Lognormal Distribution Calculator
Find lognormal probabilities, quantiles, mean, median and mode — and see how the bell of ln X becomes the skewed curve of X.
Enter the mean m and SD s of your raw (unlogged) data, and this sets μ and σ for you — μ = ln(m²/√(s²+m²)), σ = √(ln(1+(s/m)²)).
With μ = 0 and σ = 0.5, the mean is 1.1331, the median is 1.0000 and the mode is 0.7788 — mean > median > mode, as always for a right-skewed lognormal.
Calculation
The normal–lognormal bridge
Left: ln X is normal with mean μ = 0 and SD σ = 0.5 — a bell. Right: X itself is lognormal — the same bell, stretched onto the positive axis, so it skews right.
A lognormal distribution calculator finds probabilities, quantiles, mean, median and mode for a lognormal variable — the distribution of a quantity whose logarithm is normal. Enter μ and σ, the mean and SD of ln(X), and it returns P(X < x), P(X > x), P(a < X < b) or any quantile, with the mean, median, mode, variance and standard deviation, and a normal–lognormal bridge chart.
How to use the lognormal distribution calculator
- Enter μ, the mean of ln(X) — not the mean of X. If you only know the mean and SD of your raw data, use the converter below the inputs.
- Enter σ, the SD of ln(X) — a positive number. A small σ (0.1) is nearly normal; a large σ (2) is heavily skewed.
- Choose what you want: P(X < x), P(X > x), P(a < X < b), or a quantile (the value at probability p).
- Enter x (or a and b, or p) and read the headline answer, then check the results grid for the mean, median, mode, variance and SD. The ledger shows every step, and the bridge chart draws the bell of ln X beside the skewed curve of X.
- Use the log-scale check: P(ln X < 0) should equal P(X < 1) — they are the same event.
Try it now: with μ = 0 and σ = 0.5, the calculator gives P(X < 1.5) = 0.7913, a mean of 1.1331, a median of 1.0000 and a mode of 0.7788. Click “μ = 1, σ = 0.25” and the median becomes 2.7183 with a mean of 2.8046.
When should you use a lognormal distribution?
Use a lognormal distribution when your quantity is positive, multiplicative and right-skewed — when it grows by percentages rather than by fixed amounts. Incomes and stock prices are the classic examples: both are positive, both are the product of many small multiplicative factors, and both have a long right tail of very large values. It is the right distribution when:
- Your data are all positive and right-skewed, with a long tail of large values.
- The log of your data looks approximately normal — take the log and the bell appears.
- The quantity is the product of many small effects (by the central limit theorem on the log scale).
- Do not use it for quantities that can be negative or zero, or for waiting times between independent events — those are exponential.
For the family of distributions this one belongs to, see types of probability distributions.
μ and σ live on the log scale
μ and σ are the mean and SD of ln(X), NOT of X. That single sentence is the whole trap: if you take the mean of your raw data — say, incomes averaging £40,000 — and type 40000 as μ, every answer will be wrong by orders of magnitude, because μ = 40000 would mean the median of X is e40000. The converter in the calculator does the translation for you: if your raw data has mean m and SD s, then μ = ln(m²/√(s²+m²)) and σ = √(ln(1+(s/m)²)). For m = 1.1331 and s = 0.6039, that gives μ = 0 and σ = 0.5 — the worked example below. The calculator refuses a σ of 0 or less, because a standard deviation must be positive.
Mean > median > mode, always
A lognormal variable is right-skewed by construction — its long right tail pulls the mean above the median, and the median above the mode. So mean > median > mode, always, for every μ and σ: with μ = 0 and σ = 0.5, the mean is 1.1331, the median is 1.0000 and the mode is 0.7788. The gap widens as σ grows: at σ = 0.1 the three nearly agree (mean ≈ median ≈ mode), and at σ = 2 the mean is far above the median. That is why the median — not the mean — is the honest “typical value” for incomes and stock prices: a few very large values drag the mean up but leave the median where most of the data sits.
