Skip to content
Omnikalc

Taylor’s power law and sample size calculator

Enter the coefficients a and b of Taylor’s power law, the mean density, the precision wanted and the confidence level to get the expected variance, the coefficient of variation, the aggregation, the number of samples needed and the error with a given sample.

Updated on
Your inputs

Value out of range

Value out of range

Value out of range

%

Value out of range

Value out of range

Results

Expected variance a·m^b

45.948

Samples needed for the precision wanted

69

Standard deviation per sample
6.778
Coefficient of variation per sample
84.7 %
Variance-to-mean ratio (above 1 means aggregated)
5.743
Standard error of the mean with the samples taken
1.238
Relative half-width of the interval with these samples
30.3 %

The Taylor’s power law calculator gives the expected variance of counts from the mean density and the coefficients a and b, the coefficient of variation, the variance-to-mean ratio and the number of samples needed for a given precision and confidence level, with the error of a chosen sample size.

How the calculation works

  • Enter the coefficients a and b of the power law and the mean density.
  • Enter the precision wanted and choose the confidence level.
  • Enter the number of samples taken and read the sample size and the error.

Example

With a = 2.5 and b = 1.4 at a mean density of 8, the variance is 45.948, the standard deviation 6.778 and the coefficient of variation 84.7 %. A precision of 20 % at 1.96 confidence needs 69 samples. With 30 samples, the standard error of the mean is 1.238 and the interval half-width is 30.3 % of the mean.

Why it matters

The calculator helps entomologists, agronomists, crop consultants and field ecologists who plan sampling of pests, weeds, fish or plankton, want to know how many traps or plots are needed for a given precision, or check the efficiency of a sampling plan.

Good to know

The sample size is for the half-width of the interval of the mean, as a share of the mean, and not for the standard deviation. The coefficients a and b are specific to the species, the sampling unit and the habitat, so use values measured under similar conditions. The result is rounded up to a whole sample. The formula assumes that the sample mean is roughly normal, which needs a few tens of samples. A higher density usually needs fewer samples.

The Morisita calculator measures the spatial pattern, the index of dispersion calculator tests variance against the mean, the capture-recapture calculator estimates population size and the sample size calculator treats general surveys.

Frequently asked questions

What is Taylor’s power law?

Taylor’s power law (1961) states that the variance of the counts of organisms in samples is proportional to a power of their mean: variance = a × mean^b. It is found for thousands of species, from aphids to fish. The exponent b describes the aggregation: 1 for random (Poisson), above 1 for aggregated and below 1 for regular. Most species have b between 1 and 2.

How is the sample size computed?

The number of samples needed for a half-width of D (as a share of the mean) at the confidence z is n = z² × a × m^(b − 2) / D², a result of Karandinos (1976). With a = 2.5, b = 1.4, a mean of 8, D = 20 % and 95 % confidence, about 69 samples are needed. Sampling plans for pest management are built this way.

Why does the sample size depend on the density?

When b is below 2, the relative variability falls as the mean rises, so fewer samples are needed to estimate the mean precisely at high density than at low density. When the population is rare, many samples are needed. A sequential sampling plan stops as soon as the estimate is precise enough.

How do I get a and b?

Collect counts from several fields or dates covering a range of densities. For each, compute the mean and the variance, take their logarithms and fit a straight line: log variance = log a + b × log mean. The intercept gives a and the slope b. Values from the literature for the species and crop can be used for a first plan.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.