In a recent blog post we talked about four types of sampling methodologies frequently used in market insight research. In this post we’ll explore how sample sizes are determined and how to calculate for sampling errors.
Each qualitative market research project is distinct, and bigger isn’t necessarily better. While adding more people in a sample population may yield more accurate results, given the proportionate rise in costs to conduct a larger study, researchers consider four values to look for ways to maximize the accuracy while minimizing the overall project costs. The four values to define before determining the appropriate sample size for your study:
- Population Size—Approximately how many people fit your target demographic? Population size will tell you WHO you should survey to gain the types of insights you are looking for to carry out market research. For example, if you’re conducting a market study to gauge interest in the launch of a new women’s shoe collection in North America, you would want to know approximately how many women live in geographic area who fit the parameters of the shoe sizes.
- Margin of Error—Also referred to as the confidence interval, this number indicates how much lower or higher you are willing to let your sample mean fall. You will likely recognize this during political campaign season, when many polls and surveys are conducted and reported with a margin of error or +/- x%.
- Confidence Level—This percentage tells you how sure you can be that the mean falls within your margin of error or confidence interval. The more common confidence intervals fall between the range of 90-99%. Confidence levels have a related Z-score, which is a constant value required to determine this equation. For example, a 90% confidence level has a Z-score of 1.645and a 99% confidence level has a Z-score of 2.576.
- Standard of Deviation—When setting up sample populations to survey, a common number used is .5 because it ensures that your sample will be large enough.
With the above four values defined, you can now calculate the needed sample size with the following formula: Sample Size= (Z-score)² * StdDev*(1-StdDev) / (margin of error)².
If you plug in your numbers and find that the suggested sample size is too large, you can adjust by decreasing your confidence level or increasing your margin of error. The goal is to eventually get a sample population that correctly reflects your target demographic so that qualitative research can be conducted.
Once the qualitative research consultant feels confident with the sample size required to carry out the study, s/he will often hire a nationwide recruitment agency to find participants who meet the criteria of the study.