Sample Size Calculator
A tool for drafting a study plan. Its output is meant to be reviewed by the person responsible, who remains accountable for the sample size.
Driven by power. Secures the chance of clearing a target lower bound.
Result
Basis · MFDS Korea (power-based)
Positive specimens · sensitivity
minimum 149 · 10% dropout
Coefficient (1.96 + 0.84)² = 7.8400
Tolerated errors, by the exact (Clopper-Pearson) method
False negatives up to 8
At a collected size of 166, against a 90% target, by the exact method at two-sided 95%
How it was calculated
- Formula
- n= (Z₁₋α/₂ + Z₁₋β)² × P̂(1−P̂) ÷ (P̂ − P_L)²
- Values
- = (1.96 + 0.84)² × 0.95 × 0.05 ÷ (0.95 − 0.90)²
- Working
- = 7.8400 × 0.0475 ÷ 0.0025
- Rounded up
- = 148.96 → round up → 149
- Dropout correction
- n′= 149 ÷ (1 − 0.10) = 165.56 → round up → 166
Negative specimens · specificity
minimum 171 · 10% dropout
Coefficient (1.96 + 0.84)² = 7.8400
Tolerated errors, by the exact (Clopper-Pearson) method
False positives up to 3
At a collected size of 190, against a 95% target, by the exact method at two-sided 95%
How it was calculated
- Formula
- n= (Z₁₋α/₂ + Z₁₋β)² × P̂(1−P̂) ÷ (P̂ − P_L)²
- Values
- = (1.96 + 0.84)² × 0.98 × 0.02 ÷ (0.98 − 0.95)²
- Working
- = 7.8400 × 0.0196 ÷ 0.0009
- Rounded up
- = 170.74 → round up → 171
- Dropout correction
- n′= 171 ÷ (1 − 0.10) = 190.00 → round up → 190
166 positive + 190 negative
Z values are taken to two decimal places, per calculation rule 1.
Conditions
The share of collected specimens expected to drop out or be unusable. The minimum is divided by (1 − the rate).
Sample size against the target lower bound
Raising the target lower bound a little raises the sample size a lot. Use this to find the trade-off.
estimate held at 95%, coefficient 7.8400
| Target lower bound | Minimum | Collected (10% dropout) |
|---|---|---|
| 94% | 3,724 | 4,138 |
| 93% | 931 | 1,035 |
| 92% | 414 | 460 |
| 91% | 233 | 259 |
| 90%current setting | 149 | 166 |
| 89% | 104 | 116 |
| 88% | 76 | 85 |
| 87% | 59 | 66 |
| 86% | 46 | 52 |
| 85% | 38 | 43 |
Wording for the study plan
A paragraph to paste straight into a plan. It carries the track, the formula, the inputs and the result.
Basis: MFDS Korea (power-based) Formula: n = (Z₁₋α/₂ + Z₁₋β)² × P̂(1−P̂) ÷ (P̂ − P_L)² Sensitivity — estimate 95%, target lower bound 90%, confidence 95% Two-sided, power 80%, coefficient 7.8400 → minimum 149 With a 10% dropout rate → 166 collected - Sensitivity: at a collected sample size of 166, up to 8 error(s) can be tolerated while still holding the 90% target lower bound (exact method, two-sided 95%). Specificity — estimate 98%, target lower bound 95%, confidence 95% Two-sided, power 80%, coefficient 7.8400 → minimum 171 With a 10% dropout rate → 190 collected - Specificity: at a collected sample size of 190, up to 3 error(s) can be tolerated while still holding the 95% target lower bound (exact method, two-sided 95%). Total specimens: 356 (166 positive + 190 negative) Source: Ministry of Food and Drug Safety, Guideline on Clinical Performance Studies of In Vitro Diagnostic Medical Devices (안내서-0640-03 (2020-11-26)), checked 2026-07-29 This figure supports drafting the plan. The sample size is settled by the responsible reviewer.
Conditions shared by every row
7.84
Sensitivity
Press a row to add that condition to the total below. The ↗ button does not select — it opens the single calculation.
| Estimate % | Target lower bound % | Minimum n | Collected | Errors tolerated | Row actions |
|---|---|---|---|---|---|
Copy two columns out of a spreadsheet and paste into a cell to fill several rows at once.
Specificity
Press a row to add that condition to the total below. The ↗ button does not select — it opens the single calculation.
| Estimate % | Target lower bound % | Minimum n | Collected | Errors tolerated | Row actions |
|---|---|---|---|---|---|
Copy two columns out of a spreadsheet and paste into a cell to fill several rows at once.
Total for the conditions chosen
💡 Choose one row in each table and the two are added together.
Source documents
Ministry of Food and Drug Safety, Guideline on Clinical Performance Studies of In Vitro Diagnostic Medical Devices
안내서-0640-03 (2020-11-26) · Checked 2026-07-29 · Program v1.0
The final judgement rests with the responsible reviewer.
Often used together
2×2 Test Accuracy
Turns a 2×2 table into sensitivity, specificity, predictive values and likelihood ratios, each with a 95% confidence interval.
Power Analysis & Sample Size
Works out how many subjects an experimental group needs for a given power.
Bland-Altman plotter
Bland-Altman limits of agreement and Passing-Bablok regression from two methods measured on the same items.