Announcement

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

  • Binary Variable - Correlation and Scatterplot

    Dear all,

    i am estimating a fixed effects model on work performance as dependent variable. One of my explaining variables will be a binary indicator for sex 1=male == female.

    Before running the regression i would like to estimate visually, if there is an effect at all.
    Therefore I want to draw a scatterplot of sex vs performance.

    Since performance is a continuos indicator and sex binary the usual
    Code:
     scatter performance sex
    does not deliver a helpful output, as only points at 0 and 1 are rising.

    Does anyone know whether stata has a command for this issue.

    Many thanks,

    Regards, Alexander-Florian


  • #2
    If you want to visualize the distributions of performance in each sex, side by side, a simple way is

    Code:
    dotplot performance, over(sex)

    Comment


    • #3
      The dotplot command that Clyde suggests is the way I would approach it, but others might recommend comparing the empirical cumululative distribution plots. These are farily easy to do from scratch in Stata (and here), but there is at least one user-written command that might make this approach more accessible if you want to explore it, too.

      Comment


      • #4
        See also (e.g.) stripplot (SSC) (in practice a superset of dotplot), distplot (SJ) for cumulative distribution plots, qplot (SJ) for quantile plots.

        Comment


        • #5
          Clyde, Joseph, Nick,

          many thanks for your very helpful reply!

          Cheers, A.

          Comment

          Working...
          X