How this calculation works
A plain average would give the small group and large group equal weight. Combining the actual counts gives each observation equal weight.
100 × (s₁ + s₂)/(n₁ + n₂)A worked example
8 out of 10 and 30 out of 50 combine to 38 out of 60, about 63.33%.
Good to know
Use disjoint groups with comparable definitions of success.
Reference: OpenStax Prealgebra (opens in a new tab) ↗
Pool the counts before calculating the percentage
The first group has 8 successes out of 10 observations, or 80%. The second has 30 out of 50, or 60%. Together they have 38 successes in 60 observations. The combined percentage is therefore 38/60 × 100 ≈ 63.33%. Adding actual counts ensures each observation receives equal weight.
A plain average of 80% and 60% gives 70%. That calculation gives the two groups equal weight despite the second group being five times as large. You can reproduce the pooled answer by weighting the rates by group size: (80 × 10 + 60 × 50)/(10 + 50) ≈ 63.33%.
Make sure the groups can be combined
Group totals must be positive, and a success count cannot exceed its group total. The groups should be disjoint when the intended denominator is the number of unique observations. If the same person appears in both groups, adding the counts measures recorded observations and can count that person twice.
Both groups also need a compatible definition of success. Combining one group’s attendance rate with another group’s examination pass rate would create a percentage with no consistent interpretation. Use the Percentage point change calculator to compare two rates without pooling their observations. A combined rate is useful for an overall summary but can hide differences between the groups, so retain the separate rates when those differences matter.
A few questions, answered
When does the plain average work?
It agrees with pooling when the group totals are equal. It also agrees when the two group percentages happen to be identical.
Can a group have zero successes?
Yes. Enter zero successes with a positive total. Those observations increase the pooled denominator without increasing the numerator.