Relative Risk

Relative risk (RR) divides the cumulative incidence (risk) of an outcome in an exposed group by the incidence in an unexposed group. An RR of 1 means the exposure does not change risk; an RR of 2 means the exposed group has twice the risk; an RR of 0.5 means risk is halved. Because it is built directly from probabilities, RR is often more intuitive than the odds ratio.

RR is the standard effect measure for cohort studies and randomized trials, where a defined population is followed over time and incidence can be measured directly. It anchors much of how population scientists communicate exposure effects, because "twice the risk" maps cleanly onto how people reason about danger.

A key pitfall is confusing relative with absolute effects. A doubling of risk (RR = 2) is alarming if the baseline risk is 20% but trivial if it is 1 in a million. Relative risk alone hides the baseline, so it should be reported alongside the risk difference or absolute risks. RR also cannot be estimated directly from a standard case-control design.

For example, if 6% of a hypothetical screened cohort and 12% of an unscreened cohort develop late-stage disease, the RR is 6/12 = 0.5 — screening is associated with half the risk.