All entries for Monday 08 August 2005

August 08, 2005

must work harder

Note about IPW
(1) There is an example of how to use IPW in the notes from introduction-to-missing-data course. But the missing variable is a covariate, not a dependent variable. We can follow this example. Notice that the same variables are used in both the structure model and the missing-data model. Must check what Wooldridge has to say about this. Also, the coefficient of the missing covariate is still significant and of the same magnitude even though we delete some observations due to the missingness in this variable. However, the missingness of this covariate does affect the coefficient of other covariates such as "gender". This is quite unexpected.

(2) Tell Mark about the limitation of Heckman's approach which is that it wroks only with some special case of non-linear model (exponential). This is noted by Wooldridge (2002). However, nonparametric version of Heckman's approach may be intended as an relaxation of this limitation.

(3) Consider the IPW estimator in the light of Mark's comment. The model of Y given X can be non-linear as in the nonlinear LS case. The distribution of X has to satisfy certain conditions but we dont have to specify it ( again as in nonlinear LS case; look in Wooldridge (2002)). However, we have to model R and make MAR Assumption. Note that although the model of R has to be known, nonparamatric modelling is possible. Wooldridge suggests us to read the paper of Hirano, Imbens..


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