Relative Risk vs Odds Ratio: What’s the Difference?

Relative Risk vs Odds Ratio, When researchers want to understand whether an exposure, treatment, or other factor is associated with an outcome, two commonly used measures are Relative Risk (RR) and Odds Ratio (OR).

Although both describe the strength and direction of an association, they are not interchangeable. The difference becomes particularly important when the outcome is common.

This guide explains relative risk vs odds ratio, how to calculate each measure, how to interpret the results, and when to use RR or OR.

What Is Relative Risk?

Relative risk, also called the risk ratio, compares the probability of an outcome between two groups.

The basic formula is:

RR = Risk in exposed group / Risk in unexposed group

For example, suppose researchers compare two study methods and observe the following results:

• 60% of students using Method A pass
• 40% of students using Method B pass

The relative risk is:

RR = 0.60 / 0.40 = 1.50

How to Interpret Relative Risk-Relative Risk vs Odds Ratio

An RR of 1.5 means that the probability of passing is 1.5 times as high in Method A compared with Method B.

An RR can generally be interpreted as:

• RR = 1: No difference in risk
• RR > 1: Higher risk in the exposed/treatment group
• RR < 1: Lower risk in the exposed/treatment group

Relative risk is particularly intuitive because it directly compares probabilities.

It is commonly used in cohort studies and randomized controlled trials, where researchers can directly estimate the incidence or risk of an outcome.

What Is an Odds Ratio-Relative Risk vs Odds Ratio?

An odds ratio compares the odds of an outcome between two groups.

First, remember that risk and odds are different.

If 20 out of 100 people experience an event:

Risk = 20 / 100 = 0.20

But the odds are:

Odds = 20 / 80 = 0.25

In general:

Odds = Probability / (1 − Probability)

The odds ratio is:

OR = Odds in exposed group / Odds in unexposed group

For a standard 2 × 2 table:

The odds ratio is:

OR = (a × d) / (b × c)

Odds ratios are particularly important in case-control studies and are also commonly produced by logistic regression.

Relative Risk vs Odds Ratio: Example

Suppose we want to examine whether attending a tutorial is associated with passing an examination.

Step 1: Calculate the Risk

For students who attended:

Risk = 80 / 100 = 0.80

For students who did not attend:

Risk = 50 / 100 = 0.50

Therefore:

RR = 0.80 / 0.50 = 1.60

The probability of passing is 1.6 times as high among students who attended the tutorial.

Step 2: Calculate the Odds

For students who attended:

Odds = 80 / 20 = 4

For students who did not attend:

Odds = 50 / 50 = 1

Therefore:

OR = 4 / 1 = 4.00

The odds of passing are 4 times as high among students who attended the tutorial.

Notice something important:

RR = 1.60

while

OR = 4.00

Both measures indicate a positive association, but the magnitude is very different.

This happens because passing is a relatively common outcome in this example.

Why Are Relative Risk and Odds Ratio Different?

The difference comes from their denominators.

Relative risk compares:

Events / Total population

Odds compare:

Events / Non-events

When an outcome is uncommon, the number of non-events is close to the total population. Consequently, odds and probability become numerically similar, and OR and RR can also become similar.

As the outcome becomes more common, the difference between odds and probability becomes larger, and the OR can move substantially farther from 1 than the RR.

When Are OR and RR Similar?

The odds ratio can be a reasonable approximation to relative risk when the outcome is rare.

A commonly used rule of thumb is that the outcome should be relatively uncommon, although there is no universal cutoff that works for every study. Some sources use approximately 10%, while other epidemiological guidance uses broader thresholds depending on the situation.

For example, suppose:

• Risk in Group A = 2%
• Risk in Group B = 1%

Then:

RR = 0.02 / 0.01 = 2.00

Odds:

Odds A = 0.02 / 0.98 ≈ 0.0204

Odds B = 0.01 / 0.99 ≈ 0.0101

Therefore:

OR ≈ 2.02

Here, OR and RR are very close.

When Does the Odds Ratio Differ Substantially From Relative Risk?

The difference becomes more important when the outcome is common.

For example, suppose:

• Risk in exposed group = 30%
• Risk in unexposed group = 15%

Relative risk:

RR = 0.30 / 0.15 = 2.00

Now calculate the odds:

Odds exposed = 0.30 / 0.70 = 0.429

Odds unexposed = 0.15 / 0.85 = 0.176

Therefore:

OR ≈ 0.429 / 0.176 = 2.43

So the same data produce approximately:

RR = 2.00

OR = 2.43

An OR of 2.43 should not be interpreted as meaning that the risk is 2.43 times higher.

That would incorrectly treat an odds ratio as a relative risk.

How to Interpret an Odds Ratio Correctly

Suppose a logistic regression produces:

OR = 2.5

The correct interpretation is:

The odds of the outcome are 2.5 times as high in the exposed group compared with the reference group, assuming the model and other variables are appropriately specified.

It is generally incorrect to write:

“The exposed group has 2.5 times the risk.”

That statement would interpret OR as RR.

The two measures answer related but different questions.

How to Interpret Relative Risk

Suppose a cohort study produces:

RR = 2.5

A suitable interpretation is:

The probability or risk of the outcome is 2.5 times as high in the exposed group compared with the unexposed group.

This is one reason relative risk is often easier to communicate to a general audience.

RR vs OR: Quick Comparison

When Should You Use Relative Risk?

Relative risk is particularly useful when you can directly calculate the incidence or probability of an outcome in each group.

Typical examples include:

• Cohort studies
• Prospective studies
• Randomized controlled trials
• Clinical trials
• Epidemiological studies where risks can be estimated

For example, if 10 out of 100 people in the treatment group develop an outcome and 20 out of 100 in the control group develop it:

RR = 0.10 / 0.20 = 0.50

The treatment group has half the risk of the control group.

