Producing Data

3. Sources of data

3.2 Generating your own data

3.2.2. Generating data from observations

viii. Margin of error

When we perform sampling, we can calculate an estimate of how 'close' we are to the true result in the population. This is called the margin of error.

In theory, in 19 out of 20 cases, the results based on samples will differ by no more than 3% in either direction (± 3%). For smaller focus groups the potential error is greater. For example, if you wanted to find out about university students, the sampling error would probably be greater than ± 3%.

Margin of error can be described as the amount by which the percentage or proportion obtained from the sample, a sample statistic, will differ from the population percentage or proportion. As variation decreases, we can have greater confidence that the sample mean is closer to the population mean. A rough guide to margin of error can be obtained by dividing one (unity) by the square root of the number of units in the sample [6]:

Example: If you have a sample size of 150, then your margin of error would be:

If, according to your survey, 56% of people want something to happen, you can be fairly confident that your population percentage falls between 56 ± 8, that is, in a range from 48% to 64%. This range of values, 48% - 64%, represents a confidence interval.

The idea of sample statistic ± margin of error represents a confidence interval. Increasing the sample size reduces the margin of error, because approaches zero the larger n becomes.

NOTE: You might be starting to realise that the size of the sample is important in identifying the margin of error, NOT what percentage the sample is of the population.

Next - Menu