Sampling Methods 101: Probability & Non-Probability Sampling Explained Simply

Summary

This video script explores the intricate world of sampling in research, distinguishing between probability and non-probability sampling methods. It emphasizes the importance of selecting a sample that aligns with the research aims, whether they are to generalize findings or to gain in-depth insights. Probability sampling, including simple random, stratified random, and cluster sampling, is ideal for quantitative research aiming for generalizable results. In contrast, non-probability sampling, such as purposive, convenience, and snowball sampling, is more suited for qualitative research seeking rich data. The video advises researchers to consider their resources and constraints when choosing a sampling method and to be transparent about the limitations of their approach.

Takeaways

• 📚 **Sampling Defined**: Sampling is the process of selecting a subset of participants from a larger group, known as the population, for research purposes.
• 🍰 **Population vs Sample**: The entire group you're studying is the 'population', and the part you engage with is the 'sample', likened to a slice of a cake.
• 🎯 **Representative Sample**: Ideally, a sample should be representative of the population to allow for generalization of findings.
• 🤔 **Practical Challenges**: Achieving a perfectly representative sample is challenging due to various practical obstacles.
• 📊 **Probability vs Non-Probability Sampling**: There are two main approaches to sampling - probability (random) and non-probability (based on researcher's discretion).
• 🔄 **Simple Random Sampling**: Involves random selection where each participant has an equal chance of being chosen, like drawing names from a hat.
• ⚖️ **Stratified Random Sampling**: Selects participants randomly from predefined subgroups (strata) to control for the impact of large subgroups.
• 🏢 **Cluster Sampling**: Samples from naturally occurring, mutually exclusive clusters within a population, useful for large geographic areas.
• 🎯 **Purposeful Sampling**: The researcher uses their judgment to select participants based on the study's aims, often used for rare or hard-to-find populations.
• 🚪 **Convenience Sampling**: Participants are selected based on their availability or accessibility to the researcher.
• ❄️ **Snowball Sampling**: Relies on referrals from initial participants to recruit more participants, useful for hard-to-reach populations.
• 🔍 **Choosing a Method**: The choice of sampling method should be guided by research aims, resources, and practical constraints.
• 📝 **Research Aims and Resources**: Consider whether your research aims are to produce generalizable findings or to develop in-depth insights when selecting a sampling method.
• ⚖️ **Trade-offs and Limitations**: Be prepared to make compromises in your sampling method due to practical limitations and ensure to articulate these clearly.

Q & A

What is the basic concept of sampling in research?

Sampling in research is the process of selecting a subset of participants from a larger group, known as the population. The goal is to obtain a sample that can represent the larger group for the purpose of the study.

Why is it impractical to collect data from every member of the population in a study?

Collecting data from every member of the population is impractical due to the vast number of individuals involved and the associated time, cost, and logistical challenges.

What are the two main approaches to sampling?

The two main approaches to sampling are probability sampling, where participants are selected on a statistically random basis, and non-probability sampling, where participant selection is based on the researcher's discretion.

How does simple random sampling differ from stratified random sampling?

Simple random sampling involves selecting participants randomly with each participant having an equal chance of being chosen. Stratified random sampling, on the other hand, involves selecting participants randomly but from within predefined subgroups (strata) that share a common trait.

What is cluster sampling, and how does it differ from stratified random sampling?

Cluster sampling involves selecting participants from naturally occurring, mutually exclusive clusters within a population. It differs from stratified random sampling in that it focuses on a subset of clusters rather than spreading the selection across the entire population.

Why might a researcher choose non-probability sampling over probability sampling?

A researcher might choose non-probability sampling when the richness and depth of the data are more important than generalizability, or when the study aims to develop deep insights into a specific subgroup rather than drawing conclusions about the broader population.

What are some limitations of convenience sampling?

Convenience sampling is quick and easy to implement but is likely to produce a non-representative sample and is vulnerable to research bias since it is based on the availability and accessibility of participants rather than a systematic process.

How does snowball sampling work, and what type of research is it best suited for?

