• breakfastmtn@lemmy.ca
        link
        fedilink
        arrow-up
        0
        ·
        10 months ago

        1500-2000 is fine. They use random sampling and weighting. A larger sample size would decrease the margin of error but wouldn’t significantly alter the results. A margin of error of +/-3.1% is fine for this kind of snapshot of public opinion.

      • Poggervania@kbin.social
        link
        fedilink
        arrow-up
        0
        ·
        10 months ago

        Considering we have a population of 331,464,948, I would say it’s not lol

        I personally think we’re committing genocide, but I would imagine a sample size that is under 1% of the US population is small

        • Deceptichum@kbin.social
          link
          fedilink
          arrow-up
          0
          ·
          10 months ago

          A sample size of 1,659 out of a population size of 334,914,895, with a population proportion of 50% and confidence level of 95% gives an error margin of 2.41%.

          1,659 is a fine sample size.

            • sincle354@kbin.social
              link
              fedilink
              arrow-up
              0
              ·
              10 months ago

              If there’s anything I would not fault somebody about, it’s misunderstanding sampling in a statistical sense. Just about the biggest case of “plug into formula and pray” of all of my classes I took. I’m looking back and it’s just Greek letters to me now…

              • FaceDeer@kbin.social
                link
                fedilink
                arrow-up
                0
                ·
                10 months ago

                Oh yes, it’s not meant as a slight. It’s just something that humans need to be aware of. Just like how we need a forklift to move a one-ton object, we need to run calculations and trust the math when doing statistics.