Announcement

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

  • about mixed command and the random effects

    Dear professors



    When I type the command "mixed education culture ses100 $var || schids: culture, cov(unstru) mle" and how can I get the random slope effect of culture?

    the coef of culture 0,11,Is it a fixed effect? or does it involve fixed and random effects? How do I get the effect of the random slope? How do I report the random slope in my paper?
    Click image for larger version

Name:	_20230407094643.png
Views:	1
Size:	29.3 KB
ID:	1708872

  • #2
    The interpretation of a model with random slopes is that each higher-level entity (schid, in your case) has its own slope for the variable, and that the distribution of values of the slopes is normal (Gaussian) with mean equal to the coefficient shown in the fixed effects results, and variance equal to the result shown in the random effects.

    So, in your specific case, the distribution of culture effects among the different schools is normal with mean 0.11 and variance 0.0039. If you prefer to parameterize your normal distributions with mean and standard deviation, the standard deviation will just be the square root of the variance.

    You also have a random intercept at the schid level. So you could, if you wish, characterize the joint distribution of the constant terms among the different schools as bivariate normal with mean(culture effect, constant term) = [0.11, 36.7] and covariance matrix [0.0039, -.0030 \ -.0030 32.5].

    If you want estimates of the individual culture effect and intercept in each school, you can get that with -predict culture_slope constant, reffects- and Stata will calculate them and place them in those two new variables.

    Comment


    • #3
      Originally posted by Clyde Schechter View Post
      The interpretation of a model with random slopes is that each higher-level entity (schid, in your case) has its own slope for the variable, and that the distribution of values of the slopes is normal (Gaussian) with mean equal to the coefficient shown in the fixed effects results, and variance equal to the result shown in the random effects.

      So, in your specific case, the distribution of culture effects among the different schools is normal with mean 0.11 and variance 0.0039. If you prefer to parameterize your normal distributions with mean and standard deviation, the standard deviation will just be the square root of the variance.

      You also have a random intercept at the schid level. So you could, if you wish, characterize the joint distribution of the constant terms among the different schools as bivariate normal with mean(culture effect, constant term) = [0.11, 36.7] and covariance matrix [0.0039, -.0030 \ -.0030 32.5].

      If you want estimates of the individual culture effect and intercept in each school, you can get that with -predict culture_slope constant, reffects- and Stata will calculate them and place them in those two new variables.
      Thanks professor, your means is that "0.11" is the mixed effects(included fixed and random)?

      Comment


      • #4
        your means is that "0.11" is the mixed effects(included fixed and random)
        Well, that's an interesting way to think of it. But that's not what the term mixed effects refers to. It refers to the fact that the model as a whole contains both fixed and random effects. No individual effect is described as "mixed."

        Comment


        • #5
          Originally posted by Clyde Schechter View Post
          Well, that's an interesting way to think of it. But that's not what the term mixed effects refers to. It refers to the fact that the model as a whole contains both fixed and random effects. No individual effect is described as "mixed."
          thanks very much professor, I have one last question. How can I explain 0.11 to the reader in a general way?

          Comment


          • #6
            You can say that different schid's have different slopes for the education:culture relationship, and 0.11 is the mean of those slopes.

            Comment


            • #7
              Originally posted by Clyde Schechter View Post
              You can say that different schid's have different slopes for the education:culture relationship, and 0.11 is the mean of those slopes.
              thanks very much professor Schechter!

              Comment

              Working...
              X