Monday, October 09, 2023


Brian's missing a piece here, which is that fake empiricism lets bad faith actors smuggle in bullshit and call it SCIENCE.
But along the way somewhere it stopped being just a useful cudgel and came to dominate how Democrats and liberal elites think about everything, including politics. It’s all science, and we trust science. But, 1) no it’s not; and 2) it’s hard to be a spontaneous, confident politician if you believe there’s a real science to campaigning and you should never toss anything outside the formula into the mix.

The advent of poll aggregating, and the forecasting successes of Nate Silver, were benign precursors to something more insidious. On its own, poll aggregating amounted to a huge improvement over an earlier world of sweeping media narratives stitched together with yard signs, crowd sizes, and one-off surveys, whether or not they were outliers. The obsession with polling that followed the 538 boom has not been particularly healthy. It’s made polls and media about polls endogenous to politics, such that polls are often the entire basis for media narratives, and those same media narratives become drivers of subsequent poll findings. Republicans have thus taken to flooding aggregators with junk polls in uncompetitive races, in the hope of manipulating media narratives about national elections. Polling can now fulfill its own prophecies.

But it’s also fed a sense within the Democratic Party that data science, as applied to politics, is very precise and can more or less tell candidates and campaigns what to do and how to talk. That Democrats should do and say nothing of consequence until it has been surveyed and run by focus groups.
Republicans were flooding the press with junk election polls, while "progressive" polling firms were flooding party leaders with junk issue polls, or at least junk unsupported conclusions from them.