• I'm back on my BS 🤪@lemmy.autism.place
    link
    fedilink
    English
    arrow-up
    20
    ·
    3 months ago

    what‽ how is the student t distribution more important than the normal distribution?? you can’t even use the t unless you’ve confirmed that you’ve got a normal! 📈📉

    • denkrishna@midwest.social
      link
      fedilink
      English
      arrow-up
      7
      ·
      3 months ago

      Correct me if I’m wrong but isn’t the student t distribution a set of distributions that includes the normal distribution?

      Because if so, it feels a little like saying “you can’t even call something red unless you’ve confirmed that it’s crimson”

      • Bender_on_Fire@lemmy.world
        link
        fedilink
        English
        arrow-up
        11
        ·
        3 months ago

        The t-distribution approaches the normal distribution with increasing degrees of freedom. It is certainly more relevant in for example hypothesis testing, since t-Tests (variance is estimated from the data) is much more common than z-tests (variance is treated as fixed and coming from a normal distribution).

        In all of statistics or probability theory, the normal theory is however way more influential.

        Nonetheless, it’s a cool bit of history where modern statistics got its roots. As a lover of both statistics and guinness, i approve!🍻

      • driving_crooner@lemmy.eco.br
        link
        fedilink
        English
        arrow-up
        3
        ·
        3 months ago

        The t-student goes to the normal when your degrees of freedom get close to infinitum (in practice with 30 df they’re practically the same).

    • driving_crooner@lemmy.eco.br
      link
      fedilink
      English
      arrow-up
      3
      ·
      edit-2
      3 months ago

      I argue that is more important because in practice you usually don’t have enough samples (or can’t resample) to use the normal distribution.