Snowball sampling relies on referrals from initial participants to recruit additional participants. It is best suited for research involving hard-to-access populations or when the research topic is sensitive and trust is required for participants to engage.

What factors should a researcher consider when choosing a sampling method?

A researcher should consider their research aims, objectives, and questions, as well as available resources and practical constraints. The choice of sampling method should align with the research goals and be feasible within the given constraints.

Why is it important to clearly articulate the limitations of the chosen sampling method?

Articulating the limitations of the chosen sampling method is important because it provides transparency about the scope and applicability of the research findings. It also helps other researchers and practitioners understand the context and potential biases of the study.

What is the role of a sample in representing the larger population in a study?

The role of a sample is to act as a smaller, manageable part of the larger population that can be analyzed to infer insights about the entire group. An ideal sample would be representative of the population, allowing for generalization of findings.

How can a researcher ensure that their sampling method is aligned with their research aims?

A researcher can ensure alignment by carefully considering whether their research aims are more focused on generalizability or depth of insights, and by choosing a sampling method that best suits these aims. It's also important to consider the broader research methodology and practical constraints.

Outlines

00:00

📚 Introduction to Sampling in Research

This paragraph introduces the concept of sampling within a research context, emphasizing its importance for selecting a subset of participants from a larger group, known as the population. It outlines the goal of obtaining a representative sample to generalize findings. The video promises to cover various sampling methods and provide a foundational understanding to aid in making informed decisions for research projects. It also mentions the availability of free dissertation templates as a resource.

05:07

🎯 Types of Sampling: Probability vs. Non-Probability

The paragraph explains the two main approaches to sampling: probability sampling, which involves random selection, and non-probability sampling, which relies on the researcher's discretion. Probability sampling is often used in quantitative research aiming for generalizable results, while non-probability sampling is more common in qualitative research focused on in-depth data. The paragraph also lists three popular probability-based methods: simple random sampling, stratified random sampling, and cluster sampling.

10:10

🔍 Exploring Probability-Based Sampling Methods

This section delves into three common probability-based sampling methods: simple random sampling, which provides equal chances for each participant; stratified random sampling, which ensures representation from predefined subgroups; and cluster sampling, which selects participants from naturally occurring clusters. The paragraph discusses the advantages and limitations of each method and provides examples to illustrate how they might be used in practice.

15:15

🛠️ Non-Probability Sampling: Purpose, Convenience, and Snowball

The paragraph discusses three non-probability sampling methods: purposive sampling, which uses researcher judgment to select participants; convenience sampling, which chooses participants based on availability; and snowball sampling, which recruits participants through referrals. Each method is suited to specific research aims and circumstances, with a focus on qualitative research or when dealing with hard-to-reach populations. The paragraph also warns of potential biases associated with these methods.

🧐 Choosing the Right Sampling Method

The final paragraph advises on how to choose the appropriate sampling method for a study. It stresses the importance of aligning the sampling approach with the research aims and considering available resources and practical constraints. The paragraph acknowledges that compromises may be necessary and encourages researchers to clearly articulate their sampling choices and limitations. It concludes with a recap of the key points covered in the video and an invitation to engage with the Grad Coach channel for further research advice.

Keywords

💡Sampling

Sampling refers to the process of selecting a subset of participants from a larger group, known as the population, for the purpose of research. It is central to the video's theme as it underpins the discussion on how to choose the right method for a given study. The video script uses the analogy of a 'full cake' to illustrate the population and a 'single slice' to represent the sample.

💡Probability Sampling

Probability sampling, also known as random sampling, involves selecting participants based on a statistically random basis, ensuring each member of the population has an equal chance of being included. It is a key concept in the video as it is contrasted with non-probability sampling and is associated with quantitative research aiming for generalizable findings. Examples given in the script include simple random sampling, stratified random sampling, and cluster sampling.

