Announcement

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

  • country fixed effects or cluster standardised errors for cross-country analysis

    Hello,
    i am doing a master thesis on the factors associated with help-seeking behavior of domestic violence victims. I have an appended dataset of 10 countries (cross-sectional datasets) and i want to run one regression for all. I have to use a logisitc regression due to the binary nature of the dependent variable. I was initially running the regression using i.country for country fixed effects to control for unobserved heterogeneity across countries. However, i was wondering if cluster standardized errors (vce (cluster country)) should be used instead in this case.
    thank you

  • #2
    Lynn:
    you can use them both.
    Kind regards,
    Carlo
    (Stata 16.0 SE)

    Comment


    • #3
      Thank you for your answer. if i use i.country for country fixed effects, how can i represent it in the regression equation? Is putting the subscript i next to the explanatory variables correct? such as
      Logit (HS)i = β0 + β1 agei + β2 age diffi + β3children under 5i + β4edui + .......

      Comment


      • #4
        Lynn:
        it should be something like
        Code:
        β5countryi 
        Kind regards,
        Carlo
        (Stata 16.0 SE)

        Comment


        • #5
          Carlo Lazzaro Dear Carlo,

          In the same type of model explained above (cross-national, cross-sectional), can I add country-level variables? (i.e. a cross-national multilevel model with cross-sectional data)

          One comment I received said: "If you are using a country-level fixed-effects model, and if you include a time-invariant country level covariate, that covariate will fall out of the model, since country-level fixed effects make it such that you are comparing individuals within a country to each other, and therefore there will be no variation on this time-invariant characteristic. Either I am misunderstanding the model (e.g., if country level variable is time-varying) or the analysis was not conducted correctly."

          When I run my model in Stata such as

          Code:
           reg outcomevar covariates_individual_level covariates_country_level i.country [pweight=individual_level_weight], cluster(country)
          the country level covariates don't fall out the model.

          To me it was clear I could add country-level variables in such a case, but after that comment I'm wondering whether my initial understanding is correct.

          Thank you so much. Really appreciated.

          Ana

          Comment


          • #6
            Ana:
            1) it may be that the reviewer is wondering whether it makes sense to include a predictor such -i.country- that, in all likelihood, is time-invariant if there are in/outflows of individuals. At the same time, it seems that reviewer is assuming that you have run a fixed effect model. However, from your code I would say that you've run a cross-sectional regression (not a panel data fixed effect regression) and, if, my intuition is correct, it makes sense to include -i.country-.
            Usually this kind of confusion arises when the Methods section of your paper is not that clear. Hence, I would recommend you to double-check it, perhaps with the help of an experienced colleague.
            2) that said, if you have a multilevel design with citiziens nested within countries, why not following that road?
            Kind regards,
            Carlo
            (Stata 16.0 SE)

            Comment


            • #7
              Dear Carlo,
              Thank you very much for your response - very helpful re multilevel design suggestion.
              Best wishes,
              Ana

              Comment

              Working...
              X