Conversation
Replying to
✊
Quote Tweet
the biggest problem in both statistical practice and criticism of statistical practice is braindead insistence on following form rather than consideration of whether adherence to form is sufficient to produce the desired insight or necessary to produce any insight respectively
Show this thread
2
48
Show replies
Replying to
Welcome to every time a study gets posted on Reddit. A bunch of armchair statisticians come out of the woodwork
5
Replying to
Side note but being a good data scientist is really not abt showing off and 'rotating things in your head'. Its abt understanding which method applies where, what are the potential caveats and how they can be corrected for. Its an endless journey
1
7
Replying to
I think a lot of this comes from how analysing the trustworthiness of a source is taught in entry-level science.
There's a lot of "is this peer-reviewed", "is this cited by other people", etc etc. A lot of yes/no keys of credibility, rather than partial updating 1/
1
2
Replying to
I disagree with some of this, for reasons well articulated in this blog post
3
3
Replying to
Not only is limited data better than no data, it's also the best anyone will ever actually get in practice. Turns out there's a whole branch of mathematics founded to deal with this problem...
3
