We probably all know the feeling: You read an abstract of a paper and it sounds all good, so you acquire the real paper but then find that the sample or effect size is uselessly small. So you get frustrated and close the tab with the paper.

It seems to me that this situation shouldn't happen at all in that such obvious information should be required of the abstract, at least for the simpler research designs. Those of us who are reviewers could make approval conditional on that such information is reported in the abstract.

Personally, I'd like to see sample size and effect size of main outcome, and either confidence intervals or a p value (strongly preferring the first).

Thoughts?