Most marketers understand that quantitative research delivers “statistically valid” data that can be “projected” onto the market. But, how can you be confident that your survey results truly are statistically significant? The key lies in determining an acceptable level of accuracy, which requires a delicate balance between the optimal sample size and the budget for the specific research project.
The total size of the market population,?? or the universe, as researchers often call it,?? drives how large the sample size needs to be in order to achieve an acceptable degree of accuracy. Professional researchers use a mathematical formula to determine a representative sample size that will reflect the opinions and behavior of the group from which it is drawn.
Whenever possible, researchers seek sample sizes that will deliver a 95% confidence level,?? meaning that if you repeated the survey 100 times, 95 times out of a hundred, you would get the same response. From there, they’ll typically factor in a margin error of no more than a plus/minus 5 percentage points, which is the expected variation resulting from not surveying the entire market. So, for example, if the target population is 250,000 people, you can survey 400 people and stay within the +/-5% margin of error.
The numbers game behind the sample size gets a bit more complicated when the study calls for analysis and comparisons of sub-groups. For example, perhaps your company needs to compare survey data across five different geographic regions, or five different government agencies. The five different cross-tab cells require a minimum of 30,?? 40 respondents in each for the comparisons to be statistically valid. The more crosstab cells a study requires, the larger the total sample size must be.
It’s your research partner’s job to guide you on how many people need to be surveyed to get valid results from your quantitative study. At Market Connections, we frame this advice within clients’ budget parameters, acceptable margins of error and research objectives, along with our experience conducting quantitative studies in markets similar to theirs. It’s best to get this guidance before finalizing your research budget to ensure you’re being realistic. After all, the only result of conducting a quantitative study with an insufficient sample size is unreliable data.