Bookmark and Share

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: example about choice experiment datasheet


From   Arne Risa Hole <[email protected]>
To   statalist <[email protected]>
Subject   Re: st: example about choice experiment datasheet
Date   Mon, 26 Nov 2012 11:42:41 +0000

This is to be expected; if you found it surprising you need to read
the -asclogit- manual entry more carefully. Whether it's correct or
not really depends on what you are trying to do. I would suggest that
including socio-demographics in your model only makes sense if you
interact them with the alternative-specific attributes, since your
alternatives are unlabelled.

Arne

On 26 November 2012 10:43, elisabetta capobianco
<[email protected]> wrote:
> Ok Arne,
> now It's clear and I have had good results.
> Just a doubt. When I put inside the socio-demographic variables, the
> output has the four alternatives 1,2,3,4, even if I write nocons.
> I don't understand the meaning of this output considering that I have
> unlabelled alternatives.
> Is however correct this result?
>
> Thank you,
> Elisabetta
>
>
>
>
>
> 2012/11/24 Arne Risa Hole <[email protected]>:
>> Elisabetta,
>>
>> The "case" variable must a unique identifier for each choice. You can
>> generate it like this
>>
>> egen caseid = group(id scenario)
>>
>> The "alt" variable should be coded 1,2,3,4,1,2,3,4 etc. (not
>> 1,2,3,4,5,6,7,8,...).
>>
>> Arne
>>
>> On 24 November 2012 16:04, elisabetta capobianco
>> <[email protected]> wrote:
>>> Thank you Arne, I have taken the book.
>>> My problem is that I have 16 scenarios for id (n=131) with 4 different
>>> alternatives for each scenario .
>>> So in the dataset I have written for each id non 4 alternatives
>>> (1,2,3,4) but 64 alternatives. Is it correct?
>>>
>>>
>>>
>>> id   scenario      alt
>>> 1         1            1
>>> 1         1            2
>>> 1         1             3
>>> 1         1            4
>>> 1         2            5(or1?)
>>> 1         2            6(or2?)
>>> 1         2            7(or3?)
>>> 1         2            8 (or4?)
>>> 1         3            9(or1?)
>>> .
>>> .
>>> .
>>> .
>>> .
>>> 1         16           64
>>> 2         1              1
>>> .
>>> .
>>> And when I wrote:
>>> asclogit choice organic price, case(id) alternatives(car) noconst
>>>
>>> case is id or scenario? I suppose scenario. But if is scenario I don't
>>> know how to use id!
>>> And alternatives are the 64 alternatives?
>>>
>>>
>>> Thank you so much for your help.
>>>
>>> elisabetta
>>>
>>>
>>>
>>>
>>>
>>> 2012/11/23 Arne Risa Hole <[email protected]>:
>>>> Elisabetta
>>>>
>>>> I am fairly sure that you can use -asclogit- to do what you want. If
>>>> you don't want alternative-specific constants you can use the
>>>> -noconst- option. Modifying the example in the help-file slightly:
>>>>
>>>> webuse choice
>>>> asclogit choice dealer, case(id) alternatives(car) noconst
>>>>
>>>> As you can see the only coefficient which is estimated in this case is
>>>> the one for the dealer attribute, which is analogous to the organic
>>>> and price attributes in your experiment. In your case the variable
>>>> that identifies the alternatives for each case ("car" in the example)
>>>> can simply take the values 1, 2, 3 and 4 since your alternatives are
>>>> unlabelled.
>>>>
>>>> I also recommend taking a look at a basic choice modelling text before
>>>> proceeding, such as
>>>> http://books.google.co.uk/books/about/Applied_Choice_Analysis.html?id=8yZrtCCABAgC.
>>>>
>>>> Good luck!
>>>>
>>>> Arne
>>>>
>>>> On 23 November 2012 13:25, elisabetta capobianco
>>>> <[email protected]> wrote:
>>>>> I think that asclogit is not the right comand for me.
>>>>> Because I have 4 unlabeled alternatives, 16 diferrent scenarios with 4
>>>>> alternatives for 131 individuals. So my indipendent viariables are the
>>>>> attributes and dipendent variables the utility for each level of
>>>>> attribute.
>>>>> I need to know the importance of every level, not of alternatives.
>>>>> Ex.:n=131
>>>>> scenario 1)   A       B        C         D
>>>>>      organic  org   no org   no org   org
>>>>>      price       4         8         4         4
>>>>> scenario 2)   A      B         C        D
>>>>>     organic  no org  no org   org     no org
>>>>>       price     12     16         8        8
>>>>> .
>>>>> .
>>>>> .
>>>>> Perhaps with mixlogit? Because the example is on a specific
>>>>> alternative, car or bus.
>>>>> I instead need to know the coefficient of attributes and levels.
>>>>>
>>>>> I don't know if I have explained well, I hope!
>>>>>
>>>>> e
>>>>>
>>>>>
>>>>>
>>>>> 2012/11/23 Arne Risa Hole <[email protected]>:
>>>>>> -help asclogit_postestimation-
>>>>>>
>>>>>> Arne
>>>>>>
