Skip to content
Generated on: July 20, 2026 at 09:33 PM

Odds Ratio Calculator

Inputs

Exposed, outcome30
Exposed, no outcome70
Unexposed, outcome10
Unexposed, no outcome90

Generated with OneCalc · https://onecalc.app/en/math/odds-ratio

Mathematics

Odds Ratio Calculator

Calculate the odds ratio from a 2×2 contingency table with a 95% confidence interval using Woolf’s log method. Enter the four cell counts for exposure and outcome.

Inputs

Contingency table (2×2)

Results

Enter a value to see results.

Details

\begin{aligned} OR &= \dfrac{a \cdot d}{b \cdot c} \\ &= \dfrac{(30) \cdot (90)}{(70) \cdot (10)} \\ &= ? \end{aligned}
\begin{aligned} \text{95\% CI} &= \exp\!\big(\ln OR \pm 1.96 \cdot SE\big) \\ SE &= \sqrt{\tfrac{1}{a} + \tfrac{1}{b} + \tfrac{1}{c} + \tfrac{1}{d}} = ? \\ &= [\,?,\ ?\,] \end{aligned}

Embed this calculator

Preview

Paste this code into your page to show the calculator.

Share this calculation

Anyone who opens this link sees your values already filled in.

What the odds ratio measures

The odds of an event are the number of times it occurs divided by the number of times it does not. If 30 of 100 exposed subjects develop an outcome, the odds are 30 to 70, or 0.43. The odds ratio (OR) compares the odds in two groups — typically an exposed group and an unexposed group — by dividing one set of odds by the other. It is one of the most widely used effect measures in epidemiology, especially for case-control studies, and it is the quantity estimated by logistic regression.

The 2×2 table

Most odds-ratio calculations begin from a two-by-two contingency table that cross-classifies each subject by exposure and outcome:

OutcomeNo outcome
Exposedab
Unexposedcd

Here a counts exposed subjects with the outcome (cases), b counts exposed subjects without it (controls), c counts unexposed subjects with the outcome, and d counts unexposed subjects without it.

The formula

The odds of the outcome among the exposed are a/b; among the unexposed they are c/d. Their ratio simplifies to the cross-product of the table:

OR=a/bc/d=a⋅db⋅cOR = \frac{a/b}{c/d} = \frac{a \cdot d}{b \cdot c}

Interpretation

  • OR > 1 — the outcome is more likely in the exposed group; exposure is associated with increased odds.
  • OR = 1 — the odds are identical in both groups; there is no association.
  • OR < 1 — the outcome is less likely in the exposed group; exposure is associated with decreased odds, suggesting a protective effect.

The odds ratio is symmetric: the OR for outcome given exposure equals the OR for exposure given outcome, which is one reason it remains valid in case-control designs where the marginal totals are fixed by sampling.

The 95% confidence interval

A point estimate alone says nothing about precision. Because the sampling distribution of the odds ratio is skewed, the confidence interval is constructed on the natural-log scale, where the distribution of ln(OR) is approximately normal. This is Woolf's method. The standard error of ln(OR) is

SEln⁡OR=1a+1b+1c+1dSE_{\ln OR} = \sqrt{\frac{1}{a} + \frac{1}{b} + \frac{1}{c} + \frac{1}{d}}

The interval is formed by stepping 1.96 standard errors either side of ln(OR) and then exponentiating back:

95% CI=exp⁡ ⁣(ln⁡OR±1.96⋅SEln⁡OR)\text{95\% CI} = \exp\!\big(\ln OR \pm 1.96 \cdot SE_{\ln OR}\big)

Every cell must be nonzero for this formula to be defined, since each contributes a 1/cell term. When a cell is empty, analysts often add 0.5 to each count (the Haldane–Anscombe correction) before computing the interval.

