Andrew gives excellent advice about how to write a scientific paper. This is his best point:
Consider Table 2. Do you want the reader to know that in line 3, Min Obs is 894? I doubt it. If so, you should make a case for this. If not, don’t put it down. When an article is filled with numbers and words that you neither expect or want people to read, this distracts them from the content.
In other words, most tables should be figures or omitted. I would add a broader point: Don’t try to impress anyone. It gets in the way of helping them — helping them understand what you’re saying. (The classic example is B. F. Skinner, apparently insecure Harvard professor, calling one of his books The Behavior of Organisms instead of The Behavior of Animals. The book said nothing about plants.) Many tables seem more meant to impress than communicate but it isn’t just tables. That section at the end where epidemiologists talk about the “limitations” of their study: The content is so predictable, so fact-free and unhelpful that I think they are just trying to impress readers with how careful they are. So I would add to Andrew’s advice: Don’t tell people what they already know.
I also like his list of content-less words, such as very and nice. Allen Neuringer told me you should never use very and I was impressed.
That section at the end where epidemiologists talk about the “limitations” of their study: The content is so predictable, so fact-free and unhelpful. . .
. . . that one might suspect many authors only include those points so that they don’t provide an open flank to reviewers?
yeah, I agree, some tables are there just because it’s conventional. Perhaps fear is the motivation: fear of not being conventional. Then the broad lesson is: Don’t do something just because other people do it that way. Jane Jacobs told a story about people who cut the end off of their meatloaf before putting it in the oven. Why? she asked. They didn’t know. It was just how it was done. Turned out it used to be done because the oven was too small for the whole thing. Now they had a bigger oven but kept on cutting off the end. Graphs were once hard to make, so people made tables instead. Now graphs are easy but not everyone has adjusted.
I can see one good reason to have tables in there – reproduceability. Give people your raw data so they can repeat the analysis themselves.
James, that’s very rare — anything approaching raw data. I agree with you but that stuff should be in an appendix.