# Understanding Non-Probability Sampling – A Brief Guide

*by James Whaley Master Marketer and Business Builder*

When it comes to market research, probability sampling is a methodology you often see, especially in quantitative research. On the other end of the spectrum is non-probability sampling, which by the names of it lacks the element of *randomized selection *probability sampling is known for having. It’s also true that, compared to probability sampling, non-probability sampling is considered a little unconventional in market research.

The reason behind it is quite simple. Non-probability samples don’t necessarily have the same rationale that probability samples typically boast. With the latter, we know that market researchers can gauge the probability of how effectively the sample represents the population. You can calculate – or use a statistical analysis platform to gauge – the sample's confidence intervals.

Non-probability samples, on the other hand, might not be *as effective *in representing the population. Researchers prefer probabilistic sampling methods as they are more certain of the resulting sample being more accurate and rigorous.

However, that does not mean that it is completely ineffective in representing the population being considered. The fact is that probabilistic sampling in market research is more widely applicable. It suits more research settings, which certainly doesn’t throw non-probabilistic sampling out the park.

There are opportunities in social research that may benefit from non-probability sampling, both theoretically and practically. For those instances when non-probability sampling seems more feasible, it’s important to have a basic understanding of what it is, its types, and when it can be used.

# Non-Probability Sampling

## What is it?

As mentioned before, non-probability sampling is a sampling technique that doesn’t have the randomizing effect of probability sampling. Instead, samples are chosen based on their own subjectivity – it’s not random. Members of the population being considered already know their chances of being selected, unlike probability sampling, where all participants have a somewhat equal chance.

Fewer rules are governing the how-to of non-probability sampling. However, it does require researchers with expansive research expertise to use their honed observational skills and proficient judgment to choose the sample. This is why non-probabilistic sampling is used more often for qualitative research, specifically exploratory studies and pilot surveys.

## What are the types?

Generally speaking, there are five different types of non-probability samples. These can be divided into two main categories, accidental and purposive. Most types of non-probability sampling fall under the latter of the two categories. When researchers use this type of sampling, they approach it with a precise plan in mind.

Let’s get familiar with the five main types of non-probability sampling:

### 1. Convenience sampling

This type of non-probability sampling is one of the most commonly used ones. There are various other names convenience sampling is known by, including accidental and haphazard sampling.

The reason being that researchers choose their sample based on their accessibility. People are easy to recruit, and hence, researchers choose them to answer the survey, questionnaire, etc., not really considering whether their choices are representative of the entire population.

While it’s beneficial to a research’s accuracy that the sample represents all aspects of a population, there are some areas where the population is too large for the sample to be truly representative. Plus, it’s fast, cost-effective, and easily accessible.

One of the most common examples of convenience sampling is the classic street interview model many media outlets are now using. You simply walk up to people on the street because they are near and willing to answer your questions.

Similarly, some clinical tests also employ the convenience sampling method. This is mainly because clinical tests are high-risk, and the majority of a research’s population might not be willing to participate. So, they recruit those who are ready to volunteer.

### 2. Consecutive sampling

While it’s considered a distinct type of non-probability sampling, it is more of a convenience sampling sub-type. We say this because the basic principle of choosing the sample from a population is the same as convenience sampling.

The slight difference is that instead of letting go of the sample after participating in the survey, you pick that sample (it can be an individual or a group) and conduct research with them as participants for a period. This is mainly effective for research that involves many variables and/or multiple topics to cover.

Researchers conduct their study, collect and analyze the results, and then move to another sample group if needed.

### 3. Purposive Sampling

As the name suggests, this type of non-probability sampling method is approached with a specific purpose. Researchers build consumer personas with predefined traits they’re seeking and then go out looking for them. They use their judgment to deem whether the prospect is suitable or not. This is why purposive sampling is also called judgment sampling.

You might have seen this type of sampling in practice in shopping malls, where people holding clipboards stop a particular category of people and ask whether they want to participate in the research or not.

What makes this type of non-probability sampling useful is that it helps researchers find their sample quickly while meeting certain criteria. Unlike convenience sampling, this type of non-probability sampling gives researchers a bit more control over what their sample looks like.

However, there is a downside to purposive sampling. There is a chance that the researcher’s preconceived ideas about the population may interfere with the results. There are some ambiguities and other biases that can creep in and potentially skew the inference.

### 4. Quota Sampling

Fourth on the list of non-probability sampling types is quota sampling, which is quite similar to stratified sampling. It’s all about choosing fixed quotas non-randomly. To better understand this type of non-probability sampling, let’s describe its two sub-types:

**Proportional Quota Sampling** is a technique where your goal is to choose a quota representing the population proportionally, based on a major characteristic. For example, consider your target population is an urban college of 1000 students, with 45% identifying as women, 50% identifying as men, and the remaining 5% as non-binary or gender fluid.

Now let’s consider that you want a sample consisting of 200 people. If you’re using proportional quota sampling, you need 90 female-identifying, 100 male-identifying, and 10 non-binary/gender fluid-identifying students.

Gender identity is only an example of characteristics you can choose to set the quota on.

**Non-Proportional Quota Sampling** is the opposite of its counterpart. It is a bit less restrictive in terms of representation in the quota. You have to specify the minimum number of members of your quota sample within a category. Matching the proportionality of the population is not a priority here.

The goal is to have enough numbers to give some semblance of representation of the population. This type of non-probability sampling is minutely close to stratified sampling that ensures adequate population representation in smaller groups.

### 5. Expert Sampling

Expert sampling can be considered a sub-type of purposive sampling. The definition is quite easy to understand. You are essentially looking for a sample of people who have demonstrated or proven expertise in a certain area.

This type of purposive sampling works for market research on niche products that require the experience and expert opinion of industry professionals. This panel of experts allows you to defend your research and make better decisions.

### 6. Snowball sampling

Last but certainly not least is snowball sampling. You begin with a small selection of prospects, using referrals to snowball the sample into a larger size. This non-probability sampling approach is suitable for instances where participants are hard to find due to the small sample size.

As mentioned, this type of sampling can be considered as a referral program. You find one or a few participants that check all prerequisite boxes and then asking them to refer other participants that meet the same requirements.

# Closing Thoughts

While non-probability sampling might not be as actively used as other types of sampling methods in market research, it has some perks. For instance, they can be more favorable for online surveys that collect qualitative data. Plus, this methodology of sampling is cost-effective and faster to carry out. Researchers know what they are looking for in their sample, often being familiar with it.

You this simple tool to estimate your sample size.

Choosing the right type of non-probability sampling depends on your research objective, budget, and population accessibility.

#### Sponsor Ads

Created on Nov 27th 2020 13:05. Viewed 704 times.