What Is Non-Probability Sampling? | Types & Examples

Non-probability sampling is a sampling method that uses non-random criteria like the availability, geographical proximity, or expert knowledge of the individuals you want to research in order to answer a research question.

Non-probability sampling is used when the population parameters are either unknown or not possible to individually identify. For example, visitors to a website that doesn’t require users to create an account could form part of a non-probability sample.

Note
Be careful not to confuse probability and non-probability sampling.

  • In non-probability sampling, each unit in your target population does not have an equal chance of being included. Here, you can form your sample using other considerations, such as convenience or a particular characteristic.
  • In probability sampling, each unit in your target population must have an equal chance of selection.

Types of non-probability sampling

There are five common types of non-probability sampling:

Convenience sampling

Convenience sampling is primarily determined by convenience to the researcher.

This can include factors like:

  • Ease of access
  • Geographical proximity
  • Existing contact within the population of interest

Convenience samples are sometimes called “accidental samples,” because participants can be selected for the sample simply because they happen to be nearby when the researcher is conducting the data collection.

Example: Convenience sampling
You are investigating the association between daily weather and daily shopping patterns. To collect insight into people’s shopping patterns, you decide to stand outside a major shopping mall in your area for a week, stopping people as they exit and asking them if they are willing to answer a few questions about their purchases.

Quota sampling

In quota sampling, you select a predetermined number or proportion of units, called a quota. Your quota should comprise subgroups with specific characteristics (e.g., individuals, cases, or organizations) and should be selected in a non-random manner.

Your subgroups, called strata, should be mutually exclusive. Your estimation can be based on previous studies or on other existing data, if there are any. This helps you determine how many units should be chosen from each subgroup. In the data collection phase, you continue to recruit units until you reach your quota.

Tip
Your respondents should be recruited non-randomly, with the end goal being that the proportions in each subgroup coincide with the estimated proportions in the population.

There are two types of quota sampling:

  1. Proportional quota sampling is used when the size of the population is known. This allows you to determine the quota of individuals that you need to include in your sample in order to be representative of your population.
Example: Proportional quota sampling
Let’s say that in a certain company there are 1,000 employees. They are split into 2 groups: 600 people who drive to work, and 400 who take the train.

You decide to draw a sample of 100 employees. You would need to survey 60 drivers and 40 train-riders for your sample to reflect the proportion seen in the company.

  1. Non-proportional quota sampling is used when the size of the population is unknown. Here, it’s up to you to determine the quota of individuals that you are going to include in your sample in advance.
Example: Non-proportional quota sampling
Let’s say you are seeking opinions about the design choices on a website, but do not know how many people use it. You may decide to draw a sample of 100 people, including a quota of 50 people under 40 and a quota of 50 people over 40. This way, you get the perspective of both age groups.

Note that quota sampling may sound similar to stratified sampling, a probability sampling method where you divide your population into subgroups that share a common characteristic.

The key difference here is that in stratified sampling, you take a random sample from each subgroup, while in quota sampling, the sample selection is non-random, usually via convenience sampling. In other words, who is included in the sample is left up to the subjective judgment of the researcher.

Example: Quota sampling
You work for a market research company. You are seeking to interview 20 homeowners and 20 tenants between the ages of 45 and 60 living in a certain suburb.

You stand at a convenient location, such as a busy shopping street, and randomly select people to talk to who appear to satisfy the age criterion. Once you stop them, you must first determine whether they do indeed fit the criteria of belonging to the predetermined age range and owning or renting a property in the suburb.

Sampling continues until quotas for various subgroups have been selected. If contacted individuals are unwilling to participate or do not meet one of the conditions (e.g., they are over 60 or they do not live in the suburb), they are simply replaced by those who do. This approach really helps to mitigate non-response bias.

Self-selection (volunteer) sampling

Self-selection sampling (also called volunteer sampling) relies on participants who voluntarily agree to be part of your research. This is common for samples that need people who meet specific criteria, as is often the case for medical or psychological research.

In self-selection sampling, volunteers are usually invited to participate through advertisements asking those who meet the requirements to sign up. Volunteers are recruited until a predetermined sample size is reached.

