• warm@kbin.earth
    link
    fedilink
    arrow-up
    25
    arrow-down
    3
    ·
    8 天前

    1000 people is such an irrelevant data set.

    16-34 is such a massive age range, a 34 year old is completely different to a 16 year old.

    Where was the poll even posted? Media like this is what’s causing our problems. It’s just cherry picking for headlines.

      • cynar@lemmy.world
        link
        fedilink
        English
        arrow-up
        17
        ·
        8 天前

        For random bias yes. Unfortunately, that age group are notoriously difficult to poll accurately. Without knowing how the polling was done, detecting systematic bias is difficult.

        E.g. street polling on a Sunday morning, near a church, or mosque will get vastly different results to the same poll near the high street, on a Friday night.

        Also relevant. https://xkcd.com/882/

        Don’t get me wrong, it’s a worrying poll. Buta lot of studies like this disappear back into the noise, when studied by multiple groups.

        • Bad_Engineering@fedia.io
          link
          fedilink
          arrow-up
          1
          ·
          8 天前

          The study says they used online polling but didn’t elaborate further so who’s to say what biases there may or may not be. But generally ipsos does a pretty good job with their polling.

          I do agree that most of these polls tend to end up being nothing burgers, but there has been a recent trend of younger males gravitating towords the right of the political spectrum. I mainly commented to point out that a sample size of 1000 is generally recognized to be large enough for accuracy.