Announcement

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

  • Significance starts with -margins, nose post

    I am running -ivprobit- and obtaining the marginal effects using
    Code:
    margins, dydx(*) post nose
    . I need to use nose option, else the estimation takes forever.
    I am then using -esttab- to export the results. However, I cannot see the significance stars. My hunch is this happens due to the nose option. Since standard errors are not calculated, the significance stars cannot be generated.

    Is there a workaround for this? Can I "pick" the stars from the ivprobit regression and the estimates from the margins?

    Edit: it should read "stars" not starts in the title.
    Last edited by Parul Gupta; 19 Mar 2024, 01:58. Reason: Title should read "stars" not "starts"

  • #2
    No, the parameters in the regression output are related but not the same as the marginal effects and the p-value (and thus stars) can differ. So unfortunately, you have just have to remove the nose option and do something else productive while margins works.
    ---------------------------------
    Maarten L. Buis
    University of Konstanz
    Department of history and sociology
    box 40
    78457 Konstanz
    Germany
    http://www.maartenbuis.nl
    ---------------------------------

    Comment


    • #3
      Oh no, that's a bummer.

      Can you suggest a way to speed up estimation without nose option? I have 600 dummies in my regression (region fixed effects) which is probably causing slow speed.

      Comment

      Working...
      X