In this two-part series, we’ll explore the techniques and methodologies of sampling populations for market research and look at the math and formulas used to calculate sample sizes and errors.
Market research wouldn’t be possible without sampling, as it’s impossible to access every customer, whether current or future. Market researchers rely on various sampling techniques and methods to try and capture as wide range as possible the various types of customers a client is hoping to glean feedback from. Now, you may be thinking that all sampling is bunk, especially given the predictions and outcomes of both the Brexit vote and the recent U.S. Presidential election. Keep in mind that polling is different than sampling, and when market research is being carried out, more than simple questions are being asked of its sample population.
InterQ works closely with its clients to understand their objectives and then create sampling groups appropriate to the objective. We find that the best, most beneficial feedback is gleaned through a combination of qualitative AND quantitative research. Sampling methods are crucial to the quality of research, which is one of the reasons why this is better left to neutral, professional organizations, rather than done “in-house.” Choosing the right sampling technique is important so that data isn’t skewed or biased. Let’s explore sampling in more detail.
Sampling methodologies can be boiled down into two groups: probability and non-probability.
Probability (random) sampling methods allow all members of a target population to be included in the sample and isn’t encumbered by previous events in the selection process. Put another way, the selection of individuals for a sample group doesn’t affect the chance of anyone else in the targeted population to be selected. So how does a market research company go about selecting people to be included in a study? There are a number of random sampling techniques that market researches can employ, but four types of commonly used techniques include: Simple Random Sampling, Systematic Sampling, Cluster Sampling and Stratified Sampling.
Simple Random Sampling—The most commonly used sampling technique, and truly random, this method randomly selects individuals from a list of the population, with every individual having an equal chance at being selected.
Systematic Sampling—Rather than randomly selecting individuals from a population, this method is based on a system of selecting participants. For example, a market researcher may select from a list of the population every 20th person. While this allows for a controlled way to select from a target population, it may be skewed depending on how the original list is structured or organized.
Cluster Sampling—Cluster sampling is a variation on Simple Random Sampling and is often used with larger populations and across a broader geographic region. Typically, a population is segregated into clusters and then participants are randomly selected from these groups.
Stratified Sampling—This method is a conflation of Simple Random and Systematic Sampling and is often used when there are a multitude of unique subgroups that require full, randomized representation across the sampling population.
Non-probability sampling methods are less desirable and often contain sampling biases. So why would anyone choose this methodology? Budget and lack of access to a full population list are often the reason. If a researcher must go with a non-probability sampling method, he/she must be very careful when drawing conclusions, as the population is not randomized and biases inherent.
Most organizations hoping to learn more about their target populations understand that hiring third-party market research companies that are well-versed in understanding and selecting sampling populations based on the methodologies outlined above is money well spent. Market research, when done properly, is often the difference between good and great outcomes.
If you’re interested in learning more about market research, check out our qualitative research training programs from InterQ Learning Labs.