All entries for Thursday 11 August 2005

August 11, 2005

IPW

(1) Example 4.1 of Wooldridge (2002) shows that E(y|x) can be
m(x,theta), a nonlinear function. Also, the model (distribution) of Y|X can be a member of the linear exponential family yielding the quasi log-likelihood.

(2) In this light, there seems to be four components: distribution of Y|X, model for E(Y|X), distribution of X and a model of R.

(3) In the case where we have to estimate selection probabilities, Y and X are required to be observed only when S =1. Thus, X can be missing with Y???

Normally, MAR => R depends on X which is fully observed. In this case, it is still MAR but R depends on the completely observed Z and X can be missing ???


August 2005

Mo Tu We Th Fr Sa Su
Jul |  Today  | Sep
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31            

Search this blog

Favourite blogs

Galleries

Most recent comments

  • Hi im yo man my intenet namee imm 43 i woork in the toilet withh laptops this coool webbsete by yo man on this entry
  • this suck i'm out of here by katie on this entry
  • this is a stupid website im 89 years old in two weeks it's my birthday how old would i be this is an… by glf on this entry
  • What r u talking about? I think u must be confused. No No No ur major is not econometrics. U have b… by MMP on this entry
  • This blog is temporary out of service because the owner is still busy watching TV (no matter what la… by MMP on this entry

Blog archive

Loading…
Not signed in
Sign in

Powered by BlogBuilder
© MMXXI