💡Non-probability Sampling

Non-probability sampling encompasses methods where participant selection is not statistically random, relying on the researcher's discretion. This approach is highlighted in the video as being more common in qualitative research, where the depth of data is prioritized over generalizability. The script discusses purposive sampling, convenience sampling, and snowball sampling as examples.

💡Stratified Random Sampling

Stratified random sampling is a probability-based method where the population is divided into subgroups (strata) that share a common trait, and participants are then randomly selected from each stratum. The video explains that this method allows for more control over the representation of different subgroups within the sample, which is crucial for identifying differences between these subgroups.

💡Cluster Sampling

Cluster sampling involves selecting samples from naturally occurring, mutually exclusive groups within the population, known as clusters. After defining these clusters, a random selection of clusters is made, and then participants are randomly chosen from each selected cluster. The video contrasts this with stratified sampling, noting that it is more practical and economical, especially for populations spread over a large geographic area.

💡Purposive Sampling

Purposive sampling, also referred to as judgment or subjective sampling, is a non-probability method where the researcher selects participants based on their judgment related to the study's purpose. This method is highlighted in the video as being particularly useful for studies aiming to gather information from small, rare, or hard-to-find populations. However, it is also noted to be prone to researcher bias.

💡Convenience Sampling

Convenience sampling is a non-probability method where participants are selected based on their availability or accessibility to the researcher. The video script mentions that this method is quick and easy but is unlikely to produce a representative sample and is susceptible to research bias, making it a choice that must be approached with caution.

💡Snowball Sampling

Snowball sampling is a non-probability technique that relies on referrals from initial participants to recruit additional subjects. It is depicted in the video as ideal for research involving hard-to-access populations or sensitive topics, as it leverages trust within a community to gather data. However, it is also noted to be highly prone to researcher bias.

💡Research Aims

Research aims are the specific objectives or goals of a study that guide the choice of methodology, including the sampling method. The video emphasizes the importance of aligning the sampling approach with the research aims, whether the focus is on generalizable findings or in-depth insights. This alignment is crucial for selecting the appropriate sampling method.

💡Generalizability

Generalizability refers to the ability to apply the findings of a study to a larger population. The video discusses how probability-based sampling methods are more likely to produce generalizable results, which is essential for studies aiming to draw conclusions about the broader population rather than just the sample.

💡Research Bias

Research bias is a systematic error that can affect the results of a study, often due to the way participants are selected. The video script addresses the susceptibility of non-probability sampling methods, such as convenience and snowball sampling, to research bias, and the importance of being aware of and addressing these biases.

Highlights

Explaining the concept of sampling within a research context, which is selecting a subset of participants from a larger group.

Describing the population as the larger group and the sample as a subset engaged with for research purposes.

Discussing the ideal of a perfectly representative sample and its practical challenges.

Differentiating between probability sampling, which is random, and non-probability sampling, based on researcher discretion.

Highlighting the use of probability sampling in quantitative research aiming for generalizable findings.

Exploring non-probability sampling's role in qualitative research, focusing on data richness over generalizability.

Introducing simple random sampling as a method where each participant has an equal chance of selection.

Discussing stratified random sampling, which involves random selection from predefined subgroups or strata.

Describing cluster sampling, which samples from naturally occurring, mutually exclusive clusters within a population.

Presenting purposive sampling, where participants are selected based on the researcher's judgment aligned with study aims.

Explaining convenience sampling, which is based on the availability and accessibility of participants.

Introducing snowball sampling, which recruits additional participants through referrals from initial subjects.

Guiding on choosing the right sampling method by considering research aims and available resources.

Emphasizing the importance of aligning the sampling approach with the research aims and methodology.

Discussing the trade-offs involved in sampling methods and the necessity of being ready to compromise.

Stressing the importance of clearly articulating the limitations of the chosen sampling approach.

Providing actionable advice on how to plan research with the right sampling method in mind.

Encouraging transparency about study limitations to maintain the integrity of the research process.

