Announcement

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

  • What could have I possibly done wrong?

    Hello,
    I just did a regression analysis and displayed a scatterplot and the points are I can't see no pattern for nearly all the variables.
    Maybe its just like that but I have the strong feeling that I did something wrong..
    Do you have any idea what could possibly cause that?

  • #2
    Hanna:
    your only, inadvertent mistake was to forget to post what you typed and what Stata gave you back via CODE delimiters.
    In any quantitative matter, codes and numbers matter much more than tons of words aimed at describing what's going on.
    Kind regards,
    Carlo
    (Stata 18.0 SE)

    Comment


    • #3
      That is the scatterplot:
      Attached Files

      Comment


      • #4
        Hanna:
        can you please provide an example/excerpt of your data via -dataex-? Thanks.
        Kind regards,
        Carlo
        (Stata 18.0 SE)

        Comment


        • #5
          At that is the regression I did:


          Click image for larger version

Name:	Regression.png
Views:	2
Size:	35.7 KB
ID:	1478327
          Attached Files

          Comment


          • #6
            Sorry, what is dataex? (I'm pretty new using stata)

            Comment


            • #7
              Ups, the scatterplot I posted was for an other dependent variable, but for V130 it looks pretty much the same.

              Comment


              • #8
                Please type -help dataex- from within Stata and get familiar with this easy and useful command.
                Kind regards,
                Carlo
                (Stata 18.0 SE)

                Comment


                • #9
                  I downloaded the package but I don't really get how to use it:/ Is there somewhere an introduction?

                  Comment


                  • #10
                    Please see example at https://www.statalist.org/forums/for...logit-fe/page2, #17.
                    Kind regards,
                    Carlo
                    (Stata 18.0 SE)

                    Comment


                    • #11
                      Is there an introduction to dataex? Yes; as Carlo Lazzaro already mentioned

                      Code:
                      help dataex
                      is not only an introduction but as much as anyone needs to know.

                      But the scatter plot is fairly easy to interpret.

                      Your y axis variable is an integer 1 to 10, as you explained in a different thread.

                      Your x axis variable is basically 0 to 12 divided by 12.

                      The scatter plot shows that almost all of the joint possibilities for those variables exist in your data. The scatter plot tells you nothing about how often they occur.

                      Focus on the observations with say 1 on the y variable and 0 on the x variable. Clearly they all plot at the same position on the scatter plot. Either you need a different kind of plot to show frequencies too, or a table will be more informative.

                      The regression results imply a large sample but weak relationships. Even experts in your kind of social science can't easily give extra advice as only one predictor name is informative in any sense.

                      To get better advice, you need to say something more about the data.

                      Note that regression here is treating your response as interval scale and ignoring its bounds.

                      PS Your title really isn't informative. Please note advice in the FAQ on better question titles, giving data examples, etc.

                      Comment


                      • #12
                        Carlo Lazzaro
                        is there an other way to export it than safe it as a picture?

                        Comment


                        • #13
                          Hanna:
                          yes, and -dataex- is the answer.
                          Kind regards,
                          Carlo
                          (Stata 18.0 SE)

                          Comment


                          • #14
                            Did I understood it correct that you are basically saying that there might be a better way to display the scatterplot but there is not really an effect to see anyways?I also tried if a ordered logistic regression fits better but there wasn't much of a difference. Do you think that would make more sense?
                            And the other variables are things like, income, education social trust, religiosity etc. I'm really surprised why there seems to be no effect at all.

                            Comment


                            • #15
                              Carlo LazzaroI still don't get how to export it. Is there an other command? Or do I have to type in -dataex- somewhere here?

                              Comment

                              Working...
                              X