Trial Input Panel
Example Data Table
| Scenario | Mean Difference | Std. Dev. | Alpha | Target Power | Required N1 | Required N2 |
|---|---|---|---|---|---|---|
| Baseline Example | 5.0 | 12.0 | 0.05 | 0.80 | 91 | 91 |
| Higher Variability | 5.0 | 15.0 | 0.05 | 0.80 | 143 | 143 |
| Higher Power Goal | 5.0 | 12.0 | 0.05 | 0.90 | 122 | 122 |
Formula Used
This calculator uses a normal approximation for a two-arm study with a continuous endpoint and common standard deviation. First, standardized effect size is computed as:
Effect Size = |Mean Difference − Margin| / Standard Deviation
For equal or unequal allocation, required Group 1 size is estimated by:
n₁ = ((r + 1) / r) × ((Z1−α/tails + Zpower) / Effect Size)²
Then Group 2 size is:
n₂ = r × n₁
Design effect inflates both groups, while dropout adjustment divides by the retention rate. Achieved power is then estimated from the entered sample sizes:
Power ≈ Φ(δ / √(1/n₁ + 1/n₂) − Z1−α/tails)
For two-sided settings, the mirrored tail is also included. This gives a practical planning estimate for superiority-style studies.
How to Use This Calculator
- Enter the expected mean difference between intervention and control.
- Provide the shared standard deviation from pilot data or literature.
- Choose alpha and whether the test is one-sided or two-sided.
- Set the desired target power, usually 0.80 or 0.90.
- Enter planned sample sizes to estimate achieved power.
- Adjust allocation ratio if groups will be unbalanced.
- Add dropout rate and design effect for realistic enrollment targets.
- Review the summary, compare enrollment counts, and inspect the power curve.
- Download CSV for tabular records or PDF for reporting.
FAQs
1. What does clinical trial power mean?
Power is the probability of detecting a true treatment effect when it really exists. Higher power reduces the chance of missing a meaningful difference.
2. Why does standard deviation matter so much?
Larger variability makes it harder to distinguish treatment effects from noise. As standard deviation rises, required sample size usually increases quickly.
3. Why include dropout in planning?
Dropout reduces the number of analyzable participants. Inflating enrollment helps preserve target power after losses during follow-up or protocol deviations.
4. What is the allocation ratio?
It is the size of Group 2 relative to Group 1. A ratio of 1 means equal groups, while 2 means Group 2 is twice as large.
5. When should I use one-sided testing?
One-sided testing is appropriate only when effects in the opposite direction are irrelevant or scientifically implausible. Many confirmatory trials still prefer two-sided testing.
6. What does the design effect represent?
Design effect adjusts for correlation or other design features that reduce effective information. Clustered or complex studies often need this inflation factor.
7. Is this suitable for binary or survival endpoints?
No. This version targets two-arm continuous outcomes with a common standard deviation. Binary, time-to-event, and repeated-measure studies need different formulas.
8. Why show both achieved power and required size?
Achieved power evaluates your current plan, while required size shows what is needed to meet a chosen target. Together they support iteration.