Stata - How to standardize your variables?

Поділитися
Вставка
  • Опубліковано 11 гру 2024

КОМЕНТАРІ • 22

  • @lukegardner4733
    @lukegardner4733 2 роки тому +1

    you can just include ,beta in your regression to output the beta or "standardized coefficients"

    • @SteffensClassroom
      @SteffensClassroom  2 роки тому

      Haha, guess you learn every day!
      Just a few small differences: You don't obtain the standard deviation of the coefficient this way (if you need it), and you still have to calculate them manually if you wish to get the summary statistics for the standardized variables. However, if all you need is just to produce the standardized coefficients, then this works perfectly. Thank you :)

  • @felicityerrington6879
    @felicityerrington6879 Місяць тому

    thank you so much for this, this was incredibly helpful!!

  • @mast007Er
    @mast007Er 2 роки тому +2

    Thank you, very well explained and useful for my Stata university project where I'm standardising an income variable :)

  • @sachursunny5974
    @sachursunny5974 2 роки тому +2

    I was really confused seeing the mean being non-zero. Thank you for clearing it.

  • @isabelm.6640
    @isabelm.6640 9 місяців тому +1

    Again very helpful! 👏

  • @arshadkhan-cy1do
    @arshadkhan-cy1do Рік тому

    excellent sir

  • @aronibarkat8384
    @aronibarkat8384 Рік тому

    thanks! great help

  • @sairamaroof5928
    @sairamaroof5928 6 місяців тому

    Sir, I have the data of six social media platforms from 2009 to 22, so how can I create an index for each country separately or combine the countries, which are 21 countries, so they are aggregated. Also tell me how to do normalization of data

    • @SteffensClassroom
      @SteffensClassroom  6 місяців тому

      This video is about how to standardize your variables. Other inquiries can be mailed, where a session can be booked.

  • @NathanCats777
    @NathanCats777 Рік тому +1

    What if I want to standardize data on a 0-1 scale?

    • @SteffensClassroom
      @SteffensClassroom  Рік тому +1

      Hello!
      I would try something like:
      sum var replace var = (var - `r(min)') / (`r(max)'-`r(min)')
      You can extend this to a loop if you need to do this for many variables at once.

    • @NathanCats777
      @NathanCats777 Рік тому +1

      @@SteffensClassroom Thanks a lot! Very useful. On the same note, I am trying to rescale a 1 to 7 variable in an 11 point scale (0-10), plus I would like to reverse the variable so that the higher values become the lowest. Do you have any advise on that? I guess I should first subtract 1 from the variable, but then how do I "stretch" it to 10?

    • @SteffensClassroom
      @SteffensClassroom  Рік тому

      @@NathanCats777 re-scaling: use basic math ;)
      For the reverse order, there is a Stata command called revrs I believe.

    • @NathanCats777
      @NathanCats777 Рік тому +1

      @@SteffensClassroom Thanks a lot, actually revrs command did not work, but I managed to do it by abs(r(max)-var) and it worked

    • @SteffensClassroom
      @SteffensClassroom  Рік тому +1

      @@NathanCats777 Happy it worked out :)

  • @mahdouch_Mahdi_FCB
    @mahdouch_Mahdi_FCB 2 роки тому

    Is sd can help to reduce non normality distributed data ??? Thank you

    • @SteffensClassroom
      @SteffensClassroom  2 роки тому +1

      Hi! Thanks for your question. In short: no. Standardization shifts and scales a given distribution. It does however not change its shape.

    • @mahdouch_Mahdi_FCB
      @mahdouch_Mahdi_FCB 2 роки тому

      @@SteffensClassroom Thank you for your response, can you help me with any solutions to transform data to normal distribution?

    • @SteffensClassroom
      @SteffensClassroom  2 роки тому

      Hi again,
      Please mail me. Then we can discuss this further.