 The size of the sample Sample size and precision Estimating a percentage Estimating a mean

Sample size and precision

The precision of your estimates is determined by the sample size. There is a simple rule: an increased sample size will lead to more precise estimates. You can use the tool below to compute the minimal sample size to obtain a required precision.

The first step is to select the confidence level. You can choose between 95% or 99%.

Note that a larger confidence level will reduce the risk of drawing a wrong conclusion, but will also produce wider confidence intervals. So you have two options: either you draw less precisie conclusions with high conficence, or you more precise conclusion with less confindence.

 95% confidence level 99% confidence level

The next step is to select what kind of estimate you want to compute. You can choose between estimating a population percentage or a population mean:

Estimating a percentage
To be able to compute the sample size for estimating a percentage, some information is required. You have to enter this information in the input fields below. Note that you have to fill in all fields.

 The size of the population. If you do not know the population size, enter some estimate. The maximum margin of error you allow. The is the largest difference between the estimate and the true value of the percentage. Global indication of the percentage you want to estimate. If you have not the slightes idea, enter the value 50.

After you have entered all required information, click on the button Compute below.

 The sample size should at least be equal to:

Estimating a mean
To be able to compute the sample size for estimating a mean, some information is required. You have to enter this information in the input fields below. Note that you have to fill in all fields.

The size of the population. If you do not know the population size, enter some estimate.
The maximum margin of error you allow. The is the largest difference between the estimate and the true value of the mean.
A global indication of the standard deviation S. Select one of the options below using the information you have about the shape of the population distribution of your target variable. Enter the value described by the selected option.

 You have a global indication of the value of S, for example from a different survey. Enter the value S. The variable has an approximately normal distribution over an interval of known length L. Enter the value of L. The variable has an approximately homogeneous distribution over an interval of known length L. Enter the value of L. The variable has an approximately exponential distribution over an interval of known length L. There are many small values and a few large values. Enter the value of L. The value has a completely unknown distribution over an interval of known length L. Enter the value of L.

After you have entered all required information, click on the button Compute below.

 The sample size should at least be equal to: 