>>>>>> On 23 November 2012 11:05, elisabetta capobianco
>>>>>> <[email protected]> wrote:
>>>>>>> Dear Arne,
>>>>>>> i need to estimate the marginal effect of levels on attributes of my choice.
>>>>>>> Is it possible?
>>>>>>> es:
>>>>>>> about meat:
>>>>>>> attribute   level
>>>>>>> label        organic/not organic
>>>>>>> origin       local abroad
>>>>>>> prize        4/8
>>>>>>>
>>>>>>> I would know the utility for each level of each attribute.
>>>>>>>
>>>>>>> Thank you!
>>>>>>> elisabetta
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> 2012/11/21 elisabetta capobianco <[email protected]>:
>>>>>>>> Thank you so much Arne!!!
>>>>>>>> I'm creating the data set and then i'll try to run in Stata!
>>>>>>>>
>>>>>>>> Elisabetta
>>>>>>>>
>>>>>>>> 2012/11/20 Arne Risa Hole <[email protected]>:
>>>>>>>>> Elisabetta
>>>>>>>>>
>>>>>>>>> -mlogit- is not very useful to model data from choice experiments as
>>>>>>>>> it does not allow alternative-specific explanatory variables.
>>>>>>>>> -asclogit- requires your data to be in long form (one row per
>>>>>>>>> alternative); see -help asclogit- and the first example in the paper I
>>>>>>>>> mentioned in my last email.
>>>>>>>>>
>>>>>>>>> Arne
>>>>>>>>>
>>>>>>>>> On 20 November 2012 10:04, elisabetta capobianco
>>>>>>>>> <[email protected]> wrote:
>>>>>>>>>> Thank you Arne, I want to use asclogit and mlogit for my choice modelling.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> 2012/11/19 Arne Risa Hole <[email protected]>:
>>>>>>>>>>> Elisabetta
>>>>>>>>>>>
>>>>>>>>>>> As Jay says the correct data form depends on the estimation command
>>>>>>>>>>> you are planning to use, so this is a very important piece of
>>>>>>>>>>> information that you haven't given us. You may find the examples in
>>>>>>>>>>> http://www.stata-journal.com/article.html?article=st0133 useful.
>>>>>>>>>>> Generally speaking the best strategy is to import the data into Stata
>>>>>>>>>>> and then use commands like -reshape- to convert the data into the form
>>>>>>>>>>> you want, rather than trying to do this in Excel.
>>>>>>>>>>>
>>>>>>>>>>> Arne
>>>>>>>>>>>
>>>>>>>>>>>> On 19 November 2012 15:51, elisabetta capobianco
>>>>>>>>>>>> <[email protected]> wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thank you for your answer!
>>>>>>>>>>>>> I want to do a mixed logit, but I have found a data set on help file
>>>>>>>>>>>>> with just one choice set. I have 16 choice sets and I don't know how
>>>>>>>>>>>>> to fil the excel data set.
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> 2012/11/18 JVerkuilen (Gmail) <[email protected]>:
>>>>>>>>>>>>> > On Sun, Nov 18, 2012 at 4:11 AM, elisabetta capobianco
>>>>>>>>>>>>> > <[email protected]> wrote:
>>>>>>>>>>>>> >> Hi, I'm new of statalist and stata!
>>>>>>>>>>>>> >> I'm working on choice experiment about food preferences in schools.
>>>>>>>>>>>>> >> I have 4 alternatives unlabelled, 4 attributes and 4 levels, 130
>>>>>>>>>>>>> >> individuals.
>>>>>>>>>>>>> >> I have the data base in excel, but before insheet the file in stata
>>>>>>>>>>>>> >> i'm not sure that i have put the data in the right mode in the excel
>>>>>>>>>>>>> >> file.
>>>>>>>>>>>>> >> Have you got an example on a excel datasheet about choice modelling?
>>>>>>>>>>>>> >> Just an example to understand how to organize my datasheet.
>>>>>>>>>>>>> >
>>>>>>>>>>>>> > It depends on what analysis you want to use. -clogit- and procedures
>>>>>>>>>>>>> > like it that are using alternative-specific predictors use the same
>>>>>>>>>>>>> > basic format, which is specified in the help file. You can see the
>>>>>>>>>>>>> > example dataset there. Essentially it is a "long" format like
>>>>>>>>>>>>> > person-period in a growth model with time-varying covariates.
>>>>>>>>>>>>> > *
>>>>>>>>>>>>> > *   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/
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> --
>>>>>>>>>>>>> Rispetta l'ambiente: se non ti è necessario, non stampare questa mail
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>>           Le informazioni contenute in questa comunicazione e gli
>>>>>>>>>>>>> eventuali documenti allegati
>>>>>>>>>>>>>
>>>>>>>>>>>>>           hanno carattere confidenziale e sono a uso esclusivo del
>>>>>>>>>>>>> destinatario. Nel caso
>>>>>>>>>>>>>
>>>>>>>>>>>>>           questa comunicazione Vi sia pervenuta per errore, Vi informo
>>>>>>>>>>>>> che la sua diffusione
>>>>>>>>>>>>>
>>>>>>>>>>>>>           e riproduzione è contraria alla legge e prego di darmi
>>>>>>>>>>>>> prontamente avviso
>>>>>>>>>>>>>
>>>>>>>>>>>>>           e di cancellare quanto ricevuto. Grazie.
>>>>>>>>>>>>>
>>>>>>>>>>>>> *
>>>>>>>>>>>>> *   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/
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> *
