Overview
The continuity correction improves a normal approximation to a binomial count by adjusting k by half a unit.
Enter the number of trials, the success probability, the target k and a mode: 1 for at most k, or 2 for at least k.
Worked examples
| Inputs | Result |
|---|---|
| Number of Trials (n)80Success Probability (%)35Target k30Probability ofAt most k | P(X <= k) normal approximation with continuity correction: 72.11%Mean: 28SD: 4.27 |
Show 2 more examples
| Number of Trials (n)100Success Probability (%)50Target k55Probability ofAt least k | P(X >= k) normal approximation with continuity correction: 18.41%Mean: 50SD: 5 |
| Number of Trials (n)40Success Probability (%)20Target k5Probability ofAt most k | P(X <= k) normal approximation with continuity correction: 16.15%Mean: 8SD: 2.53 |
P(X <= k) normal approximation with continuity correction by target k
With the calculator's starting values, the P(X <= k) normal approximation with continuity correction is 72.11%. The table shows how it changes with the target k.
| Target k | P(X <= k) normal approximation with continuity correction |
|---|---|
| 15 | 0.17% |
| 20 | 3.94% |
| 25 | 27.89% |
| 30starting value | 72.11% |
| 35 | 96.06% |
| 40 | 99.83% |
| 45 | 100% |
Other inputs: Number of Trials: 80 n · Success Probability: 35%. Change them in the calculator above for your own numbers.
How it works
A binomial count has mean n × p and standard deviation √(n × p × (1 − p)). The continuity correction moves k by 0.5 to bridge the gap between whole-number counts and the smooth normal curve: k + 0.5 for 'at most' and k − 0.5 for 'at least'. The approximation works well when n × p and n × (1 − p) are both at least about 10.
The formula
Practical tips
- 80 trials at 35% have a mean of 28 and an SD of 4.27, so P(X ≤ 30) ≈ 72.11% with the correction.
- For small n or p close to 0 or 1, use the exact binomial distribution instead.
- Without the correction, results for small n can be noticeably off.
Frequently asked questions
Why add 0.5?
A whole number k covers the interval from k − 0.5 to k + 0.5 on the continuous curve.
When is the approximation good?
When np and n(1 − p) are both at least about 10.
Can I get the exact answer?
Yes, our binomial distribution calculator gives exact values.