Announcement

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

  • Specify reference category to be omitted from a set of multicollinear variables

    Dear all,

    I am trying to run the following regression:

    Code:
    svy: tpoisson lnincome lnexp count_00_05 - count_g37, ll(0)
    where count_00_05, count_05_010, count_10_13,...count_g37 are variables that count the number of times temperature occurs in a certain bin. When I run this regression, Stata automatically drops a bin (because they are perfectly collinear) but it does so arbitrarily. Is it possible to specify the bin to be dropped (as reference category)?

    Even when I drop one myself, Stata drops another one from the the bins included.

    I would be grateful for any suggestion. Thank you.

    Sincerely,

    Chiara
    Last edited by Chiara Piazzo; 11 Dec 2017, 02:40.

  • #2
    Chiara:
    see -help fvvarlist-.
    Kind regards,
    Carlo
    (Stata 15.1 SE)

    Comment


    • #3
      Dear Professor Carlo,

      Thank you for your reply. If I understand correctly, this is for specifying the reference group for factor/group variables?

      However, I have a series of variables which are temperature bins. Is it possible to apply something similar to ib3 in this case? Thank you again!

      Sincerely,

      Chiara

      Comment


      • #4
        Chiara:
        yes, you're correct.
        Unfortunately, there's no a similar tool for continuous variables,
        You can omit the independent variable you want just avoiding to include it among predictors and see it it breaks the collinearity.
        As an aside, please call me Carlo, as all on (and many more off) the list do. Thanks.
        Kind regards,
        Carlo
        (Stata 15.1 SE)

        Comment


        • #5
          Dear Carlo,

          Thank you again for your advice. Unfortunately, the collinearity issue remains. I will keep trying, thanks again!

          Sincerely,

          Chiara

          Comment


          • #6
            Chiara:
            you may want to take a look at -estat vce, corr- to spot the culprit(s) and (re)specify your regression model accordingly.
            Kind regards,
            Carlo
            (Stata 15.1 SE)

            Comment

            Working...
            X