Self-selection or volunteer sampling involves two steps:

  1. Publicizing your need for subjects
  2. Checking the suitability of each subject and either inviting or rejecting them
Example: Self-selection sampling
Suppose that you want to set up an experiment to see if mindfulness exercises can increase the performance of long-distance runners. First, you need to recruit your participants. You can do so by placing posters near locations where people go running, such as parks or stadiums.

Your ad should follow ethical guidelines, making it clear what the study involves. It should also include more practical information, such as the types of participants required. In this case, you decide to focus on runners who can run at least 5 km and have no prior training or experience in mindfulness.

Keep in mind that not all people who apply will be eligible for your research. There is a high chance that many applicants will not fully read or understand what your study is about, or may possess disqualifying factors. It’s important to double-check eligibility carefully before inviting any volunteers to form part of your sample.

Snowball sampling

Snowball sampling is used when the population you want to research is hard to reach, or there is no existing database or other sampling frame to help you find them. Research about socially marginalized groups such as drug addicts, homeless people, or sex workers often uses snowball sampling.

To conduct a snowball sample, you start by finding one person who is willing to participate in your research. You then ask them to introduce you to others.

Alternatively, your research may involve finding people who use a certain product or have experience in the area you are interested in. In these cases, you can also use networks of people to gain access to your population of interest.

Example: Snowball sampling
You are studying homeless people living in your city. You start by attending a housing advocacy meeting, striking up a conversation with a homeless woman. You explain the purpose of your research and she agrees to participate. She invites you to a parking lot serving as temporary housing and offers to introduce you around.

In this way, the process of snowball sampling begins. You started by attending the meeting, where you met someone who could then put you in touch with others in the group.

When studying vulnerable populations, be sure to follow ethical considerations and guidelines.

Purposive (judgmental) sampling

Purposive sampling is a blanket term for several sampling techniques that choose participants deliberately due to qualities they possess. It is also called judgmental sampling, because it relies on the judgment of the researcher to select the units (e.g., people, cases, or organizations studied).

Purposive sampling is common in qualitative and mixed methods research designs, especially when considering specific issues with unique cases.

Note
Unlike random samples—which deliberately include a diverse cross-section of ages, backgrounds, and cultures—the idea behind purposive sampling is to concentrate on people with particular characteristics, who will enable you to answer your research questions.

The sample being studied is not representative of the population, but for certain qualitative and mixed methods research designs, this is not an issue.

Common purposive sampling techniques include:

These can either be used on their own or in combination with other purposive sampling techniques.

Maximum variation sampling

The idea behind maximum variation sampling is to look at a subject from all possible angles in order to achieve greater understanding. Also known as heterogeneous sampling, it involves selecting candidates across a broad spectrum relating to the topic of study. This helps you capture a wide range of perspectives and identifies common themes evident across the sample.

Example: Maximum variation sampling
You are researching what first-year students think of their study program. You are more interested in nuance than generalizable findings, so you decide to pursue a qualitative approach.

You draw your sample using maximum variation sampling, including students who performed poorly, students who excelled, and students in the middle. You recruit and interview students until you have reached a saturation point.

Homogeneous sampling

Homogeneous sampling, unlike maximum variation sampling, aims to achieve a sample whose units share characteristics, such as a group of people that are similar in terms of age, culture, or job. The idea here is to focus on this similarity, investigating how it relates to the topic you are researching.

Example: Homogeneous sampling
You are researching the long-term side effects of working with asbestos. You determine “long-term” to mean 20 years or longer. Using homogeneous sampling, only people who worked with asbestos for 20 years or longer are included in your sample.

Typical case sampling

A typical case sample is composed of people who can be regarded as “typical” for a community or phenomenon. A typical case sample allows you to develop a profile of what would generally be agreed as being “average” or “normal.”

Typical case samples are often used when large communities or complex problems are investigated. In this way, you can gain an understanding in a relatively short time, even if you are not familiar with what’s going on yourself.

Example: Typical case sampling
Suppose you want to evaluate the level of care provided by physiotherapists to clients at a certain clinic. To develop a typical case sample, you interact closely with both therapists and clients in order to develop a set of criteria of what is “typical,” or average.

For physiotherapists, this could include years of professional experience, educational background, etc. For patients, criteria can include their age, or how often they have visited the clinic in the past year. By comparing the two typical case samples, you can conclude whether the average physiotherapist has the expertise needed to meet the average client’s needs.

