[Limdep Nlogit List] Choice based sampling and panel data

Zainab Jamal zainab_jamal at yahoo.com
Tue Sep 18 08:12:49 EST 2007

  I wanted to estimate a binomial logit model and a mixed logit model.
  The data has a panel structure - individuals with responses to multiple campaigns.
  The dependent variable has two values: responded/ did not respond to an email campaign.
  I select a sample with equal number of responders and non-responders.
  If you go by individuals: its 50:50 in the sample and 5:95 in the population.  
  How do I create the wts? Do I wt all the observations of an individual similarly - deflate or inflate depending on whether the person is a responder or a non-responder?
  Or do I wt. differently each observation for each individual? if so, how do I do it? 
  Any help would be really helpful.

More information about the Limdep mailing list