Correlation

Researchers should take care that they know what questions they're really asking of their data, say two Johns Hopkins scientists.

We default to cause-and-effect thinking because we want to maintain control over our lives, but some things just don't have clear answers.