Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Obtaining random intercept for each group from xtlogit or xtmelogit
From
Phil Clayton <[email protected]>
To
[email protected]
Subject
Re: st: Obtaining random intercept for each group from xtlogit or xtmelogit
Date
Thu, 21 Jul 2011 06:55:04 +1000
Yes, from the model you've specified each country should have the same intercept. Given how tiny those standard deviations are I would guess this is simply a computer precision issue.
Phil
On 21/07/2011, at 2:26 AM, Orian Brook wrote:
> Brilliant, thank you
>
> I was expecting the intercepts to be common across countries - however, some
> have standard deviations (tiny, but still). Am I misunderstanding?
>
> Summary of random effects for
> NATION COUNTRY: _cons
> (SAMPLE) Mean Std. Dev. Freq.
>
> BELGIUM .08866321 . 1040
> DENMARK .85474342 1.179e-07 1008
> GERMANY W .44680032 . 1006
> GERMANY E .66576046 . 528
> GREECE -.70365345 . 1000
> SPAIN .0415894 . 1006
> FINLAND .69681209 . 1040
> FRANCE -.02973256 . 1031
> IRELAND .19429855 1.667e-08 1000
> ITALY -.05471432 . 1000
> LUXEMBOUR .94192523 9.375e-08 500
> NETHERLAN 1.0726261 0 1000
> AUSTRIA .14302817 . 1011
> PORTUGAL -.53366905 . 1013
> SWEDEN 1.2025346 . 1011
> GREAT BRI .64745915 . 1009
> NORTHERN .32294551 . 301
> CYPRUS (R -.52049023 . 500
> CZECH REP .73380858 . 1060
> ESTONIA .51977873 0 1001
> HUNGARY .11004449 . 1000
> LATVIA .6622172 . 1006
> LITUANIA .10387274 1.520e-08 1029
> MALTA .13686518 . 500
> POLAND -.42845824 . 1000
> SLOVAKIA .4025459 . 1094
> SLOVENIA .4127627 . 1015
> BULGARIA -.67533928 . 1009
> ROMANIA -.59633559 7.265e-08 1028
> CROATIA -6.7901573 0 1000
>
> Total -.02248435 1.4072492 27746
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Phil Clayton
> Sent: 20 July 2011 14:57
> To: [email protected]
> Subject: Re: st: Obtaining random intercept for each group from xtlogit or
> xtmelogit
>
> After xtmelogit try:
> predict rintercept, reffects
>
> Phil
>
> On 20/07/2011, at 11:10 PM, Orian Brook wrote:
>
>> Dear all
>>
>> I'm have data on a binary outcome which is organised as respondents
>> within countries. What I really want to know is the random intercept
>> for each country, but I can't see how to get it - could anyone point
>> me? (I was previously not using a multilevel approach and had a dummy
>> variable for each country, to get the country effect - but I realised
>> I really should try a multilevel approach).
>>
>> The syntax I'm using is
>>
>> xi:xtmelogit QA4_PERF VD8 i.D10 VD11 i.C14 i.QB3 i.D25 || COUNTRY:
>> xi:xtlogit QA4_PERF VD8 i.D10 VD11 i.C14 i.QB3 i.D25, i(COUNTRY)
>>
>> They get the same log-likelihood results, though naturally xtmelogit
>> takes longer to run, I realise it has other options available.
>> Thanks for any advice.
>>
>>
>>
>> *
>> * For searches and help try:
>> * http://www.stata.com/help.cgi?search
>> * http://www.stata.com/support/statalist/faq
>> * http://www.ats.ucla.edu/stat/stata/
>
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
>
> *
> * For searches and help try:
> * http://www.stata.com/help.cgi?search
> * http://www.stata.com/support/statalist/faq
> * http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/