Note that the purpose of typical case sampling is to describe and illustrate what is typical to those unfamiliar with the setting or situation. The purpose is not to make generalized statements about the experiences of all participants. In other words, typical case sampling allows you to compare samples, not generalize samples to populations.

Extreme (deviant) case sampling

Extreme (or deviant) case sampling uses extreme cases of a particular phenomenon (outliers). This can mean remarkable failures, successes, or crises, as well as any event, organization, or individual that appears to be the “exception to the rule.” Extreme case sampling is most often used when researchers are developing best-practice guidelines.

Note that extreme case sampling usually occurs in combination with other sampling strategies. The process of identifying extreme or deviant cases usually occurs after some portion of data collection and analysis has already been completed.

Example: Extreme (deviant) case sampling
You are studying serial killers. You identify a few cases where the serial killer was female. These cases are outliers, i.e., cases that stand out in your sample. In an effort to develop a richer, more in-depth understanding of the phenomenon, you decide to select these outliers and analyze them further.

Critical case sampling

Critical case sampling is used where a single case (or a small number of cases) can be critical or decisive in explaining the phenomenon of interest. It is often used in exploratory research, or in research with limited resources.

There are a few cues that can help show you whether or not a case is critical, such as:

  • “If it happens here, it will happen anywhere”
  • “If that group is having problems, then all groups are having problems”

It is critical to ensure that your cases fit these criteria prior to proceeding with this sampling method.

Example: Critical case sampling
You want to know how well people understand a new tax law. If you ask tax professionals and they do not understand it, then it’s likely laypeople won’t either. Alternatively, if you ask people from other professional fields, irrelevant to taxes or law, and they do understand it, then it’s safe to assume most people will.

In other words, your critical cases could either be those with relevant expertise or those who have no relevant expertise.

Expert sampling

Expert sampling involves selecting a sample based on demonstrable experience, knowledge, or expertise of participants. This expertise may be a good way to compensate for a lack of observational evidence or to gather information during the exploratory phase of your research.

Alternatively, your research may be focused on individuals who possess exactly this expertise, similar to ethnographic research.

Example: Expert sampling
You are interested in teaching methods for children with special needs in your district, and you want to conduct some exploratory research. Using expert sampling, you can contact special education teachers who work at schools in your district, gathering your data via surveys or interviews.

Non-probability sampling examples

There are a few methods you can use to draw a non-probability sample, such as:

Social media

Suppose you are researching the motivations of digital nomads (young professionals working solely in an online environment). You are curious what led them to adopt this lifestyle.

Since your population of interest is located all over the globe, it clearly isn’t feasible to conduct your study in person. Instead, you decide to use social media, finding your participants through snowball sampling.

You start by identifying social media sites that cater to digital nomads, such as Facebook groups, blogs, or freelance job sites. You ask the administrators for permission to post a call for participants with information about your research, encouraging readers to share the call with peers.

River sampling

You are part of a research group investigating online behavior and cyberbullying, in particular among users aged 15 to 30 in your state. You are collecting data in two ways, using an online survey.

You first place a link to your survey in an online news article about cyber-hate published by local media. Second, you create an advertising campaign through social media, targeted at users aged 15 to 30 and linking back to your survey. To entice users to participate, a prize draw (movie tickets) is mentioned in all ads. The survey and the campaign are active for the same length of time.

These two data collection methods are river samples. The name refers to the idea of researchers dipping into the traffic flow of a website, catching some of the users floating by.

Note
Keep in mind that river samples can suffer from coverage bias.

Street research

You are interested in the level of knowledge about myocardial infarction symptoms among the general population.

For a week, you stand in a shopping mall and stop passersby, asking them whether they would be willing to take part in your research. To try to allow as broad a range of respondents as possible to be included, you interview equal numbers of people from Monday to Friday during working hours.

Receive feedback on language, structure and formatting

Professional editors proofread and edit your paper by focusing on:

  • Academic style
  • Vague sentences
  • Grammar
  • Style consistency

See an example

Probability vs. non-probability sampling

Sampling methods can be broadly divided into two types:

  • Probability sampling: When the sample is drawn in such a way that each unit in the population has an equal chance of selection
  • Non-probability sampling: When you select the units for your sample with other considerations in mind, such as convenience or geographical proximity

Probability sampling

For many types of analysis, it is important that the statistical analysis is conducted from a random probability sample from the population of interest. For the sample to qualify as random, each unit must have an equal chance (i.e., equal probability) of being selected.

