• Womble@piefed.world
      link
      fedilink
      English
      arrow-up
      8
      arrow-down
      1
      ·
      edit-2
      1 day ago

      I wish all the people who harp on about sample size would at least google it first before commenting that they know better than academics whose job this is.

      • misk@piefed.social
        link
        fedilink
        English
        arrow-up
        8
        arrow-down
        1
        ·
        16 hours ago

        For a binary choice where you expect distribution to be relatively similar you can do it with less than 400 while having 95% confidence and 5% margin of error. That’s just statistics and probability.

          • misk@piefed.social
            link
            fedilink
            English
            arrow-up
            5
            arrow-down
            1
            ·
            16 hours ago
            1. We are machines made out of meat.
            2. It’s irrelevant if the choice was made at random or if it was based on some deep spiritual insight. We’re measuring outcomes, not the rationale.