I occasionally see studies that (at least as presented in news articles and blog entries) group together both an obvious/easy/common category and an unusual/obscure/rare category, and then present the results in terms of the unusual category.

Like: This survey shows that an astonishing 90% of dining-room-table-related injuries are caused by tables falling from airplanes or people bumping their elbows on the table at dinner. Conclusion: We must ban tables from airplanes!

Or: This survey shows that an astonishing 99% of table injuries are either fatal or very minor. Conclusion: Tables should be illegal!

(I made up these table-accident examples entirely at random; I don't intend them as a metaphor for any specific real-life thing.)

I think most of the time when I see this, the conclusion is being drawn by an overzealous blogger rather than by the original study. But the study seems to be setting up for this kind of conclusion by grouping the two things together in the first place. I'm sure in most cases the people doing the study had a good reason for grouping things the way they did. But in cases where the implications of the results for the two parts are very different, it might be wise for the people writing up the study to keep the two parts separate.

Join the Conversation