The question of how many people you should recruit to take your survey is one that comes up a lot.
It comes down to what you want to use the survey results for, and how serious you are about the data you’re collecting.
Generally people I work fall into one of two camps:
1. Need some quick informal/ directional feedback
If you just want quick informal feedback then in general anywhere from 10-100 will do. 50 is good, 100 is better, and anything over 100 just gives added reliability.
My most popular gig addon is the +80 responses, which yields a 100 person sample, which is probably ideal for most people’s needs.
2. Need highly reliable data that will support a business plan or academic study
In this case, you’re going to need to run some math, to ensure that your survey results are representative and extrapolatable to your target population.
We’ll have to discuss some statistics concepts here, which may not be immediately clear at first, but are important to understand to answer this question for yourself:
Your confidence interval, or margin of error, is what you normally see in political polls that say how confident the survey results are to extrapolate out to the entire population. For example, if 52% of people prefer toothpaste brand B, and your confidence interval is +-5, then you can be sure that if you had asked the entire population, anywhere from 47% to 57% of people would prefer toothpaste brand B.
Your confidence level is what it sounds like – how confident you can be sure of the results. It represents how often the true percentage of the population who picks an answer would live int he confidence interval. So for instance a 95% confidence level, means you could be 95% certain. Most market research studies go with a 95% confidence level, although with a large enough sample, you could go to 99%, or certainly do less than 95%.
If you need a statistically valid survey, then please contact me, so I can help you calculate the correct number to be representative of your target population.
Leave a Reply