How to identify your population of interest

Read on to learn more about the concept of a population of interest, why it is important, and how to find yours. Or, if you’d like us to do the legwork, we can do that too! Feel free to reach out to one of our experts and we’ll have a population of interest created for you in a flash. 

So, what do we mean when we refer to a population of interest? Simply put, a population of interest is a specified group of individuals or institutions from whom you are attempting to draw conclusions. However, don’t be confused by the word population. Your population of interest is a subset of the general population (or population of inference) that shares some common characteristics of interest to you. 

For example, if you’re a business to consumer (B2C) retailer of freshly cut flowers, you want to know the types of floral designs your customers love, and those that are less appealing to them,  your population of interest is probably people who regularly purchase floral bouquets. If, on the other hand, you’re a business to business (B2B) vendor of information technology services and you want to know more about your customer segment, your population of interest might be small and medium sized enterprises (SMEs). 

One question you might have is: what are the parameters of the population of interest? In other words, what are the inclusion and exclusion criteria, and who should be incorporated and left out? Defining the breadth of your population of interest can be a challenge, but your starting point should be the characteristics, qualities and traits shared by the population from whom you want to gather information or who you want to learn more about. So, when thinking about who to include in your population of interest, you’ll first need to start thinking about their likely traits and characteristics. We have some more guidance below on the specific steps you should take to define your population of interest. But first, let’s take a look at the importance of choosing the right population of interest.

Imagine you’re the owner of a grocery store, thinking of moving into online sales. In order to make sure that there’s a market for online grocery shopping among your existing customers, and to identify their needs, you decide to conduct a survey. You design a survey and hand one to every individual coming into the store between 9am-11am on a Tuesday morning. 

What is the likelihood that the findings from this research study will tell you what you need to know about the likely demand and preferences for online grocery shopping? You might be surprised, but in fact, a survey like this might not tell you very much. Why? Probably, the population of interest is not quite right in this case. That’s because people who do their weekly shopping at the grocery store on Tuesday mornings likely differ in demographic terms from people who do their shopping online. For instance, seniors tend to do their shopping on midweek mornings, while busy professionals—a demographic who are more likely to express a demand for an online shopping option tend to leave their shopping days until the weekend. Gathering data from midweek shoppers could lead you to the conclusion that there is little demand for your proposed online solution, when in fact, there is—just among a different population.

This is why it’s especially important to choose the right population of interest. Gathering data from the wrong population of interest can lead you to erroneous or skewed conclusions. It can mean the difference between a useful, valid research study, and an expensive waste of time.  

Let’s take a deeper look at the steps involved in selecting a population of interest, so you can make sure you get yours right first time.

We recommend taking the following steps when identifying a population of interest.

As our example of the grocery shoppers shows, one of the most important steps to take when choosing a population of interest is establishing the population parameters. In other words, you will need to define and agree upon a set of criteria that you expect each member of the population to share. Researchers tend to determine both inclusion criteria and exclusion criteria when defining their population of interest.

Criteria that all members of the population of interest must share. For example, if you are trying to ascertain the level of demand for an online grocery sales survey, one criterion that you might determine to be important is that the population of interest contains individuals who assume primary responsibility for grocery shopping in their respective households.

Criteria that exclude individuals from being included in your population of interest. It is important to remember that exclusion criteria are not just the opposite of inclusion criteria. Rather, there may be individuals or institutions that meet your inclusion criteria but should still be excluded from the population of interest because of some other characteristic. In our grocery store example, you might exclude senior citizens from the population of interest because even if they're responsible for doing the shopping in their household because of the low takeup of internet shopping among this population more generally. 

The next stage of the process is to apply these criteria in advance. This means taking steps to ensure that the sample from which you gather data is the same as your population of interest. 

Our grocery store owner could have taken the step of surveying shoppers on a Saturday morning instead of a Tuesday morning in order to reduce the chances that the sample included the wrong individuals. However, if you’re using an online survey approach, you can also build in filter questions to make sure that your population of interest meets your inclusion criteria, and that individuals who do meet the exclusion criteria are removed from the sample. At the beginning of your survey, you can ask respondents explicitly whether they meet your criteria, and those that do not can be directed to the end of the questionnaire and thanked for their time. 

Finally, once you’ve established the population of interest, it's time to use an appropriate sampling technique to extract a sample for inclusion in the study. Although the population of interest is a subset of the population of inference, it will still probably be too large for you to feasibly gather data from each and every member. That’s where sampling comes in. 

