personally, I'm not comfortable leaving it to people's judgment to decide when they're fooling themselves and when they're not—not for experimentalists and not for modelers. the whole point of fooling one's self is that you don't know when you're doing it.
-
-
Replying to @talyarkoni @IrisVanRooij and
Yes and THAT is the fundamental source of our disagreement. You're saying that you don't trust us to do our science. And therefore you'd like to tell us how we should be doing it. But we know that it is easy to fool ourselves. That is why we use comp. models.
2 replies 2 retweets 11 likes -
Replying to @bradpwyble @talyarkoni and
I think we HAVE to let people make their own mistakes in science. Imposing strict rules and guidelines is always going to have a cost. The key is developing a way to find those mistakes before they become cultural traditions.
1 reply 1 retweet 4 likes -
Replying to @bradpwyble @IrisVanRooij and
why do we have to let them make mistakes we could easily avoid? seriously, what's so hard about omitting inferential statistics if you're not willing to explain what was exploratory and what was confirmatory? you can do it, why can't others?
1 reply 0 retweets 0 likes -
Replying to @talyarkoni @bradpwyble and
re: cultural traditions, I totally agree. but what exactly do you think is going to shift the tradition towards not reporting exploratory work as confirmatory, if not people like me pointing out why it's problematic? is it going to happen by itself?
1 reply 0 retweets 0 likes -
Replying to @talyarkoni @IrisVanRooij and
Fine, go ahead and do it in areas that you have expertise in. But when you come into our corner and try to tell us that what we're doing is wrong and we suggest that you misunderstand our methods, maybe give it some more consideration.
1 reply 3 retweets 9 likes -
Replying to @bradpwyble @IrisVanRooij and
that's fair; I promise I will do that. and in return, maybe you can give some more consideration to the possibility that some of the practices currently widespread in modeling are actually problematic from a scientific standpoint, as is true for many practices in other fields.
2 replies 0 retweets 1 like -
Replying to @talyarkoni @bradpwyble and
as always, I appreciate our discussions; I hope you find them useful too. I'm sorry we didn't come to an agreement in this case, but I think I understand your position better. gotta run now—have a good rest of the weekend!
1 reply 0 retweets 2 likes -
Replying to @talyarkoni @IrisVanRooij and
Believe me, I have taken your concerns seriously and this debate has given me some room to flex my understanding of the philosophy of science. I feel more well grounded in understanding my own intuitions than I did two days ago. I should go as well, house won't paint itself
1 reply 0 retweets 3 likes -
Replying to @bradpwyble @talyarkoni and
I'm surprised nobody brought up that modelers hate inferential stats (it's a really common private conversation between modelers) and they have them imposed on their work by non-modelers during peer review. So I'm hoping such conservations (if this is being read) can stop this.
1 reply 3 retweets 11 likes
Stop telling modelers to report irrelevant inferential stats in their papers that have nothing to add and only serve as a gatekeeping mechanism. We don't like it. We just do it because we're desperate to get published.
-
-
Replying to @o_guest @bradpwyble and
I've never done it myself, but it's so common. We all complain about and laugh about how non-modelers think try can tell us how to evaluate our models.
1 reply 1 retweet 4 likes -
Replying to @o_guest @bradpwyble and
As
@tom_hartley has already done a great thread on how marginalised we are, I won't go into it, but seriously... Not every modeler is able to tell reviewers "no, I won't add an irrelevant ANOVA". In this thread you're talking to people who can, but ECRs need their papers, etc.1 reply 4 retweets 8 likes - 1 more reply
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.