Re: st: mlogit outlier

[Friedrich Huebler <huebler@rocketmail.com>]

Somsupa,



Are you sure that mlogit is the proper approach for your problem? What is the theory behind "country" as the dependent variable in your model?


Friedrich



----- Original Message ----

From: somsupa Nopprach <nmayecon31@yahoo.com>

To: statalist@hsphsun2.harvard.edu

Sent: Monday, October 23, 2006 6:57:31 AM

Subject: st: mlogit outlier



Dear Statalisters,

 I am now running mlogit model that has four

categories. My problem is that my Pseudo R2 is 1 while

my loglihood is -2.676e-13. So I think it should be

some problems such as outlier. I have checked my data

by deleting data that has independent value exceed and

lower than 3time of standard deviation.

    However I still face the same problem. could

anyone suggest me what I will do or how I will check

outlier in mlogit model.



  The following is my model estimation.



Thank you very much.

Somsupa



mlogit loca lnprod mmr  gdpdeflator_2000 k_l                lnaggl

,nolog base(3) robust        

                        

Multinomial logistic regression                Number of obs   =  

     453        

                Wald chi2(15)   =          .        

                Prob > chi2     =          .        

Log pseudolikelihood = -2.676e-13                Pseudo R2       =

    1.0000        

                        

loca    Coef.    Robust    z    P>z     [95% Conf. Interval]        

        Std. Err.                

Thailand                             

lnprod    24.36669    1.398603    17.42    0    21.62548    27.1079

mmr    -1.161723    0.1949555    -5.96    0    -1.543829    -0.7796175

gdpdefl~2000    -6.010435    0.3225363    -18.63    0    -6.642595

-5.378276

k_l    -41.4972    0.460219    -90.17    0    -42.39921    -40.59519

lnaggl    168.1013    9.164816    18.34    0    150.1386    186.064

_cons    -484.5797    .    .    .    .    .

                        

Malaysia                        

lnprod    -8.564474    0.6935952    -12.35    0    -9.923896

-7.205052

mmr    -6.412616    0.0790822    -81.09    0    -6.567615    -6.257618

gdpdefl~2000    -0.1889368    0.0214649    -8.8    0    -0.2310073

-0.1468664

k_l    53.92499    0.8171914    65.99    0    52.32332    55.52665

lnaggl    -26.08898    1.041421    -25.05    0    -28.13013    -24.04784

_cons    213.2393    6.827779    31.23    0    199.8571    226.6215

                        

Indonesia                        

lnprod    21.32817    3.32382    6.42    0    14.81361    27.84274

mmr    7.19923    0.4604767    15.63    0    6.296713    8.101748

gdpdefl~2000    -9.331812    0.5171074    -18.05    0    -10.34532

-8.3183

k_l    -70.35383    1.866545    -37.69    0    -74.01219    -66.69547

lnaggl    255.675    14.29446    17.89    0    227.6584    283.6916

_cons    -713.84    10.95042    -65.19    0    -735.3025    -692.3776

                        

(loca==Philippines is the base outcome)       





*
*   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/