Survey sample size: What it is and how to calculate it

While planning a survey, you may have encountered the term ‘sample size’. The sample represents your overall population, and calculating the size ensures you reach out to enough people and ensure the overall population is represented. Calculating the sample size comes in handy when you’re short of time, and your target audience is large; waiting for 100,000 customers to respond and analyzing the data can be tedious. Survey tools and marketing platforms might suggest upgrading in order to reach out to so many customers 🙂

One quick and easy way to calculate your sample size is by using Slovin’s formula.

A sidebar in statistics

If this section puts you off, we’re not surprised, however, bare with us as this is important while calculating the sample size. Confidence level and error tolerance are two important things to remember while calculating your sample size.

The confidence level measures how sure you are that subsequent surveys of the same sample would yield similar results. A confidence level of 100% means the responses will be the same even on repetition. Needless to say, you’d never publish your results if your confidence level is 0.

Since people are rarely consistent 100% of the time, expecting our statistics to work the same is unrealistic. To account for this, we introduce a margin of error. The more confident you are in your results, the smaller your margin of error. 

The error tolerance level looks at the probability of being wrong, considering the confidence level. Error tolerance is 1- the confidence level. If the confidence level is 98%, the error tolerance is 1-(98/100) or .02

Calculating the sample size with Slovin’s formula

Slovin’s formula, which may or may not have been proposed by the statistician, Slovin, who may or may not have existed (seriously, look it up), is best used when little is known about the distribution of the population’s behaviour (does anybody else sense the irony here?).

Here’s Slovin’s formula: 

Slovin's formula for calculating the sample size: n = N/(1+ N.e^2) where n = sample size N = population size e = error tolerance level

Put in action, let’s say your segment has 100,000 people, and your error tolerance level is .05. The right sample size for your survey would be 100,000/1 + 100,000 x .0025 = 398.4. Round it off because you’re not going to be able to survey .4 of a person, leaving you with a sample size of 398, which is the number of responses you should aim to have. With the sample size calculated, all that’s left is reaching out to your target audience.

Check out our companion blog here to learn more about identifying the right target audience.