Calculator
Normal Curve Preview
Example Data Table
| # | x | μ | σ | z | Percentile (≈) |
|---|---|---|---|---|---|
| 1 | 78 | 70 | 8 | 1.0000 | 84.13% |
| 2 | 55 | 60 | 10 | -0.5000 | 30.85% |
| 3 | 92 | 80 | 6 | 2.0000 | 97.72% |
| 4 | 68 | 68 | 12 | 0.0000 | 50.00% |
| 5 | 40 | 50 | 5 | -2.0000 | 2.28% |
Formula Used
The z-score standardizes a value by expressing it in standard deviations:
z = (x − μ) / σ
The percentile is the left-tail probability under the standard normal curve:
Φ(z) = P(Z ≤ z)
This calculator uses a stable approximation for Φ(z) and an inverse-CDF approximation for percentile-to-z conversion.
How to Use This Calculator
- Select a Calculation Mode (raw score, z, percentile, between, or critical).
- Fill in the required inputs shown for that mode.
- Choose a Shaded Region Preview to match your question.
- Click Submit to view results above the form.
- Use Download CSV or Download PDF to export results.
Article
1) What the curve shows
The standard normal curve centers at 0 and uses a spread of 1. Every z-score marks a point on that curve, so areas become probabilities. In this calculator, the left area equals Φ(z) and the right area equals 1 − Φ(z). The total area is always 1.0000, which is why left and right add to 1.
2) Converting a raw score
Use z = (x − μ) / σ. If x=78, μ=70, and σ=8, then z=1.0000. That matches the example table and places the score one standard deviation above average. If x equals μ, then z=0 and you sit at the center peak. Negative z values mean the score falls below μ.
3) Percentiles with real numbers
A percentile is simply 100×Φ(z). For z=0, Φ(z)=0.5000 (50th percentile). For z=1, Φ(z)≈0.8413 (84.13rd percentile). For z=2, Φ(z)≈0.9772 (97.72nd percentile). For z=−2, Φ(z)≈0.0228, showing how fast tails shrink. These anchors help sanity‑check outputs quickly.
4) Tail areas and p-values
Right-tail probability is 1−Φ(z). For z=1.645, Φ(z)≈0.9500, so the upper-tail area is about 0.0500. The two-tailed p-value is 2×min(left,right). For z=1.96, Φ(z)≈0.9750, giving a two-tailed p-value near 0.0500. For z=2.576, Φ(z)≈0.9950, which corresponds to α≈0.01 two-tailed.
5) Area between two cutoffs
When you enter z1 and z2, the calculator orders them and computes Φ(z2)−Φ(z1). For example, between −1 and 1 the area is about 0.6827, meaning roughly 68.27% of values lie within one standard deviation of the mean. Between −2 and 2 the area is about 0.9545, a common “95% rule” reference.
6) Critical z from alpha
Critical values define rejection regions. For α=0.05 two-tailed, each tail holds 0.025, so the cutoffs are approximately ±1.96. For α=0.05 one-tailed (upper), the cutoff is about 1.645. For α=0.10 one-tailed, the cutoff is about 1.2816. The tool reports these and previews the shaded tails for quick interpretation.
7) Using the shaded preview
Pick “Left of z” for percentiles, “Right of z” for upper-tail risk, “Between 0 and z” to read symmetry, and “Two tails” for two-sided tests. The preview uses the standard curve, but you can still interpret it for any normal model after standardizing. If the curve shading looks wrong, double-check the selected mode and sign of z carefully.
FAQs
How is z different from a raw score?
A raw score is in original units. A z-score is unitless and shows how far x is from the mean in standard deviations, using z=(x−μ)/σ.
What does “Left-tail area” represent?
It is Φ(z), the probability that a standard normal value is less than or equal to z. As a percent, it is the percentile.
When should I use the two-tailed p-value?
Use it for two-sided hypotheses where extreme values on either side matter. The calculator returns 2×min(left-tail,right-tail).
Why does the percentile-to-z mode reject 0 or 100?
Inverse CDF requires a probability strictly between 0 and 1. Exact 0 or 1 would map to infinite z, which is not usable in practice.
What is the meaning of critical z for alpha?
Critical z is the cutoff where the tail area equals α (one-tailed) or α/2 (two-tailed). Values beyond that cutoff fall in the rejection region.
Can I trust results for very large |z| values?
For typical statistics tasks the approximation is reliable. Extremely large |z| makes tail areas tiny, so rounding and display precision become more noticeable.