View Single Post
  #80   ^
Old Tue, Jan-23-18, 05:43
inflammabl's Avatar
inflammabl inflammabl is offline
Senior Member
Posts: 2,371
 
Plan: Atkins
Stats: 296/220/205 Male 71 inches
BF:25%?
Progress: 84%
Location: Upstate SC
Default

Quote:
Originally Posted by cotonpal
The first is that when you do large observational studies the purpose should be to generate hypotheses for further testing using randomized controlled studies since the large observational studies cannot determine causation which is the purpose of the controlled experimentation.

Agree 8000%. Personally I am appalled at the lack of models which are used in every other discipline to relate the microscopic, i.e. what a particular enzyme, hormone, etc. does, to the macroscopic, i.e. weight, cholesterol, etc. going up and down. There was an article written in the WSJ a few years ago "Great Scientists Don't Need Math" which was a truly amazing justification of Biochemists ignorance of math. It stuns me that the author can still walk through the lab building with his chin up.

Quote:
She also said that scientists are supposed to be always looking for things that refute their hypotheses not simply looking for ways to confirm in. They are supposed to be diligently looking for where they might have got things wrong. When instead data that refutes their hypotheses are simply hidden or discarded that is not science that is bias.

Jean

Eh. How about this... "scientists are supposed to be always looking for things that refute other scientists' hypotheses not simply looking for ways to confirm in." And I would change this too, "When instead data that refutes their hypotheses are simply hidden or discarded that is not science that is lying." But they don't lie 99% of the time. Practically what they do is stop taking data if the data is taking them in a direction they don't like. For instance, if they've taken 11 samples and their data shows a low to medium confidence in an alternative hypothesis they don't take the next 17 samples which would give them high confidence in an alternative hypothesis. They honestly feel they are wasting their time. Honestly. Besides if there is no general model, if none has ever been created, then there is no general hypothesis to argue with anyway.
Reply With Quote