It is well understood that even skilled researchers regularly misinterpret p-values in all sorts of ways, consisting of confusion of practical and statistical significance, dealing with non-rejection as approval of the null hypothesis, and interpreting the p-value as some sort of replication probability or as the posterior likelihood that the null hypothesis is real …
It is shocking stunning these errors mistakes appear hard-wired into statisticians thinking, and this suggests that our profession really truly requires look at how it teaches the interpretation analysis statistical analyticalReasonings We presume that, to make progress in pedagogy, statisticians will have to provide up some of the claims we have implicitly been making about the effectiveness of our methods …
It would be nice great the statistics profession occupation offering a good excellent service the significance testing problem issue we just needed required convey communicate more clearly. Rather, we have to accept that we live in a world penetrated by authentic unpredictability and that it takes a lot of variation to make great inductive inferences.
Running 10 replicative experiments does not make you as sure of your inductions as when running 10 000 differed experiments– even if the possibility worths are the very same.
To Keynes, expectations are a concern of weighing possibilities by degrees of belief, beliefs that typically have preciously little to do with the kind of stochastic probabilistic estimations made by the logical agents as designed by contemporary social sciences.
from Lars Syll
It is well understood that even knowledgeable researchers routinely misinterpret p-values in all sorts of ways, including confusion of useful and statistical significance, dealing with non-rejection as approval of the null hypothesis, and interpreting the p-value as some sort of duplication possibility or as the posterior possibility that the null hypothesis is true …
It is shocking stunning these errors seem appear hard-wired into statisticians thinking, and this suggests recommends our profession occupation actually to look at how it teaches the interpretation analysis statistical analyticalReasonings We presume that, to make progress in pedagogy, statisticians will have to offer up some of the claims we have actually implicitly been making about the effectiveness of our approaches …
It would be nice great the statistics data occupation offering a good solution option the significance testing screening and we just needed to convey it more clearly. To put it another way, its not that were teaching the right thing inadequately; regrettably, weve been teaching the incorrect thing all too well.
Andrew Gelman & & John Carlin
Teaching both economics and stats, yours genuinely cant however see that the statements “give up a few of the claims we have implicitly been making about the efficiency of our techniques” and ” its not that were teaching the ideal thing inadequately; sadly, weve been teaching the wrong thing all too well” obviously use not just to data …
And the service? Not– as Gelman and Carlin also highlight– to reform p-values. Instead, we need to accept that we live in a world penetrated by genuine uncertainty which it takes a lot of variation to make great inductive inferences.
Sounds familiar? It certainly should!
The standard view in stats– and the axiomatic probability theory underlying it– is to a large level based upon the rather simple idea that more is much better. As Keynes argues in his seminal A Treatise on Probability ( 1921 ), more of the same is not what is essential when making inductive reasonings. Its a question of more however various– i.e., variation.
y & & w )does not make w irrelevant. , knowing that the probability is unchanged y & & w) another evidential weight ( weight of argument). Running 10 replicative experiments does not make you as sure of your inductions as when running 10 000 varied experiments– even if the probability values are the exact same.
Keynes rather thinks that we base our expectations on the self-confidence or weight we put on different occasions and alternatives. To Keynes, expectations are a concern of weighing possibilities by degrees of belief, beliefs that often have preciously little to do with the kind of stochastic probabilistic estimations made by the reasonable agents as designed by modern-day social sciences.
Like this: Like Loading …
