Suppose that people constantly get soaked by rain. Then, somebody invents the umbrella. A few people start using it, and it seems to work. Now you can complain that the umbrella isn't perfect in particular ways. But it sure is better than being soaking wet.
-
-
Replying to @EJWagenmakers @richarddmorey
I do not think this is the point. The point is about culture: How umbrella advocates (UA) think everyone should use umbrellas, even when other researchers (OR) are inside a building and dry. They have a roof and doing great work on their computers.
2 replies 0 retweets 6 likes -
OR say they’re fine without umbrellas for much of their work. Umbrellas make sense when go outside & it is raining. But UA better not push OR to go outside just so they can use umbrellas or push them to use umbrellas inside just in case, ...
1 reply 0 retweets 4 likes -
... or say that if OR do not go in an umbrella-needed situation then they are not doing good science. (Btw, I do not think this culture problem is specific to preregistration, but pervades psych science and open science also for other types of tools.)
2 replies 1 retweet 4 likes -
The only way IMHO the umbrella metaphor has any value here is if you also include the fact the "umbrella" is causing a high incidence of injuries to other people while also keeping the user dry.
1 reply 0 retweets 1 like -
Replying to @o_guest @IrisVanRooij and
To wit, something that might work well for one person might not for others. We live in a world where things we do to improve our side of things has to be balanced against how it affects others.
1 reply 0 retweets 1 like -
Sure. But if you present a p-value it only has meaning in a confirmatory setting. Exploratory research introduces a multiple comparisons problem with the number of comparisons unknown. And I believe there are problems for Bayesians too.
1 reply 0 retweets 0 likes -
I don't really do that kind of research, so I can't really comment on it in neither good faith nor with the required practical experience. I don't run analyses with p-values nor with Bayes factors, for example.
1 reply 0 retweets 1 like -
Replying to @o_guest @EJWagenmakers and
While I have opinions (of course), I also don't think it's my place to tell you how do your science. I am not a big fan of prescriptivism — in most cases, I think it's misplaced. And I don't think science is a monolith.
1 reply 1 retweet 2 likes -
Replying to @o_guest @EJWagenmakers and
I know you haven't said this here yourself, but I wanted to super clarify where I am coming from. I think it's a tricky situation where there is a culture clash on quite a few different levels.
1 reply 0 retweets 0 likes
I'm both a comp modeller (who doesn't care — on some level — what stats people do) and I do not believe in telling people how to do their science generally (so I don't think it's useful to dictate in a very rigidi way to others how to do stuff).
-
-
Replying to @o_guest @EJWagenmakers and
I'm way more interested in stopping the prescriptivist culture, than I am in p-values or BFs. I feel like people telling others how to do research that the former don't even do, and are not experts in, is a strange and perhaps slippery slope.
2 replies 1 retweet 5 likes -
Replying to @o_guest @EJWagenmakers and
Hope that helps clarify my position. I think yours is that you are very interested in having a "what methods are appropriate" debate for the stats people use on their empirical data. And I don't really have any skin in that game.
0 replies 0 retweets 0 likes
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.