• commie@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    3
    ·
    1 year ago

    registered voters are not the same as likely voters, nor actual past voters. you made a claim that you simply can’t prove and none of the data you’ve provided is, in fact, proof for your claim.

    • silence7@slrpnk.netOPM
      link
      fedilink
      arrow-up
      3
      ·
      1 year ago

      Likely voter models don’t work well enough to look at 1-3% kinds of numbers of voters more than a year out from election day. Sorry.

      Using actual voters from 2020 is tough because we had two different third parties there: the Greens who siphoned votes off of Biden, and the Libertarians who siphoned a larger number of votes off of Trump. So you see polls showing the combined effect (slightly beneficial to Biden) but not the separate impact of the Green party candidate.

      Absolute proof isn’t something that really exists in the social sciences, which is why you’re never going to find it, the most you find is several decent converging lines of evidence, as we have here.

      • commie@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        2
        ·
        1 year ago

        the Greens who siphoned votes off of Biden and the Libertarians who siphoned a larger number of votes off of Trump.

        you can’t prove this at all. just because e those people did vote for libertarians or greens does t mean they would have voted for anyone else. in fact, given the option, they did NOT vote for someone else.

      • commie@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        2
        ·
        1 year ago

        Absolute proof isn’t something that really exists in the social sciences

        this platitude isn’t even true. lots of things can be proven false in social sciences. the fact that you are (quixotically) defending an unprovable hypothesis doesn’t mean there aren’t disprovable hypotheses which are possible.

      • commie@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        2
        ·
        1 year ago

        Likely voter models don’t work well enough to look at 1-3% kinds of numbers of voters more than a year out from election day. Sorry.

        your claim was about past elections. the data you provided was about a potential future election. you still don’t seem to be able to understand what was wrong with your claim.

          • commie@lemmy.dbzer0.com
            link
            fedilink
            English
            arrow-up
            1
            arrow-down
            2
            ·
            1 year ago

            it’s not the volume: it’s teh quality and relevance. you haven’t given me any relevant data to support your claim.