Market research sampling is the process of taking the entire population you want to describe — such as all U.S. residents over the age of 18 — and selecting a smaller portion to represent this population. If sampling is done properly, then the market research study can save time and expense while still accurately describing the population as a whole. The same principle is applied to quality control — since inspecting every item coming off an assembly line in detail would be impractical, the quality control department selects a sample of one out of X completed units to undergo thorough testing.
Learn how your market research team can benefit from smart sampling methods and what possible sampling methods they have at their disposal by reading on:
Ideally, research samples should be both representative and chosen at random.
Representative samples conform to demographic trends in the actual data. For instance, only around 30 percent of Americans have a bachelor’s degree or higher. If your sample is supposed to represent the U.S. population, but 60 percent of it has college degrees, you will end up with a non-representative sample.
Some populations will have different demographics than larger populations as a whole, though, so if you intend to sample Boulder, Colorado and you do end up with 60 percent college graduates, then that is okay! Make sure to review the expected demographics and weigh your sample’s demographics against it to determine viability.
Another important concept in sampling is randomness. Without having a somewhat random selection process, the potential for selection bias or flawed sampling can occur. For this reason, true random selection is prized but not always attainable.
One inherent source of bias or misrepresentation is to begin selecting your sample from a population list that is itself not representative. An excellent example would be a study that wants to compare whether single people eat more red meat than dating couples. If the sample is chosen only from a list of couples who entered a particular vegan restaurant, then the results will likely state that single people eat overwhelmingly more red meat than couples.
Obtaining high-quality, representative lists can be difficult, providing a sticking point for many market research projects. Try to find government-sourced lists or third-party lists procured from a healthy mix of sources, and always evaluate whether one source of population identities was used too heavily.
Make sure that your sample size is large enough to account for your desired margin of error, and avoid hand-picking your individual samples if you can avoid it.
Once you have identified the correct sample for your survey, it is time to start screening respondents to determine if they meet the sample. For that, check out our handy guide to demographic and screening surveys. There's a lot of great practical info there about how to obtain the sample you want--plus a free screening survey template.