2 ms·
>But that doesn't answer my original question. The problem here being referred to was people disproving the null hypothesis to indirectly validate the alternati
by eli_gottlieb 12y ago
>But that doesn't answer my original question. The problem here being referred to was people disproving the null hypothesis to indirectly validate the alternative hypothesis, right?
>
>Is that still Popperianism? Because as far as I understand it, Popper only tells to keep trying to disprove the actual hypothesis/hypotheses.
I actually did answer, but it was kind of buried. Yes, using p-values to disprove the null hypothesis is Popperianism, at least insofar as it holds that you can't ever really provide quantitative validation to your alternative hypothesis and have to sort of wave your hands at its non-falsification as if that meant something. It seems almost but not quite like positivism because Popperians desperately want to validate things but are committed to an ideology telling them it's categorically impossible to do so. So they conjecture really fervently instead.
If this all sounds kinda stupid, well, I did give a whole rant on Bayesianism, which does let us quantify validation of hypotheses directly.
- sa1 12y agoWell, even Bayesian inference effectively leads to a similar problem - We have x% confidence in a theorem because it has y% simplicity. Does it take away the need to conjecture new theories which may turn out to be simpler? Thanks for continuing to answer. I am learning new stuff.