Dear stata users,
I have a group of 6 commodities and I want to model the choice behaviour
of consumers. I have household wise data on these commodities. Now the
problem is that more than 30% of the consumers in my sample are not
consuming any of the commodities I am cosidering. If I take any one
commodity from the sample more than 50% does not have consumption of that
commodity. Also consumers who are consuming these commodities some time
consume more than one of this commodity. If I want to model the probabilty
of consuming say x1 over x2 (odds ratio) can multinomial logit model be
applied here. Since My data is highly hetroscedastic and it violates
normality assumptions would it be apropriate to use some nested logit
model (conditional logit) instead? Any suggestions would be appreciated.
Thanks,
Rijo John.
Research Scholar
Mumbai, India
*
* 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/
