Table 2 Hypothetical results from an observational study

Risk factor value, x	Population A, without intervention			Population B, with intervention			Odds ratio
Risk factor value, x	r	n	Risk estimate*	r	n	Risk estimate*	Odds ratio
1	10	100	0.1	10	200	0.05	0.47
2	60	300	0.2	45	300	0.15	0.71
3	180	600	0.3	125	500	0.25	0.77
All levels	250	1000	0.25	180	1000	0.18	0.66

*Note that the risks in population A if they had had the intervention would be half the observed estimates, and in population B if they had not had the intervention would be twice the observed estimate.