Transcripts

In this video, we are going to unpack the  jargon-filled world of sampling and sampling  

methods. We will explain what sampling is explore  the most popular sampling methods and unpack how  

to choose the right sampling method for your  study. By the end of this video, you will have  

a clear foundational understanding of sampling  so that you can make informed decisions for your  

research project. By the way, if you are currently  working on a dissertation, thesis or research  

project be sure to grab our free dissertation  templates to help fast-track your write-up.  

These tried and tested templates provide a  detailed roadmap to guide you through each  

chapter section by section. If that sounds helpful  you can find the link in the description below.

To kick things off let us start with the what. So  what exactly does sampling mean within a research  

context? Well at the simplest level sampling is  the process of selecting a subset of participants  

from a larger group. For example, if your research  aimed to assess US consumers’ perceptions about  

a particular brand of laundry detergent you  would not be able to collect data from every  

single person that uses laundry detergent yeah  good luck with that. But you could plausibly  

collect data from a smaller subset of this group.  In technical terms, the larger group is referred  

to as the population and the subset the group you  will actually engage with is called the sample.  

Put another way you can look at the population as  a full cake and the sample as a single slice of  

that cake. Can you see what my mind is on? Well  in an ideal world, you would want your sample to  

be perfectly representative of the population as  that would allow you to generalise your findings  

to the entire population. In other words, you  would want to cut a perfect cross-sectional  

slice of that cake such that the slice reflects  every layer of the cake in exact proportion.  

Unfortunately achieving a truly representative  sample is a little trickier than slicing a cake  

as there are many practical challenges and  obstacles to achieving this in a real-world  

setting. Thankfully though you do not always  need to have a perfectly representative sample.  

It largely depends on the specific research  aims of each study so do not stress yourself  

out about that just yet. By the way, if you want  to learn more about research aims, objectives  

and questions check out our explainer video up  here or click on the link in the description.  

All right with the concept of sampling broadly  defined let us look at the different approaches  

to sampling to get a better understanding  of what it all looks like in practice.

At the highest level, there are two approaches to  sampling. Probability sampling and non-probability  

sampling. Within each of these, there are  a variety of sampling methods which we will  

explore a little later. Probability sampling  involves selecting participants or any other  

unit of interest on a statistically random basis  which is why it is also called random sampling.  

In other words, the selection of each individual  participant is based on a predetermined process  

and not the discretion of the researcher.  As a result of this process-driven approach  

probability sampling achieves a random sample.  Probability-based sampling methods are most  

commonly used in quantitative research in other  words numbers and statistics-based research.  

This is especially true for studies where the  aim is to produce generalisable findings. In  

other words to produce findings that allow you  to draw conclusions about the broader population  

of interest, not just the sample itself. Right now  let us look at the second approach non-probability  

sampling. Non-probability sampling as the name  suggests consists of sampling methods in which  

participant selection is not statistically  random. In other words, the selection of  

individual participants is based on the discretion  and judgment of the researcher rather than on a  

predetermined process. Non-probability sampling  methods are commonly used in qualitative research  

where the richness and depth of the data are  more important than the generalisability of the  

findings. If that all sounds a little too  conceptual and fluffy do not worry. Next,  

we will take a look at some actual sampling  methods to make it a little more tangible.

To kick things off we will look at three popular  probability-based random samplings. Specifically,  

we will explore simple random sampling, stratified  random sampling and cluster sampling. Importantly  

this is not a comprehensive list of all possible  probability sampling methods these are just the  

three most common ones. So if you are interested  in adopting a probability-based sampling approach  

be sure to explore all of the options. First up  we have got simple random sampling. Simple random  

sampling involves selecting participants  in a completely random fashion where each  

participant has an equal chance of being  selected. Basically, this sampling method  

is the equivalent of pulling names out of a hat  except that you can do it digitally. For example,  

if you had a list of five hundred people you  could use a random number generator to draw a  

list of fifty numbers each number reflecting  a participant and then use that data set as  

your sample. Thanks to its simplicity simple  random sampling are easy to implement and as  

a consequence is typically quite cheap and  efficient because the selection process is  

completely random you can generalise  your results fairly reliably. However,  

this also means it can hide the impact of large  subgroups within the data which can result in  

minority subgroups having little representation  in the results if any at all. This may or may  

not be an issue depending on what you are trying  to achieve. So it is important to always consider  

your research aims and research questions when you  are deciding which sampling method to use. We will  

explore that a little later in more detail. Next  in line, we have got stratified random sampling.  

