• Zaktor@sopuli.xyz
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    1
    ·
    1 year ago

    Unless you want to claim the type of Gen Z who would answer a phone poll would bias the results this is irrelevant.

    • Rapidcreek@reddthat.com
      link
      fedilink
      arrow-up
      3
      arrow-down
      5
      ·
      1 year ago

      Controlling the sample and excluding parts of the sample very much screws the results of a poll.

      • Zaktor@sopuli.xyz
        link
        fedilink
        English
        arrow-up
        2
        arrow-down
        2
        ·
        1 year ago

        That’s not how sampling works. Unless the excluded group is biased toward a result it doesn’t change how representative it is. Are you claiming that the type of Gen Z who would answer a call is biased for or against Biden?

        • Rapidcreek@reddthat.com
          link
          fedilink
          arrow-up
          2
          arrow-down
          4
          ·
          1 year ago

          You don’t have the ability to determine bias in a straight poll, yet if you exclude 9 of 10 on the sample you’ve created a bias because they are not representative.

          • Zaktor@sopuli.xyz
            link
            fedilink
            English
            arrow-up
            3
            arrow-down
            2
            ·
            1 year ago

            That’s again not how sampling bias works. If you called these people and asked what their favorite color was, you wouldn’t have an unrepresentative sample. It’s only if your sample is itself biased toward a response. Very small samples might have larger error bars, but they aren’t inherently unrepresentative unless the method of sampling is corelated with a response.