Select Page
Report an error

4.12 How big should my survey sample be?

Imagine if a million people walked through Lobster Land’s doors every year. Would it be possible to ask every single person for their opinion about the park? The answer is, quite simply, no. So the next logical question is, “How big should my sample size be?” Well, it depends on the following:

  • The margin of error you are willing to accept   
    • Since it is not possible to get the opinion of every single visitor to Lobster Land, or to test a new drug on all suitable patients, we have to extrapolate from a smaller sample. Naturally, this means there will be errors in our reporting. This is what is known as the margin of error.  It is always reported in survey results.
    • Scenario: 90% of visitors polled said they would like Lobster Land to start selling chocolate ice cream. The margin of error is +-3%, the confidence level is 95%. What do these findings mean?
    • It means that if you had asked ALL of the park’s visitors the same question, you can be 95% confident that between 87% and 93% of them would have given the same answer.
    • Please note that the margin of error is not to be confused with a confidence interval. It is equal to half the width of the confidence interval.
  • The alpha level (confidence level) you set
    • The confidence level tells you how confident you can be that the entire population would have selected an answer within a certain range.
    • The higher the sampling confidence level, the larger your sample size needs to be.
  • Population size
    • There are many formulas available to calculate the ideal sample size depending on what you know about the population. When a population size is known, one can take advantage of an online sample size calculator to determine the ideal sample size. When a population size is unknown, researchers may use Cochran’s formula given a desired level of precision, confidence level, and the estimated proportion of the attribute present in the population.

While a bigger sample size will mitigate the margin of error and generate a more accurate picture, beyond a certain point, the increase in accuracy is so small it is not worth the time and expense required to collect more data. That said, margins of error and survey findings are only reliable if the sampling is well conducted.