Oh dammit I forgot 4). Hmm. This is such a big and important question and I should have some ready answer for it.
I suppose my most general answer, and it’s not all that recent, is that I’ve become MUCH less trusting of the scientific literature in loads of fields, especially social sciences, because I’ve become much more aware of the statistical problems. But that’s a bit of a dodge, isn’t it.
And relatedly I think in Covid times I’ve become less happy with the public-health-institutions model of scientific/health evidence, of thinking “there isn’t an RCT supporting it” equals “it doesn’t work”; I’ve become much more of a Bayesian, or at least I try to think in terms of probabilities and best guesses and cost-benefit analyses rather than “this works” and “this has not been rigorously shown to work ergo we will say it doesn’t work”. I was already on that route I think but it’s become very obvious in the last year
Oh dammit I forgot 4). Hmm. This is such a big and important question and I should have some ready answer for it.
I suppose my most general answer, and it’s not all that recent, is that I’ve become MUCH less trusting of the scientific literature in loads of fields, especially social sciences, because I’ve become much more aware of the statistical problems. But that’s a bit of a dodge, isn’t it.
And relatedly I think in Covid times I’ve become less happy with the public-health-institutions model of scientific/health evidence, of thinking “there isn’t an RCT supporting it” equals “it doesn’t work”; I’ve become much more of a Bayesian, or at least I try to think in terms of probabilities and best guesses and cost-benefit analyses rather than “this works” and “this has not been rigorously shown to work ergo we will say it doesn’t work”. I was already on that route I think but it’s become very obvious in the last year