Worked example

Using the default table — a = 30, b = 70, c = 10, d = 90:

OR=30×9070×10=2700700=3.8571SEln⁡OR=130+170+110+190=0.398495% CI=exp⁡ ⁣(ln⁡3.8571±1.96×0.3984)=[ 1.767, 8.422 ]\begin{aligned} OR &= \frac{30 \times 90}{70 \times 10} = \frac{2700}{700} = 3.8571 \\ SE_{\ln OR} &= \sqrt{\tfrac{1}{30} + \tfrac{1}{70} + \tfrac{1}{10} + \tfrac{1}{90}} = 0.3984 \\ \text{95\% CI} &= \exp\!\big(\ln 3.8571 \pm 1.96 \times 0.3984\big) \\ &= [\,1.767,\ 8.422\,] \end{aligned}

The exposed group has nearly four times the odds of the outcome, and because the interval lies entirely above 1, the association is statistically significant at the 5% level.

Odds ratio versus relative risk

The odds ratio is often confused with the relative risk (RR), which compares probabilities rather than odds. The two coincide only when the outcome is rare. As the outcome becomes more common, the odds ratio moves farther from 1 than the relative risk, so an OR reported as if it were an RR can overstate the effect.

The choice of measure follows the study design. Cohort and cross-sectional studies sample by exposure status and can estimate risks directly, so the relative risk — see the Relative Risk Calculator — is usually preferred, and a significant risk difference can be translated into the Number Needed to Treat (NNT) Calculator. Case-control studies sample by outcome status, fixing the number of cases and controls, which makes risks inestimable; the odds ratio is then the natural and valid measure. The rare-disease assumption — that the outcome occurs in well under roughly 10% of each group — is what lets the odds ratio approximate the relative risk in case-control work.

Frequently Asked Questions (FAQ)

What is an odds ratio?

An odds ratio (OR) compares the odds of an outcome between two groups. Odds are the number of events divided by the number of non-events. From a 2×2 table with cells a (exposed cases), b (exposed controls), c (unexposed cases), and d (unexposed controls), the odds ratio is OR = (a·d) / (b·c). OR = 1 means the odds are the same in both groups; OR > 1 means the outcome is more likely with exposure; OR < 1 means it is less likely.

How is the odds ratio different from relative risk?

Relative risk (RR) compares probabilities — the chance of the outcome in each group — while the odds ratio compares odds. RR can be computed directly only when the sample reflects the true population at risk, as in a cohort study.

In case-control studies the number of cases and controls is fixed by design, so risks cannot be estimated and the odds ratio is used instead. When the outcome is rare, the OR closely approximates the RR; when the outcome is common, the OR is farther from 1 than the RR.

How do I read the 95% confidence interval?

The confidence interval gives the range of odds-ratio values compatible with the data. The width reflects precision: small samples and rare events give wide intervals. The key reference value is 1 (no association). If the interval excludes 1 — for example [1.77, 8.42] — the association is statistically significant at the 5% level. If the interval includes 1 — for example [0.6, 1.9] — the data are consistent with no effect.

When does the odds ratio overstate the relative risk?

The odds ratio always lies farther from 1 than the corresponding relative risk, and the gap grows as the outcome becomes more common. The rare-disease assumption — that the outcome occurs in well under about 10% of each group — keeps the OR close to the RR. When the outcome is frequent, reporting an OR as if it were an RR can substantially exaggerate the apparent effect, so the two measures should be distinguished carefully.

Recommended Next

Relative Risk Calculator

Calculate relative risk (risk ratio) from a 2×2 table with a 95% confidence interval. Enter the exposed and unexposed counts with and without the outcome.

 Learn more
Number Needed to Treat (NNT) Calculator

Calculate the number needed to treat (NNT) from control and treatment event rates, along with absolute risk reduction (ARR) and relative risk reduction (RRR).

 Learn more
Bayes’ Theorem Calculator

Compute the posterior probability P(A|B) from a prior, a true-positive rate (sensitivity), and a false-positive rate. Useful for medical-test and diagnostic reasoning.

 Learn more
Chi-Square Goodness-of-Fit Test

Run a Pearson chi-square goodness-of-fit test. Enter observed category counts and the counts expected under your null hypothesis to get the chi-square statistic, degrees of freedom, and p-value.

 Learn more

200+ calculators · 10 languages · 100% free

Was this calculator helpful?

No

How can we improve this calculator?
Powered by OneCalc ↗