(Sorry for the delay in answering this, but it's Christmas and I need to
get my head straight before answering technical questions to the best of
my mental ability!)
John Fulton wrote:
> I'm using clogit as McFadden's choice model, and
> I want to identify the effects of a common feature of a subset of my
> alternatives.
> Then I want to interact this with individual characteristics to see if
> different choosers
> pick different broad classes of alternatives.
>
> To start, I assume it is reasonable to enter common group features into
> my data as a single dummy variable.
> E.G:
> alt: char:
> 1 0
> 2 1
> 3 0
> 4 1
> alternatives 2 and 4 possess the attribute, while 1 and 3 do not.
>
> Then, I interact the "char" variable with an indicator variable for the
> chooser, e.g., race. This chooser-characteristic is constant within
> "group", but because the indicator variable for the alternatives is not,
> the coefficient doesn't drop out.
>
> I am worried, because as I read the literature, I have not found anyone
> employing this approach. Is this legitimate? Please help me find ways
> to think about this.
Hmmm, others may think otherwise, but I'm not sure it is a legitimate
approach. I know that when I reconfigure datasets for McFadden-choice
model-fitting, I always -expand- them by the number of alternatives and
then proceed from there.
CLIVE NICHOLAS |t: 0(044)7903 397793
Politics |e: [email protected]
Newcastle University |http://www.ncl.ac.uk/geps
Whereever you go and whatever you do, just remember this. No matter how
many like you, admire you, love you or adore you, the number of people
turning up to your funeral will be largely determined by local weather
conditions.
*
* 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/