The grocery store owner might, for instance, survey every 10th customer who comes into the store on a Saturday morning, or only those that use trolleys instead of handbaskets. As we’ll see below, the precise approach you use will depend on your research needs and have implications for your research findings.

The process of how to choose a sample set from the population of interest is loosely as follows:

Decide on your exclusion and inclusion criteria, and make sure you have a clear justification for them. For instance, if you’re interested in customers’ experiences of impulse buying and buyers’ remorse, you might decide only to include participants that have had any experience of an impulse buying situation within the preceding four weeks in order to maximize the chances that respondents have good recall of their expectations and perceptions of historic purchasing experiences.

You should always expect that your estimates of a sample are inexact. It is rarely possible to know with any certainty the size of the sample, so assume a certain margin of error. 

Relatedly, recognize that populations do vary in terms of size. If your population of interest is large, this does not necessarily mean that you need to gather a very large sample, if the population is relatively stable.

Finally, recognize that a significant proportion of your sample is not going to respond, no matter how many times you contact them and not matter how many incentives you offer. For some online surveys, a response rate of 20% is acceptable. You may need to increase your sample size in order to accommodate high levels of refusal.

As we outlined above, it is rarely practical to gather data from each and every member of your population of interest, and samping is usually required. Sampling renders a mechanism for gathering data without surveying the whole targeted population. This is cheaper and more time-saving than surveying the entire population. Furthermore, taking a sample will reduce survey fatigue among your population of interest, which increases the chances that members will complete your survey—especially useful if you will be conducting multiple studies of the same population of interest. Remember that your population of interest is the entire unit of people you consider for the study, while your sample is a subset of this group that represents the population. 

When using sampling, you will usually gather data from the sample, and then generalize the conclusions to the entire population. 

There are two main types of sampling techniques: probability or random sampling procedures and non-probability sampling procedures. Within each category, there are different approaches that can be used. Let’s take a look at a few examples. 

A probability sampling procedure is one in which every member of the population has an equal chance of being included in the final sample. It is considered by researchers to be the paragon approach because it yields a sample that is representative of the wider population This is important because there is a close relationship between the extent to which the sample is representative, and the generalizability of the conclusions that are yielded from the study. 

Some types of sampling approach are:

Using simple random sampling, every population member has an equal chance of being included in the final sample, because each member is selected completely randomly. Some researchers like to use a number generator to help with simple random sampling.

Cluster sampling first organizes population members into clusters on the basis of some shared characteristic. For example, you might cluster the sample on the basis of income level, or household size. Then, a random sample is selected.

Using this approach, you would select every nth individual from your population of interest.  For example, you might decide to send a survey to every 10th person in a list sorted by alphabetical order.

Stratified random sampling involves first dividing your population of interest into predefined, distinguishable segments before selecting a random sample from each. 

Sometimes it is not possible to use a random sampling procedure. Importantly, probability sampling procedures require there to be a complete, comprehensive and correct population database from which the random sample can be identified and drawn. This is rarely possible in marketing research, and particularly that examining consumer behavior. The alternative to the probability sampling procedure is the non-probability sampling procedure.  This procedure does not yield a sample that is representative of the population because an unsystematic procedure is used to identify the sample. Nevertheless, because it is more convenient and practicable, much business and marketing research draws on non-probability methods of sampling. 

Using a convenience sampling procedure, you select a sample of subjects not because they're representative but because it is convenient to use them. For example, customers who happen to come into your store are more convenient for selection in a survey than locating prospective customers who have never visited your store before. As you might expect, the key benefit of this procedure is the convenience of the method.

In judgmental sampling, you will develop your sample on the basis of your specific needs. For instance, if you are looking to survey vegans, you will only include individuals in your sample who follow a vegan diet. In this way, the sample is constructed based on the purpose of the research, hence this approach is often called purposive sampling.

A snowballing strategy is where an initial respondent is first contacted and asked to recommend others who are willing to take part in the research. Over time, the size of the sample grows, similar to a snowball increasing in size as it rolls down a hill. 

Quota sampling means establishing quotas of the number or proportion  of individuals of different characteristics for inclusion in the study. For example, you might want a 50/50 split of women and men, or you might want to include a certain proportion of respondents who live in rural areas. The sample is constructed until the quota is met.

Ready to identify your population of interest? Use SurveyMonkey’s Audience panel solution to help construct the parameters quickly and easily. If you already have a population in mind, we can help with that too. Get started today.