When you use a random selection method (e.g., a drawing) and ensure that you have a sufficiently large sample, your sample is more likely to be representative, and the results generalizable.

Example: Probability sampling
You are researching what motivates students to study medicine at a certain university. From student records, you see that there are 1,700 students enrolled in total.

Due to time restrictions, a sample of 150 students is deemed sufficient. After obtaining the student registration list, you use it as your sampling frame, and with the help of an online random number generator, you draw a simple random sample.

Non-probability sampling

Non-probability sampling designs are used when the sample needs to be collected based on a specific characteristic of the population (e.g., people with diabetes).

Unlike probability sampling, the goal is not to achieve objectivity in the selection of samples, or to make statistical inferences. Rather, the goal is to apply the results only to a certain subsection or organization. These are used in both quantitative and qualitative research.

Advantages and disadvantages of non-probability sampling

It is important to be aware of the advantages and disadvantages of non-probability sampling and to understand how they can play a role in your study design.

Advantages of non-probability sampling

Depending on your research design, there are advantages to choosing non-probability sampling.

  • Non-probability sampling does not require a sampling frame, so your subjects are often readily available. This can make non-probability sampling quicker and easier to carry out.
  • Non-probability sampling allows you to target particular groups within your population. In certain types of research, it is vital that certain units be included in your sample. For example, many kinds of medical research rely on people with a specific health issue.
  • Although it is not possible to make statistical inferences from the sample to the population, non-probability sampling methods can provide researchers with the data to make other types of generalizations from the sample being studied.

Disadvantages of non-probability sampling

Non-probability sampling has some downsides as well. These include the following:

  • Non-probability samples are extremely unlikely to be representative of the population studied. This undermines the generalizability of your results.
  • Non-probability samples are at risk of several kinds of bias:
    • As some units in the population have no chance of being included in the sample, undercoverage bias is likely.
    • Furthermore, since the selection of units included in the sample is often based on ease of access, sampling bias is common as well.
    • While the subjective judgment of the researcher in choosing who makes up the sample can be an advantage, it also increases the risk of researcher bias.

You can mitigate the disadvantages of non-probability sampling by describing your choices in the methodology section of your dissertation. Specifically, explain how your sample would differ from one that was randomly selected and mention any subjects who might be excluded or overrepresented in your sample.

Frequently asked questions about non-probability sampling

What is a sampling method?

When your population is large in size, geographically dispersed, or difficult to contact, it’s necessary to use a sampling method.

This allows you to gather information from a smaller part of the population (i.e., the sample) and make accurate statements by using statistical analysis. A few sampling methods include simple random sampling, convenience sampling, and snowball sampling.

What is a sampling frame?

A sampling frame is a list of every member in the entire population. It is important that the sampling frame is as complete as possible, so that your sample accurately reflects your population.

What is sampling?

A sample is a subset of individuals from a larger population. Sampling means selecting the group that you will actually collect data from in your research. For example, if you are researching the opinions of students in your university, you could survey a sample of 100 students.

In statistics, sampling allows you to test a hypothesis about the characteristics of a population.

What is stratified sampling?

In stratified sampling, researchers divide subjects into subgroups called strata based on characteristics that they share (e.g., race, gender, educational attainment).

Once divided, each subgroup is randomly sampled using another probability sampling method.

What is the difference between stratified and cluster sampling?

Stratified and cluster sampling may look similar, but bear in mind that groups created in cluster sampling are heterogeneous, so the individual characteristics in the cluster vary. In contrast, groups created in stratified sampling are homogeneous, as units share characteristics.

Relatedly, in cluster sampling you randomly select entire groups and include all units of each group in your sample. However, in stratified sampling, you select some units of all groups and include them in your sample. In this way, both methods can ensure that your sample is representative of the target population.

Is this article helpful?
Kassiani Nikolopoulou

Kassiani has an academic background in Communication, Bioeconomy and Circular Economy. As a former journalist she enjoys turning complex scientific information into easily accessible articles to help students.