• defunct_punk@lemmy.world
    link
    fedilink
    arrow-up
    7
    arrow-down
    5
    ·
    edit-2
    1 year ago

    those sample sizes are of zero scientific value.

    Tell me you never took statistics without telling me you never took statistics. A sample size of ~1,100 randomly selected people is a statistically accurate sample size for the US population likely voters. Of that, the 18-29 YO demographic is accurately represented by the 179/168 surveyed.

    • Hypx@kbin.social
      link
      fedilink
      arrow-up
      4
      arrow-down
      2
      ·
      1 year ago

      The main issue is that no one can generate random samples anymore. Not via landlines anyways. The problem is that scientific polling is nearly impossible now.

    • Heresy_generator@kbin.social
      link
      fedilink
      arrow-up
      2
      arrow-down
      4
      ·
      edit-2
      1 year ago

      Neither 179 nor 168 is enough of a sample size to draw conclusions; if you don’t understand that then your statistics professor completely failed.

      The total sample doesn’t matter when you’re only pulling out a small part of it. If I polled 10,000 people but only 20 of them were over 100 I can’t go around saying shit like “A majority of people over 100 say” this or that.