Figure 1

Examples of (A) confounding and (B) reverse causality in observational epidemiology. (A) The arrows denote the direction of proposed causality and the cross denotes that the postulated direct link between yellow teeth and lung cancer is false. (B) The arrows denote the direction of proposed causality and the cross denotes that the postulated direct link between C-reactive protein (CRP) and CHD is false and in fact the current evidence suggests that CHD raises levels of CRP (ie, the arrow goes in the opposite direction).