Someone writes whether the data actually shows what it purports to show: https://statmodeling.stat.columbia.edu/2020/01/09/filters-be...
> The whole thing is driven by one data point and a linear trend which makes no theoretical sense in the context of the paper (from the abstract: “Air testing conducted inside schools during the leak (but before air filters were installed) showed no presence of natural gas pollutants, implying that the effectiveness of air filters came from removing common air pollutants”) but does serve to create a background trend to allow a big discontinuity with some statistical significance.
I’m reminded of the walkback of scientific studies showing massive benefits from giving kids in third world countries deworming medications: https://www.theguardian.com/society/2015/jul/23/research-glo....
My beef with Matt Y.’s worldview of “scientifically driven public policy” is that the costs and benefits of public policy interventions are so devilishly difficult to study that you can’t meaningfully use them on realistic time scales to drive policy. This is an exceedingly simple hypothesis—filtering air improves test scores—that can easily be tested while controlling for confounding factors. But even then it’s hard!