• @Eiri
    link
    33 months ago

    It’s just one poll. There’s a possibility that pollster’s methods have some sort of bias, or that there happened to be an error in sampling as there’s always a possibility there will be.

    There are statistical models that take in various polls, remembering and taking into account the kinds of mistakes each given pollster tends to make, and calculates the chances of each candidate. That would be a (probably) more accurate look at Harris’s chances.

    The only one I know is Qc125/338Canada. It was developed mostly in Quebec and then fine-tuned for other Canadian races, before finally being applied to a few other countries, including the US. That history might make it weaker at predicting the outcome of an American race compared to a home-grown model, though; I don’t really know.

    • @[email protected]
      link
      fedilink
      23 months ago

      Democrats have won some major statewide elections in recent years in Georgia, so no, it’s not necessarily an outlier.

      • @Eiri
        link
        13 months ago

        Oh that wasn’t what I meant. Independently of where this specific poll lies, it’s just that any single poll is just a limited dataset.

        I was just recommending using broader analyses as a potentially more accurate election result predictor.