Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: A logit approach that allows alternative specific variables


From   Peer Skov <peer_skov@hotmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: A logit approach that allows alternative specific variables
Date   Mon, 8 Dec 2008 12:22:15 +0000

Hi Arne
Thanks for your reply!
If it’s okay I have follow question regarding marginal effects. 
I have read in the FAQ: http://www.stata.com/support/faqs/stat/mfx_unsuit.html that -mfx- is unsuitable for calculating marginal effects in the -clogit-. 
However in your answer to Jim http://www.stata.com/statalist/archive/2006-08/msg00899.html 
you provide him two alternatives to do so. 
I was wondering if they also apply to my example?
 
Thanks again for the help.
 
Peer



 



> Date: Wed, 3 Dec 2008 11:27:10 +0000
> From: arnehole@gmail.com
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: A logit approach that allows alternative specific variables
> 
> Hi Peer
> 
> Try this:
> 
> asclogit choice life_earnings sigp sige, case(id) alternatives(edu) noconst
> 
> which should give you the same result as:
> 
> clogit choice life_earnings sigp sige, group(id)
> 
> Note that -clogit- can be used both to estimate the McFadden model
> *and* the fixed effects logit (i.e. a binary logit for panel data with
> the fixed effects conditioned out of the log-likelihood) which is a
> source of some confusion.
> 
> Hope this helps.
> 
> Arne
> 
> On 03/12/2008, Peer Skov  wrote:
>> Dear statlisters,
>>
>> I am confused; which logit command is right for the job. I have a choice situation where each individual has the possibility to choose among four alternatives. I have calculated individual specific expected income for each of these four choices. Furthermore I have also calculated two types of income risk however these two measures are alternative specific. Hence the income risk concerned with choice 1 is the same for all individuals. My initial idea was to use asclogit along the lines
>>
>> asclogit choice life_earnings sigp sige, case(id) alternatives(edu)
>>
>> But the parameters to sigp and sige can not be identified since the model estimates alternative specific constants.
>> My second approach was to use the clogit command however I'm uncertain if this specification is appropriate. As I understand the clogit or fixed effect logit deals with individual changes over time and this is not the case for my data.
>>
>> clogit choice life_earnings sigp sige, group(id)
>>
>> Do any of you have suggestion on logit specification that is appropriate when I also want to include alternative specific(case constant) measures?
>> Best,
>> Peer
>>
>>
>>
>>
>> _________________________________________________________________
>> News, entertainment and everything you care about at Live.com. Get it now!
>> http://www.live.com/getstarted.aspx
>> *
>> * 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/
_________________________________________________________________
Connect to the next generation of MSN Messenger 
http://imagine-msn.com/messenger/launch80/default.aspx?locale=en-us&source=wlmailtagline
*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index