Confounding in epidemiological studies: why "independent" effects may not be all they seem.

Confounding in epidemiological studies: why "independent" effects may not be all they seem.