Stratified random sampling is similar to the  previous method but it kicks things up a notch.  

As the name suggests stratified sampling involves  selecting participants randomly but from within  

certain predefined subgroups that share a common  trait these are called strata. For example, you  

could stratify a given population based on gender  ethnicity or level of education and then select  

participants randomly from each group. The benefit  of this sampling method is that it gives you  

more control over the impact of large subgroups,  large strata within the population. For example,  

if a population comprises eighty percent males  and twenty females you may want to balance this  

skew out by selecting an equal number of male  and female participants. This would of course  

reduce the representativeness of the sample but  it would allow you to identify differences between  

subgroups. So once again you need to think about  your research aims as well as the nature of the  

population that you are interested in so that you  can make the right sampling choice. Next up let us  

look at cluster sampling. As the name suggests  this sampling method involves sampling from  

naturally occurring mutually exclusive clusters  within a population. For example, area codes  

within a city or cities within a country once  the clusters are defined a set of clusters is  

randomly selected and then a set of participants  are randomly selected from each cluster. Now you  

are probably wondering how is cluster sampling  different from stratified random sampling. Well,  

let us look at the previous example where each  cluster reflects an area code in a given city.  

With cluster sampling, you would collect data  from clusters of participants in a handful of area  

codes let us say five neighbourhoods. Conversely,  with stratified random sampling, you would need to  

collect data from all over the city. In other  words many more neighbourhoods. Either way,  

you would still achieve the same sample size  let us say two hundred people, for example. But  

with stratified sampling, you do not need to do a  lot more running around as participants would be  

scattered across a more vast, geographic area.  As a result cluster sampling is often the more  

practical and economical option especially when  the population spans a large geographic area. If  

that all sounds a little mind-bending you can  use the following general rule of thumb. If a  

population is relatively homogeneous cluster  sampling will often do the trick. Conversely,  

if a population is generally quite heterogeneous  in other words diverse stratified sampling will  

often be more appropriate. If you are enjoying  this video so far please help us out by hitting  

that like button. You can also subscribe for  loads of plain language, actionable advice. If  

you are new to research check out our free  dissertation writing course which covers  

everything that you need to get started on your  research project. As always links are down in the  

description. Right now that we have looked at a  few probability-based sampling methods it is time  

to explore some non-probability methods. The three  that we will be looking at are purposive sampling,  

convenience sampling and snowball sampling. First  up we have got purpose of sampling also known  

as judgment or subjective sampling. The names  here provide some clues as the sampling method  

involves the researcher selecting participants  using their own judgment based on the purpose  

of the study that is to say the research aims.  For example, suppose your research aims were  

to understand the perceptions of hyper-loyal  customers of a particular retail store in that  

case you could use your judgment to engage with  frequent shoppers as well as rare or occasional  

shoppers and then analyse the resultant data  to understand what perceptions and attitudes  

drive the two behavioural extremes frequent  versus rare shopping. Purpose of sampling is  

often used in studies where the aim is to gather  information from a small population especially  

rare or hard-to-find populations as it allows  the researcher to target specific individuals who  

have unique knowledge or experience. Naturally,  this sampling method is quite prone to research  

bias and judgment error and it is unlikely to  produce generalisable results. So it is best  

suited to studies where the aim is to go narrow  and deep rather than broad. Next up let us look  

at convenience sampling. As the name suggests  with this method participants are selected  

based on their availability or accessibility. In  other words, the sample is selected based on how  

convenient it is for the researcher to access it  as opposed to using a predefined or objective or  

consistent process. Naturally, convenience  sampling provides a quick and easy way to  

gather data as the sample is selected based on  the individuals who are readily available and  

willing. This makes it an attractive option if you  are particularly tight on resources or time but as  

you would expect this sampling method is unlikely  to produce a representative sample and will of  

course be vulnerable to research bias. So it is  important to approach it with caution. By the way,  

if you want to learn more about research bias  check out our video about that up here or you  

can hit the link in the description. Last  but certainly not least we have the snowball  

sampling method. This method relies on referrals  from initial participants to recruit additional  

participants. In other words, the initial subjects  from the first small snowball and each additional  

subject recruited through a referral are added to  the snowball making it larger as it rolls along.  

Snowball sampling is often used in situations  where it is difficult to identify and access a  

particular population. For example, people with a  rare medical condition or members of an exclusive  

group. It can also be useful in cases where  the research topic is sensitive or taboo and  

people are unlikely to open up unless they are  referred to by someone that they trust. Simply  

put snowball sampling is ideal for research that  involves reaching hard-to-access populations but  

keep in mind that once again it is a sampling  method that is highly prone to researcher bias  

and is unlikely to produce a representative  sample. So make sure that it aligns with your  

research aims and research questions before  you start rolling your snowball down the hill.

Now that we have looked at a few popular sampling  methods both probability and non-probability based  

the obvious question is how do I choose the right  method of sampling for my study? This is a big  

question and we could do an entire video covering  this but I will try to cover the essentials here.  

When selecting a sampling method for your research  project you will need to consider two important  

factors your research aims and your resources. As  with all research design and methodology choices,  

your sampling approach needs to be guided by  and aligned with your research aims objectives  

and research questions in other words your  golden thread. I know I am starting to sound  

like a stuck record here but this alignment  is really important specifically you need to  

consider whether your research aims are primarily  concerned with producing generalisable findings  

in which case you will likely opt for a  probability-based sampling method or if  

they are more focused on developing rich deep  insights in which case a non-probability based  

approach could be more practical. Typically  quantitative studies lean towards the former  

while qualitative studies lean towards the  latter. So be sure to consider your broader  

methodology as well. The second factor you need  to consider is your resources and more generally  

the practical constraints at play. For example,  if you have easy free access to a large sample at  

your workplace or university along with a healthy  budget to help attract your participants that will  

open up multiple options in terms of sampling  methods. Conversely, if you are cash-strapped,  

short on time and do not have unfettered  access to your population of interest you may  

be restricted to convenience or referral-based  methods. Importantly you need to be ready for  

trade-offs. You will not always be able to  utilise the perfect sampling method for your  

study and that is okay. Much like all the other  methodological choices you will make as part of  

your study, you will often need to compromise and  accept trade-offs when it comes to sampling. Do  

not let this get you down though as long as your  sampling choices are well explained and justified  

and the limitations of your approach are clearly  articulated you will be on the right track. Every  

study has its limitations so do not try to hide  yours. By the way, if you want to learn more  

about research limitations we have got videos  covering that too. Links in the description.

All right we have covered a lot of ground  in this video. Let us quickly recap the key  

takeaways. Sampling is the process of defining  a subgroup a sample from the larger group of  

interest in the population. The two overarching  approaches to sampling are probability sampling,  

random sampling and non-probability sampling.  Popular probability-based sampling methods  

include simple random sampling, stratified  random sampling and cluster sampling. Popular  

non-probability-based sampling methods include  purpose of sampling, convenience sampling and  

snowball sampling. When choosing a sampling method  you need to take into account your research aims,  

objectives and question as well as your resources  and other practical constraints. Keep these points  

in mind as you plan your research and you  will be on the path to sampling success.

If you got value from this video please hit  that like button to help more students find  

this content for more videos like this check  out the Grad Coach channel and subscribe for  

plain language, actionable research tips and  advice every week. Also if you are looking for  

one-on-one support with your dissertation,  thesis or research project be sure to check  

out our private coaching service where we hold  your hand throughout the research process step  

by step. You can learn more about that and book  a free initial consultation at gradcoach.com.