Neyman n_h = n·N_h·σ_h/Σ
- Created by
- Renato Passos, Eng. de Software
- Reviewed by
- Renato Passos, Eng. de Software
Last updated: Apr 18, 2026
About this calculator
This calculator determines the optimal sample size for each stratum in a stratified sampling study using Neyman allocation. The method distributes the sample proportionally to the size and variability of each stratum, minimizing the variance of the total estimate. The formula used is n_h = (n × N_h × σ_h) / Σ(N_h × σ_h), where n_h is the sample size for stratum h, N_h is the population size of the stratum, and σ_h is the standard deviation of the variable of interest in that stratum.
The tool is ideal for heterogeneous populations that can be divided into distinct subgroups (strata) with varying levels of variability. For instance, in market research or epidemiological studies, where representative samples must be collected from the most dispersed groups. This enhances the accuracy of final estimates, especially in resource-constrained scenarios.
Ensure that N_h and σ_h values for each stratum are known beforehand. Neyman allocation requires a clear understanding of population structures and variability patterns. If a stratum has a standard deviation of zero, it will not contribute to the allocation since there is no variability to capture.
This approach is preferable to methods like equal allocation (same number of samples per stratum) when strata differ significantly. However, if all strata have the same variability or size, the result will resemble proportional allocation. Always round final values to whole numbers, adjusting as needed.
Frequently asked questions
What is this calculator used for?
It determines sample sizes for each stratum in stratified sampling using Neyman allocation, prioritizing larger and more variable strata.
When to use instead of proportional allocation?
Use it when strata have unequal variability. Optimal allocation reduces total variance, while proportional allocation ignores variability.
What if the standard deviation for a stratum is unknown?
Use historical estimates or pilot tests. Missing σ_h may invalidate optimal allocation.
Why round the results?
Sampling requires whole numbers. Round calculated values and adjust remainders to keep total n.
Does the method work with few strata?
Yes, but it's most effective with three or more strata. With only two, the difference between proportional and optimal allocation will be smaller.