# st: Re: Multinomial logit

 From "Scott Merryman" <[email protected]> To <[email protected]> Subject st: Re: Multinomial logit Date Wed, 25 Feb 2004 19:43:33 -0600

```Meherun,

I believe -bs- will produce the results you want if you supply the correct
exp_list, in this case it should be _b _se.

For example,

. use http://www.stata-press.com/data/r8/sysdsn3.dta
(Health insurance data)

. bs "mlogit insure age " _b _se , nobc nope

command:      mlogit insure age
statistics:   b_age      = [Prepaid]_b[age]
b_cons     = [Prepaid]_b[_cons]
b_1age     = [Uninsure]_b[age]
b_1cons    = [Uninsure]_b[_cons]
se_age     = [Prepaid]_se[age]
se_cons    = [Prepaid]_se[_cons]
se_1age    = [Uninsure]_se[age]
se_1cons   = [Uninsure]_se[_cons]

Bootstrap statistics                              Number of obs    =       615
Replications     =        50

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
b_age |    50 -.0083061  .0034922   .005469  -.0192965   .0026843   (N)
b_cons |    50  .3128955 -.1757527  .2138954  -.1169433   .7427343   (N)
b_1age |    50 -.0036252  .0030272    .01249  -.0287248   .0214744   (N)
b_1cons |    50 -1.710738 -.1441363  .5947692  -2.905972   -.515505   (N)
se_age |    50  .0059433 -.0017707  .0027628   .0003913   .0114954   (N)
se_cons |    50  .2768637 -.0821143  .1288932   .0178431   .5358842   (N)
se_1age |    50  .0112765  8.77e-06  .0007574   .0097545   .0127985   (N)
se_1cons |    50  .5281215   -.00135  .0308724   .4660811   .5901618   (N)
------------------------------------------------------------------------------
Note:  N   = normal

Hope this helps,
Scott

----- Original Message -----
From: <[email protected]>
To: <[email protected]>
Sent: Wednesday, February 25, 2004 12:09 PM
Subject: st: Multinomial logit

> Hi all,
> I am trying to estimate a multinomial logit model where y takes on more than 3
values.
> For some of the outcome (say y=3), the number of observations is really small,
Stata estimates the coeffcient but but puts a "." in place of standard error. I
want to Bootstrap my way out of this. Did any of you do bootstrap in MNL
framework?
>
> If you use the canned BS command it just gives the BS estimated for the first
outcome (Y=1). I think I need to write a ado file. Any suggestions on how to go
>
> Thanks,
>
> meherun
>

*
*   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/
```

• References: