  Total |    231,191      100.00

When I run an mlogit model as follows:

mlogit moth_wtgainbybmi i.race age i.education i.smoking height i.parity married, rrr, if plural==1

the model automatically sets 'weight gain high', the third category in the above variable as the base outcome. I however would like the second category, 'weight gain appropriate' to be the 'base outcome'. The three categories above are coded as 0, 1, 2. 

Even if I recode the variable, such that weight gain appropriate==1, weight gain low==2 and weight gain high==3, 
the latter, 'weight gain high' still gets set as the 'base outcome' in the mlogit model.

How do I change my reference category in the above model?

(and strangely enough, I do NOT have the same problem if I use another categorical variable as a dependent variable in the above model).


Thanks,

/Amal


*
*   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/