Normal Approximation Calculator

Estimate P(a ≤ X ≤ b) for a Binomial(n, p) distribution using the normal approximation. Supports continuity correction and one- or two-sided bounds.

Leave blank for one-sided upper probabilities.

Leave blank for one-sided lower probabilities.

Mean (μ = np): 20.0000

Standard deviation (σ = √np(1 − p)): 3.4641

Adjusted bounds: [14.5, 25.5]

Probability estimate: 88.7649%

z-score (lower): -1.5877

z-score (upper): 1.5877

How to Use This Calculator

  1. Specify the number of Bernoulli trials and the success probability.
  2. Choose lower and upper bounds to define the event of interest.
  3. Decide whether to apply continuity correction to improve accuracy.
  4. Interpret the approximate probability and z-scores, keeping in mind the normal approximation assumptions.

Formula

Binomial mean: μ = np

Binomial standard deviation: σ = √[np(1 − p)]

Normal approximation: X ≈ N(μ, σ²)

Continuity correction: adjust lower bound by −0.5, upper bound by +0.5

P(a ≤ X ≤ b) ≈ Φ((b + 0.5 − μ) / σ) − Φ((a − 0.5 − μ) / σ)

Full Description

For large n, the binomial distribution is well-approximated by a normal distribution with matching mean and variance. Continuity correction improves accuracy by aligning discrete probabilities with continuous areas under the curve.

The approximation performs best when np and n(1 − p) are both at least 5–10. For extreme probabilities or small n, consider exact binomial calculations.

Frequently Asked Questions

When is the normal approximation acceptable?

Rules of thumb require np ≥ 5 and n(1 − p) ≥ 5. Larger values yield better accuracy.

Why use continuity correction?

Because the binomial distribution is discrete, adjusting boundaries by 0.5 aligns the discrete probability mass with the continuous normal curve.

Can I compute one-tailed probabilities?

Yes. Leave one bound blank to obtain P(X ≤ b) or P(X ≥ a).

How accurate is the approximation?

Accuracy depends on n and p. For critical decisions, compare with the exact binomial CDF.