Relative Risk (2×2) — cohort / prospective risk ratio
VerifiedAdvanced & specialized
Run this test straight away on a free built-in teaching dataset — no data of your own needed — or bring your own. Either opens the guided workspace: variable setup, assumption diagnostics, results with effect sizes and confidence intervals, figures, and APA-ready reporting.
Loading teaching datasets…
Or use your own dataset
Loading your datasets…
Compares the probability of a binary event between two groups (exposed vs unexposed).
Valid in prospective cohort / RCT designs where exposure PRECEDES outcome — RR is biased on case-control data and the engine refuses to compute it on designs where the marginal disease prevalence is fixed. Reports RR, Katz log-CI, absolute risk difference (RD), and number-needed-to-treat (NNT). The engine surfaces a primary-voice qualifier when the unexposed baseline > 50% — RR is ceiling-bounded by 1/p_unexp in that regime.
Worked example
How much more likely is the outcome with exposure?
Relative risk from a 2×2 table compares the probability of the outcome between exposed and unexposed — best for cohort / prospective designs.
Exposure raised the risk by 70% (RR = 1.7, 95% CI [1.2, 2.4], p = .003).
The relative risk showed exposure increased outcome risk, RR = 1.7, 95% CI [1.2, 2.4], p = .003.
When to use it
- Randomised controlled trial (RCT)A vaccine RCT randomises 5,000 participants to vaccine or placebo and follows them for 6 months.
- Prospective cohort study (observational)An occupational cohort of 2,000 chemical-plant workers is followed for 10 years.
When NOT to — use instead
- Case-control study (retrospective)Marginal disease prevalence is fixed by design in case-control sampling — RR is mathematically not identifiable. → Odds Ratio (2\u00d72) \u2014 case-control / cross-sectional
- Cross-sectional with no temporal directionWithout a clear exposure-then-outcome temporal order, RR is at best a prevalence ratio — usually clearer to report OR or PR explicitly. → Odds Ratio (2\u00d72) \u2014 case-control / cross-sectional
- Common outcome with severe ceiling effectsWhen unexposed baseline is high (>50%), RR is structurally bounded; report OR or absolute risk difference instead. → Odds Ratio (2\u00d72) \u2014 case-control / cross-sectional
- Multivariable adjustment neededCrude 2×2 RR cannot adjust for confounders. → Binomial Logistic Regression \u2014 odds ratios + ROC AUC + classification
Hypotheses
Parameter tested: relative risk RR
Assumptions (and what to do if they fail)
Check: Min cell count.
If violated: Zero cells break the log-CI.
Check: c / n0.
If violated: Rare event: RR becomes unstable and similar to OR.
Ready to run a Relative Risk (2×2) — cohort / prospective risk ratio on your own data?
Guided setup, automatic assumption checks, effect sizes, figures and an APA write-up.
Run this test →