Timeline for How can I avoid being "the negative one" when giving feedback on statistics?
Current License: CC BY-SA 3.0
8 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 18, 2017 at 15:25 | comment | added | zelanix | As a "data scientist", this answer did make me chuckle. | |
| Jan 17, 2017 at 15:18 | comment | added | dtldarek | @Jeff I stand corrected. I knew "pedant" is a pejorative term, but understood it as "excessively concerned with formalism, accuracy, and precision" (quoted from Wikipedia) and that was mine intended meaning (I did mention not being a jerk). Only now I have checked there is another meaning which is probably the one you intended (one who makes an ostentatious and arrogant show of learning; also from Wikipedia, but lots of dictionaries agree). | |
| Jan 17, 2017 at 14:57 | comment | added | Jeff | @dtldarek Pedantic is a pejorative term. It implies, for example, arrogance or superiority in your approach. | |
| Jan 17, 2017 at 10:46 | comment | added | dtldarek | @Jeff Could you give an example that makes the distinction between being pedantic and being thorough? One I can think of is whether to check validity of some widely accepted work that we are basing on, but in some cases even being thorough would mean going deep into that too, so I am not sure. In particular, if that work makes some implicit assumption (e.g. otherwise some theorems authors use do not work), then we should include that assumption in our paper (e.g. similarly to [42], for Theorem 7 to work we assume that XYZ). If you don't then you get just a bunch of false results. | |
| Jan 17, 2017 at 9:05 | comment | added | Jeff | @dtldarek There's a difference between being pedantic and being thorough and correct in your work. And since the OP's question was about how to best engage with his colleagues, it's an important difference. | |
| Jan 16, 2017 at 22:13 | comment | added | Pete L. Clark | I think this is a good answer: in practice, whether the data justifies the conclusion has a strong academic-cultural component. This means that you can come off as an ignorant jerk if you just walk in the room and assume you're right and everyone else is wrong. Of course, whether the data justifies the conclusions is also a statistical issue. So I think the OP's goal should be to try to contribute to the discussion of the validity of the results in a useful way. (It seems unlikely that the role will be to convince the rest of the group to go back to the drawing board...) | |
| Jan 16, 2017 at 21:39 | comment | added | dtldarek | Being a complete pedant is actually the right thing to do, esp. in statistics. That does not mean being a complete jerk. It means that any additional assumptions should be stated and only claims that really follow from data and used theorems are valid. Even that one slight generalization, which sounds really reasonable can be only a hypothesis or potential explanation (which is still worth including in a paper), but not a result. | |
| Jan 16, 2017 at 20:52 | history | answered | Jeff | CC BY-SA 3.0 |