As De Boer puts it, if there's any possibility at all that an education result came from sample bias, it's definitely sample bias. The standard narrative for "

As De Boer puts it, if there's any possibility at all that an education result came from sample bias, it's definitely sample bias. The standard narrative for "better schools" is that a trial program works great, everyone gets excited, and then it declines to mean the instant it's expanded. If they had lotteried people in entirely at random and still gotten better outcomes, I'd at least be curious. https://fredrikdeboer.com/2017/03/29/why-selection-bias-is-the-most-powerful-force-in-education/ https://fredrikdeboer.com/2017/03/29/why-selection-bias-is-t...