Announcement

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

  • Dynamic random effects probit - unobserved heterogeneity and number of right predictions

    Hi,

    I am working with a rather large panel featuring firm level data. I am trying to analyze path dependence innovation investments. My endogenous variable is binary. I am using Stata 13 and fitted a dynamic random effects discrete choice model. I am further using Wooldrige's "Simple Solution to the Initial Condition Problem" (2005). The latter aside, it basically resembles the following:

    Code:
    webuse nlswork
    
    xtset idcode year
    
    *First model
    xtprobit msp l.msp race, vce(robust)
    
    *Second model
    xtprobit msp l.msp race grade, vce(robust)
    I now would like to do two things:

    (1) I would like to know, how much variance is explained with the random effect?
    (2) I am fitting a few models and would like to know, how many right predictions I get. Is that possible as a postestimation?

    Any help is greatly appreciated!
    Thank you
    /R
Working...
X