Geometric Standard Deviation Calculator

Enter positive numbers to compute the geometric standard deviation.

How it is calculated

Enter at least two positive numbers separated by commas or spaces; zero, a negative number or text gives no result, because logarithms need positive values. The engine works on natural logarithms: first the mean of the logs, μ = Σ ln(xᵢ) ÷ n, then the variance of the logs, σ² = Σ (ln xᵢ − μ)² ÷ n. The geometric mean is e^μ and the geometric standard deviation is GSD = e^σ. The variance divides by n, not n − 1, so this is the population version; a sample-based GSD of the same data is somewhat larger. A GSD of 1 means all values are equal, and it is always at least 1 because it is a multiplicative factor, not a distance. Results are rounded to two decimals.

Worked example

Take the values 2, 4, 8, 16. Their natural logs are 0.693, 1.386, 2.079 and 2.773, with a mean of 1.733. The deviations from that mean are −1.040, −0.347, 0.347 and 1.040; squaring and averaging them gives 0.6006, whose square root is 0.775. So GSD = e^0.775 = 2.17, and the geometric mean is e^1.733 = 5.66. Read together: typical values lie roughly between 5.66 ÷ 2.17 ≈ 2.6 and 5.66 × 2.17 ≈ 12.3.

Try this example →

Common mistakes

  • Including a zero or a negative number. Logarithms are not defined for them, so the tool returns nothing; remove them or check whether the data is really log-normal.
  • Typing a decimal comma such as 2,5. Commas separate values here, so it is read as the two numbers 2 and 5 and the geometric mean changes; write 2.5.
  • Reading the GSD as an amount to add and subtract. It is a factor: the range is the geometric mean divided and multiplied by it, not shifted by it.
  • Expecting a GSD below 1. The value is always at least 1, and exactly 1 only when every value is identical.
  • Comparing the result with a sample-based figure from a statistics package. This tool divides by n, so a package that divides by n − 1 will report a slightly larger GSD.

About this calculator

The geometric standard deviation (GSD) is the multiplicative counterpart to the ordinary (additive) standard deviation, used specifically for data that follows a log-normal distribution — data that is skewed and only takes positive values, where ratios between values matter more than raw differences. Instead of adding and subtracting a spread around a mean, GSD multiplies and divides: roughly two-thirds of the data falls between the geometric mean divided by the GSD and the geometric mean multiplied by the GSD.

This measure is standard in fields where quantities are naturally log-normal, such as environmental and industrial hygiene, where airborne particle sizes and chemical exposure levels are almost always reported with a geometric mean and geometric standard deviation rather than an arithmetic one, and in biology and pharmacology when analyzing concentrations or dose-response data that span orders of magnitude.

Was this helpful?

You can pick up to 3 reactions.

Comments

  • Be the first to comment.

Popular calculators

All Calculators