How many response do you need for your survey?
Whether your using our Surmetric platform or just conducting informal surveys at school or work, you will likely have to determine how many responses you need to collect before you have a statistically valid sample. The number of responses required depends on two basic factors. How large is your population and how accurate does your survey need to be?.
The population size is the total number of individuals you consider to be relevant to your study. The difficulty in coming up with this number can vary greatly. Sometimes it is very well defined, “Number of customers who purchased product XYZ”. Other estimates can be more difficult, “Number of customers who shopped but did not buy anything”. Both of these populations could provide valuable insights.
The accuracy is usually expressed as a +/- margin of error. The inverse of the margin of error is the confidence level. An error margin of +/-5% has a confidence level of +/-95%. Think of confidence level as being the likelihood that if the survey was conducted numerous times using random samples, the results would statistically be the same. A margin of error between 3% and 10% is the norm.
As you can see from the table above, the number of response for populations above 10,000 do not fluctuate. The one thing the table does not capture is the validity of the collection mechanism. Ideally all respondents should be randomly selected and the study should take care to remove any bias introduced by the collection process.