The normal–lognormal bridge
Take the log of a lognormal variable and you get a normal one: if X is lognormal with parameters μ and σ, then ln X is normal with mean μ and SD σ. That is the bridge the chart draws — the bell of ln X beside the skewed curve of X. It is also why every lognormal probability is a normal probability in disguise: P(X < x) = Φ((ln x − μ)/σ), and the p-th percentile is exp(μ + σ·zp) where zp is the normal quantile. If you need the normal side of the bridge, use the normal distribution calculator.
Worked examples
Example 1 — μ = 0, σ = 0.5
With μ = 0 and σ = 0.5: the median is exp(0) = 1.0000, the mean is exp(0 + 0.25/2) = exp(0.125) = 1.1331, the mode is exp(0 − 0.25) = 0.7788, the variance is (exp(0.25) − 1)·exp(0.25) = 0.3647 and the SD is 0.6039. P(X < 1.5) = Φ((ln 1.5 − 0)/0.5) = Φ(0.8109) = 0.7913, P(0.5 < X < 2) = Φ(1.3863) − Φ(−1.3863) = 0.8343, and the 90th percentile is exp(0.5·1.2816) = 1.898. Log-scale check: P(ln X < 0) = Φ(0) = 0.5 — because ln X is centred on 0, half its mass is below 0, and X < 1 exactly when ln X < 0.
Example 2 — μ = 1, σ = 0.25
With μ = 1 and σ = 0.25: the median is exp(1) = 2.7183 and the mean is exp(1 + 0.03125) = 2.8046. The mode is exp(1 − 0.0625) = 2.5536, the variance is 0.5073 and the SD is 0.7122.
Example 3 — nearly normal and heavily skewed
At σ = 0.1 (with μ = 0), the distribution is nearly normal: the mean is 1.0050, the median is 1.0000 and the mode is 0.9900 — the three almost agree, and the curve is nearly a bell. At σ = 2, it is heavily skewed: the mean is 7.3891, the median is still 1.0000 and the mode is 0.0183 — the mean is more than seven times the median, and the mode has collapsed towards zero.
Frequently asked questions
Are μ and σ the mean and SD of X?
No. μ and σ are the mean and standard deviation of ln(X), the logarithm of X — not of X itself. The mean of X is exp(μ + σ²/2), the median is exp(μ) and the mode is exp(μ − σ²). Entering the raw mean of your data as μ is the #1 error with this distribution: if your raw data has mean m and SD s, use μ = ln(m²/√(s²+m²)) and σ = √(ln(1+(s/m)²)) instead.
Why is the mean bigger than the median?
Because a lognormal variable is right-skewed: its long right tail pulls the mean above the median, and the median above the mode, so mean > median > mode always. With μ = 0 and σ = 0.5, the mean is 1.1331, the median is 1.0000 and the mode is 0.7788. The gap widens as σ grows — at σ = 2 the mean is 7.3891 while the median is still 1.0000. That is why the median is the honest “typical value” for incomes and stock prices.
Source: the lognormal distribution follows the definition in the NIST/SEMATECH e-Handbook of Statistical Methods, Lognormal Distribution. All examples on this page were recomputed with Python (scipy.stats), last updated 19 November 2026.
See all 91 free statistics tools on Empirical Rule Calculators.
Next steps
🧮
● Live
Normal Distribution Calculator
Take the log of a lognormal variable and you get a normal one.
Calculator
Try Calculator →
📊
● Live
Exponential Distribution Calculator
Another right-skewed distribution — waiting times with a different tail.
Calculator
Try Calculator →
📘
● Live
Types of Probability Distributions
See where the lognormal sits in the family of distributions.
Article
Read Article →
More Tools to Explore
Keep going with the calculators that pair with the lognormal distribution — the normal bridge, other skewed distributions and z-scores.
Normal Distribution Calculator
Find areas, percentiles and z-scores under the normal curve.
Exponential Distribution Calculator
Work out exponential probabilities, quantiles and the mean from a rate or a mean waiting time.
Uniform Distribution Calculator
Work out uniform probabilities, quantiles, mean and variance for continuous and discrete uniform distributions.
Geometric Distribution Calculator
Find geometric probabilities for trials-until-success or failures-before-success.
Binomial Probability Calculator
Count successes with replacement, where the success probability stays fixed.
Z-Score Calculator
Turn any raw score into a z-score and read off its percentile in seconds.
Explore All Tools→
Browse all 91 free statistics tools — every calculator on this site in one place.