Say you’re a researcher. You’ve spent months collecting and analyzing data, crunching numbers and composing tables, but it’s all been for naught. That hypothesis you set out to test (say, that a link exists between a specific food and a known indicator of cancer risk) didn’t pan out. In your investigation at least, you found no such link.

You have achieved what in scientific circles is called a “null finding.” And, in a very real sense, that’s not nothing.

Null findings don’t make headlines, and often don’t even get published. (The tendency of journal editors to publish results that seem “new” over those that find no association — or that simply accord with previously published results — is a source of publication bias, which over time can distort the general scientific opinion on a given subject.)

But there is an important difference between a null finding capable of closing the book on a given question — that says, essentially “There’s no there there, move along.” — and a null finding that says simply “We need more and better data before we can make a judgment.”

On those rare occasions when the media do pick up on a null finding, there is a tendency to mistake one kind for another. Let’s take a look at a recent null finding and see what it really has to say. Continue reading →