Florida is nearing toss-up status as poll shows Trump’s lead nearly vanished
www.independent.co.uk
Race and age of Florida respondents seemed to correlate with which candidate they supported
  • @TropicalDingdong
    16•2 months ago

    A bigger issue than MOE is structural bias.

    Here is FL 2020:

    Dem’s face a self-imposed structural disadvantage in both inter and intrastate models.

    R+3 in FL should be read more accurately as R+6 or R+7 based on the best most recent structural bias measurement we have. The article is weekend whacking material.

    • @[email protected]OP
      14•2 months ago

      That’s assuming the polling error goes the same way. That’s not a given at all especially as many pollsters have made methodology changes such as some doing much heavier rural sampling

      Polling error has gone both directions in the past. Dems were underestimated by polls in 2012 for instance

      • @TropicalDingdong
        1•2 months ago

        You should base it on the data we have. The data we have says the polling bias for FL leans +3-4 for Republicans.

        You dont get to just “wish” it were some other way and base expectations around that.

        • @[email protected]OP
          5•2 months ago

          Polling error has historically moved in inconsistent direction. Data goes back further than 2020. In 2012, Democrats were underestimated in florida by ~2 points. Romney was up 1.5% in Florida poll average vs Obama winning Florida by 0.9%

          Assuming it certain to go that way is not a given either. My point is that you cannot be certain about it

          • @TropicalDingdong
            -3•2 months ago

            My point is that you cannot be certain about it

            Yeah and thats not really a point. Everything has uncertainty. We have to and do make judgements in the face of uncertainty of reality all the time.

            If you choose to live in a fact based reality rather, this is the thing we have.

            • @[email protected]OP
              3•2 months ago

              That’s not how your earlier comments are phrased. The earlier comments declare that this is a given structural bias and that it will always exist. How is entirely ignoring the 2012 election any more real than saying we can’t be sure?

              • @TropicalDingdong
                0•2 months ago

                declare that this is a given structural bias and that it will always exist

                You just lack reading comprehension.

                The quote is:

                best most recent structural bias measurement

                The previous comments said, “the best most recent estimate of structural bias”, which was Trump v Biden 2020. Its the best because its not a simulation or modeling. Its two measured values. I’ve seen simulations and statistical models to estimate things like structural bias, but none of them are as good as a measurement. We should use the measurement.

                I get it. You’ve got an axe to grind. And at this point you might be better off inside a warm fantastic cocoon where Harris is crushing it and is going to win FL and TX. It might be the last light of joy you get to experience.

                • @[email protected]OP
                  2•2 months ago

                  My response was more so to the “you don’t get to ‘wish’” part. It could go the same way, it could not. It’s not consistent year to year. Assuming it is when long term data does not support that, isn’t helpful

                  Over the long term, there is no meaningful partisan statistical bias in polling. All the polls in our data set combine for a weighted average statistical bias of 0.3 points toward Democrats. Individual election cycles can have more significant biases — and, importantly, it usually runs in the same direction for every office — but there is no pattern from year to year

                  https://fivethirtyeight.com/features/weve-updated-our-pollster-ratings-ahead-of-the-2020-general-election/

                  No where am I claiming that Harris definitely will necessarily be underestimated, I am saying it is possible. Or perhaps even just underestimated by less. Dismissing the possibility out of hand by N=1 is what I am responding to