• Jtotheb@lemmy.world
    link
    fedilink
    English
    arrow-up
    2
    ·
    17 hours ago

    What’s neat about an open poll like the one they ran is that it includes most of the people who care about weighing in. Maybe you can read an AI summary about statistically relevant sample sizes? It might be more useful to learn about the math, or to skip the summary and find a useful website, like this calculator from Qualtrics that will show you that a sample size of 175,000 is more than 10 times what you’d need to achieve a roughly 99% level of confidence with a 1% margin of error.

    • Fizz@lemmy.nz
      link
      fedilink
      English
      arrow-up
      1
      ·
      5 hours ago

      Yes that is a good sample size but its biased by the people weighing in caring most about the issue. If it was 175000 users sampled randomly from the DDG user base then yes it would be a good predictor but it wasnt.