When Should You Use Odds Ratio?

Odds ratios are especially useful when the study design does not allow direct estimation of risk.

The classic example is a case-control study.

In a case-control design, researchers start by selecting people with the outcome and people without the outcome. Because the number of cases and controls is determined by the study design, the original population risk cannot generally be calculated directly.

The odds ratio is therefore the natural measure of association for this design.

OR is also the standard association measure produced by conventional logistic regression.

Why Logistic Regression Uses Odds Ratios

Logistic regression models a binary outcome, such as:

• Disease / No disease
• Purchased / Did not purchase
• Passed / Failed
• Responded / Did not respond

The coefficients from a logistic regression model are naturally interpreted on the odds scale.

After exponentiating a coefficient:

OR = eᵝ

the result represents an odds ratio associated with a one-unit change in the predictor, holding other model variables constant.

For example, if:

β = 0.693

then:

OR = e⁰·⁶⁹³ ≈ 2.00

This means the odds are approximately doubled for the specified change in the predictor, conditional on the other variables in the model.

Relative Risk and Odds Ratio for Protective Effects

The same concepts apply when the exposure is associated with a lower risk.

Suppose:

RR = 0.60

This means the risk in the exposed group is 60% of the risk in the comparison group.

The relative reduction in risk is:

1 − RR = 1 − 0.60 = 0.40

or 40%.

For odds ratios, interpretation remains on the odds scale.

For example:

OR = 0.60

means the odds of the outcome are 40% lower, not necessarily that the risk is 40% lower.

This distinction becomes especially important when the outcome is common.

Confidence Intervals Should Be Reported

RR and OR should generally be reported with their confidence intervals.

For example:

RR = 1.50 (95% CI: 1.10–2.05)

or:

OR = 2.20 (95% CI: 1.30–3.70)

The confidence interval provides information about the precision of the estimated association.

For both RR and OR:

• A value of 1 indicates no association.
• A confidence interval that excludes 1 is commonly interpreted as statistically significant at the corresponding two-sided level.

However, statistical significance should not be considered separately from effect size, study design, uncertainty, and practical importance.

Don’t Forget Absolute Risk

Relative measures can sometimes make an effect appear more dramatic than it is.

Suppose a treatment reduces risk from:

2% → 1%

The relative risk is:

RR = 1% / 2% = 0.50

That represents a 50% relative reduction.

But the absolute risk reduction is only:

2% − 1% = 1 percentage point

This is why reporting the underlying risks alongside RR or OR can make results much easier to understand.

For clinical or practical decision-making, absolute risk measures can provide important context alongside relative measures.

Common Mistakes When Comparing RR and OR

Mistake 1: Treating OR as RR

An OR of 3 does not necessarily mean the risk is three times higher.

Mistake 2: Assuming OR and RR Are Always Similar

They can be close for rare outcomes but can diverge substantially when outcomes are common.

Mistake 3: Ignoring Study Design

Case-control studies generally require odds ratios because risk cannot be directly calculated from the sampled cases and controls.

Mistake 4: Reporting Only the Ratio

RR or OR should ideally be accompanied by confidence intervals and enough information about baseline risks to understand the practical magnitude.

Mistake 5: Ignoring Absolute Risk

A large relative effect can correspond to a small absolute change when the baseline risk is very low.

Relative Risk vs Odds Ratio: Which Is Better?

There is no single measure that is always better.

The appropriate measure depends on the research question, study design, outcome frequency, and statistical model.

A useful rule is:

Use RR when you can directly estimate risk and want an intuitive comparison of probabilities.

Use OR when the study design requires it, particularly in case-control studies, or when using conventional logistic regression.

For common outcomes, be particularly careful about interpreting an OR as though it were an RR.

Frequently Asked Questions

Is relative risk the same as odds ratio?

No. Relative risk compares probabilities, while odds ratio compares odds. They can be numerically similar for rare outcomes but may differ substantially when the outcome is common.

Is an odds ratio higher than relative risk?

For positive associations, the OR is often farther from 1 than the RR when the outcome is common. However, the exact relationship depends on the underlying risks and direction of association.

When is odds ratio approximately equal to relative risk?

OR and RR are often close when the outcome is rare or when the association is modest. The approximation becomes less reliable as the outcome becomes more common and the effect becomes larger.

Can relative risk be used in a case-control study?

Not directly from the usual case-control sampling design because the numbers of cases and controls are selected by the researcher rather than reflecting the underlying population risks. The odds ratio is therefore the conventional measure for case-control studies.

Does an OR of 2 mean twice the risk?

No. An OR of 2 means the odds are twice as high. It does not necessarily mean the probability or risk is twice as high.

Should I report RR or OR?

Choose the measure that matches the study design and research question. When risks can be directly estimated, RR is often easier to interpret. OR is appropriate for case-control studies and is the standard output of logistic regression.

Final Takeaway

Relative Risk and Odds Ratio are both important measures of association, but they measure different quantities.

Relative risk compares probabilities, while odds ratio compares odds.

The distinction may be small when the outcome is rare, but it can become substantial when the outcome is common.

The most important points to remember are:

• RR = risk in one group / risk in another group
• OR = odds in one group / odds in another group
• RR is often easier to interpret
• OR is commonly used in case-control studies
• Logistic regression typically reports ORs
• OR and RR are not interchangeable
• The difference becomes more important as the outcome becomes common
• Confidence intervals should be reported
• Absolute risks should be considered alongside relative measures

Ultimately, the best measure is not simply the one that produces the largest or most impressive number. It is the measure that correctly represents the study design, research question, and underlying data.