Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • how to obtain the correct AIC / BIC values after a model using [pw=...]

    Hi everyone,

    In general, I use -estat ic- or -fitstata- to obtain the bic & aic values after the logit model.

    However, when the model includes the option [pw=...], the commands above would give the bic/aic values in very huge magnitude, such as 1.418e+08 for the simple reason that the log pseudolikelihood fo the model is very tiny.

    Anyway, if I just type -estat ic- after the model with the option [pw=...], I would get some abnormal values of BIC & AIC.

    How could I fix this problem ?

    Thanks.

  • #2
    Are AIC and BIC legit after fitting a model with sampling weights?

    Comment


    • #3
      does anyone know how to fix this problem?

      Comment

      Working...
      X