Sample size and significance tool

Click here for free tools and know-how materials from the Mud Valley™ strategy & brand marketing community.

Question - how do I know how many people I should interview to obtain statistically significant results?

Answer - the issue is less whether the results are significant, more whether the difference in the results between two surveys is significant. The general rule is that the larger your sample size, the more sensitive it is to recognising small differences.

Key action points

N.B. - as far as we know this is about the only tool in the market that does this simple task simply. At Mud Valley, we find it invaluable, and use it constantly.

There is a general misconception that statistical significance tests tell you whether the results of a survey are significant or not.

What statistical significance testers actually tell you is:

• the likely margin of error in the results, as in "there is a 5% plus/minus on your result of 65%, so the real-world result could be anywhere between 60% and 70%"
• whether the results of the second survey are statistically different from those of the first survey

In more detail ....

This two sample significance calculator will tell you whether there is a significant difference between the results of two surveys. So if the first survey said that 65% of the sample had heard of you, then the next survey said 70% of the sample hard heard of you, have you increased your awareness levels in the market, or is the difference in the results within the statistical margin of error?

By turning the tool "on its head", you can also judge what sample size you need by putting in mock results to see when the differences become significantly different as you increase your sample size.

For this calculator you need only two pieces of information: