We need to stop pretending models are reality.

in #covid4 years ago

image.png

Dozens of government officials and media talking heads use phrasing like “we must”, ”if we don’t”, ”people who do/don’t” to precede other phrases like “lock down”, ”wear a mask”, ”restrict xxxx activity,” as though statistical modelling has made those statements foregone conclusions.

The entire country has, by force of law, changed most of its economic behavior based on various models. I can think of one or two executive leaders who have publicly stated they don’t trust models and even they have made some policy decisions based on them.

I have conversations with people nearly every single day, who cite models as if they are factual reality -- and I've had these conversations nearly every day for 4 months straight.

The entire political class is also invested in this narrative, using models to justify sweeping policy decisions as if the models were, themselves, truth.

Even as we have seen that these models have been ludicrous failures when stacked up against real world observations, people are still referring to the models as if they have always been inherently accurate. If you look at many of the "studies" that purport to show that lockdowns saved hundreds of thousands or even millions of lives, you'll see that they rely on these models to make those claims -- which itself is a form of circular reasoning... question begging.

This seems totally insane to me, intellectually.

These kinds of studies are starting from the premise that the models were/are fact and that the disparate results we've seen in reality are only attributable to lockdowns -- but as Sweden's example fairly clearly shows, that's almost certainly not true. To the contrary, what is far more likely is that the models themselves were radically off the mark.

But of course, since the confidence intervals of these models have been massive, virtually any result could fall "inside" the range, which has generally made them useless for that reason.

And yet, they've been the core of why the policy shift was so severe most everywhere around the world.

So yeah. People are taking these models as if they are fact.