>>>>>>>>>>> *   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/
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> --
>>>>>>>>>> Rispetta l'ambiente: se non ti è necessario, non stampare questa mail
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>           Le informazioni contenute in questa comunicazione e gli
>>>>>>>>>> eventuali documenti allegati
>>>>>>>>>>
>>>>>>>>>>           hanno carattere confidenziale e sono a uso esclusivo del
>>>>>>>>>> destinatario. Nel caso
>>>>>>>>>>
>>>>>>>>>>           questa comunicazione Vi sia pervenuta per errore, Vi informo
>>>>>>>>>> che la sua diffusione
>>>>>>>>>>
>>>>>>>>>>           e riproduzione è contraria alla legge e prego di darmi
>>>>>>>>>> prontamente avviso
>>>>>>>>>>
>>>>>>>>>>           e di cancellare quanto ricevuto. Grazie.
>>>>>>>>>>
>>>>>>>>>> *
>>>>>>>>>> *   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/
>>>>>>>>>
>>>>>>>>> *
>>>>>>>>> *   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/
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>> Rispetta l'ambiente: se non ti è necessario, non stampare questa mail
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>           Le informazioni contenute in questa comunicazione e gli
>>>>>>>> eventuali documenti allegati
>>>>>>>>
>>>>>>>>           hanno carattere confidenziale e sono a uso esclusivo del
>>>>>>>> destinatario. Nel caso
>>>>>>>>
>>>>>>>>           questa comunicazione Vi sia pervenuta per errore, Vi informo
>>>>>>>> che la sua diffusione
>>>>>>>>
>>>>>>>>           e riproduzione è contraria alla legge e prego di darmi
>>>>>>>> prontamente avviso
>>>>>>>>
>>>>>>>>           e di cancellare quanto ricevuto. Grazie.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Rispetta l'ambiente: se non ti è necessario, non stampare questa mail
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>           Le informazioni contenute in questa comunicazione e gli
>>>>>>> eventuali documenti allegati
>>>>>>>
>>>>>>>           hanno carattere confidenziale e sono a uso esclusivo del
>>>>>>> destinatario. Nel caso
>>>>>>>
>>>>>>>           questa comunicazione Vi sia pervenuta per errore, Vi informo
>>>>>>> che la sua diffusione
>>>>>>>
>>>>>>>           e riproduzione è contraria alla legge e prego di darmi
>>>>>>> prontamente avviso
>>>>>>>
>>>>>>>           e di cancellare quanto ricevuto. Grazie.
>>>>>>>
>>>>>>> *
>>>>>>> *   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/
>>>>>>
>>>>>> *
>>>>>> *   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/
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Rispetta l'ambiente: se non ti è necessario, non stampare questa mail
>>>>>
>>>>>
>>>>>
>>>>>           Le informazioni contenute in questa comunicazione e gli
>>>>> eventuali documenti allegati
>>>>>
>>>>>           hanno carattere confidenziale e sono a uso esclusivo del
>>>>> destinatario. Nel caso
>>>>>
>>>>>           questa comunicazione Vi sia pervenuta per errore, Vi informo
>>>>> che la sua diffusione
>>>>>
>>>>>           e riproduzione è contraria alla legge e prego di darmi
>>>>> prontamente avviso
>>>>>
>>>>>           e di cancellare quanto ricevuto. Grazie.
>>>>>
>>>>> *
>>>>> *   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/
>>>>
>>>> *
>>>> *   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/
>>>
>>>
>>>
>>> --
>>> Rispetta l'ambiente: se non ti è necessario, non stampare questa mail
>>>
>>>
>>>
>>>           Le informazioni contenute in questa comunicazione e gli
>>> eventuali documenti allegati
>>>
>>>           hanno carattere confidenziale e sono a uso esclusivo del
>>> destinatario. Nel caso
>>>
>>>           questa comunicazione Vi sia pervenuta per errore, Vi informo
>>> che la sua diffusione
>>>
>>>           e riproduzione è contraria alla legge e prego di darmi
>>> prontamente avviso
>>>
>>>           e di cancellare quanto ricevuto. Grazie.
>>>
>>> *
>>> *   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/
>>
>> *
>> *   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/
>
>
>
> --
> Rispetta l'ambiente: se non ti è necessario, non stampare questa mail
>
>
>
>           Le informazioni contenute in questa comunicazione e gli
> eventuali documenti allegati
>
>           hanno carattere confidenziale e sono a uso esclusivo del
> destinatario. Nel caso
>
>           questa comunicazione Vi sia pervenuta per errore, Vi informo
> che la sua diffusione
>
>           e riproduzione è contraria alla legge e prego di darmi
> prontamente avviso
>
>           e di cancellare quanto ricevuto. Grazie.
>
> *
> *   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/

*
*   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index