Sampling Techniques: Probability vs. Non-Probability Sampling Explained
Every researcher eventually faces a daunting reality: you cannot survey every single person in your target population. Whether you’re studying small business owners in Nairobi, university students in Kisumu, or farmers in rural Kiambu, the population is simply too large, too dispersed, or too expensive to reach entirely.
So what do you do? You sample.
Sampling is the process of selecting a subset of your population to represent the whole. Choose your sample well, and you can draw accurate conclusions about thousands of people from just a few hundred respondents. Choose poorly, and your entire research could be flawed from the start.
This article will explain the two main families of sampling techniques—probability sampling and non-probability sampling—and help you decide which one is right for your thesis. By the end, you’ll understand the difference between random and purposive selection, when to use each method, and how to justify your choice to your supervisor.
If you find the research methodology chapter overwhelming, you’re not alone. Many students struggle to select the right sampling technique for their study. At Proposal Writers Kenya, we help students craft clear, well-justified methodology sections that impress supervisors. But first, let’s master the fundamentals of sampling.
What Is Sampling and Why Do You Need It?
Sampling is the process of selecting a smaller group (your sample) from a larger group (your population) to participate in your research. The goal is to learn about the population by studying the sample.
Consider this example: You want to understand the study habits of all 10,000 students at your university. Interviewing every single student would take months and cost a fortune. Instead, you select 300 students who represent the larger population. If you choose them correctly, their responses will closely mirror what the entire 10,000 would say.
The key concepts you need to understand are:
Target population: The entire group you want to study (e.g., all registered nurses in Kenya)
Accessible population: The portion you can realistically reach (e.g., nurses working in Nairobi County hospitals)
Sample: The individuals you actually collect data from
Sample size: The number of people in your sample
The ultimate goal of sampling is generalizability—the ability to take your findings from your sample and apply them confidently to your entire population.
The Two Main Families of Sampling Techniques
All sampling techniques fall into one of two families: probability sampling or non-probability sampling. The distinction comes down to one critical factor: random selection.
| Feature | Probability Sampling | Non-Probability Sampling |
|---|---|---|
| Random selection | Yes | No |
| Every member has a known chance | Yes | No |
| Generalizable results | Yes | Limited |
| Common in | Quantitative research | Qualitative research |
| Risk of bias | Low | Higher |
Probability sampling is the gold standard when you need results that represent the entire population. Non-probability sampling is more practical and common in qualitative research where depth matters more than breadth.
Probability Sampling Explained
Probability sampling means every member of your population has a known, non-zero chance of being selected. This randomness is what allows you to use statistics and generalize your findings.
Simple Random Sampling
This is the purest form of probability sampling. Every member of the population has an equal chance of being selected—like a lottery.
How it works: You obtain a complete list of your population (called a sampling frame), then use a random number generator or the old-fashioned fishbowl method to select your sample.
Example in Kenyan context: You have a list of 2,000 registered voters in a Nairobi ward. You assign each voter a number, then use a random number generator to select 200 participants.
Advantages: Completely unbiased, easy to understand.
Disadvantages: You need a complete, accurate list of every population member.
Stratified Random Sampling
This technique ensures specific subgroups are represented proportionally.
How it works: You divide your population into subgroups (strata) based on a characteristic (e.g., gender, year of study, county), then randomly sample from each subgroup.
Example in Kenyan context: A study on academic performance at Kenyatta University. You divide students by year of study (1st, 2nd, 3rd, 4th year), then randomly select 70 students from each year to ensure all levels are represented.
Advantages: Ensures minority groups aren’t excluded.
Disadvantages: You need to know the characteristics of your population in advance.
Cluster Sampling
This technique is ideal for geographically dispersed populations.
How it works: You divide the population into natural clusters (e.g., schools, villages, hospitals), randomly select clusters, then sample everyone within those clusters.
Example in Kenyan context: Studying primary school students across Kenya. Instead of listing every student, you randomly select 20 schools (clusters) and survey all students in those schools.
Advantages: Cost-effective and practical for large geographical areas.
Disadvantages: Higher sampling error than simple random sampling.
Systematic Sampling
This is a simpler alternative to simple random sampling.
How it works: You select every kth member from a list. To find k, divide your population size by your desired sample size.
Example in Kenyan context: You have a list of 1,000 patients at a Kisumu hospital and need 100 participants. You select every 10th patient on the list (1,000 ÷ 100 = 10).
Advantages: Simple to implement.
Disadvantages: Hidden patterns in the list can introduce bias.
When to Use Probability Sampling
Use probability sampling when:
You need results that generalize to the entire population
You are conducting quantitative research (surveys, experiments)
You have access to a complete sampling frame
Your supervisor expects statistical generalizability
Non-Probability Sampling Explained
Non-probability sampling means not every population member has a chance to be selected. The researcher deliberately chooses participants based on specific criteria, availability, or convenience.
While you cannot calculate sampling error or generalize with confidence, these techniques are valuable for qualitative research and hard-to-reach populations.
Purposive Sampling
The researcher deliberately selects participants who have specific characteristics or knowledge relevant to the study.
How it works: You define criteria for inclusion, then intentionally find people who meet those criteria.
Example in Kenyan context: Studying successful women entrepreneurs in Mombasa. You deliberately select women who have operated a business for over five years and employ at least ten people.
Advantages: You get exactly the right participants for your research question.
Disadvantages: Researcher bias can influence selection.
Snowball Sampling
Existing participants recruit future participants from their networks.
How it works: You find one or two initial participants, ask them to refer others, and continue until you reach your desired sample size.
Example in Kenyan context: Studying street-connected youth in Nakuru. You find one participant through a local NGO, who introduces you to others, who introduce you to more.
Advantages: Accesses hidden or hard-to-reach populations.
Disadvantages: Sample may be biased by social networks.
Convenience Sampling
You select whoever is easiest to reach.
How it works: You approach available people until you have enough responses.
Example in Kenyan context: Standing outside a lecture hall and surveying the first 50 students who walk out.
Advantages: Fast, cheap, and easy.
Disadvantages: Highest risk of bias; least generalizable.
Quota Sampling
You set quotas for subgroups, then non-randomly fill them.
How it works: You decide you need 50 male and 50 female respondents. Then you stand in a shopping mall and approach people until you have 50 of each gender.
Example in Kenyan context: A market research study needing equal representation of high-income and low-income shoppers in Nairobi.
Advantages: Ensures subgroup representation without a sampling frame.
Disadvantages: Selection within quotas is still biased.
When to Use Non-Probability Sampling
Use non-probability sampling when:
You are conducting qualitative research (interviews, focus groups)
Your population is hidden or hard to reach
You are doing exploratory or pilot research
Depth of understanding is more important than generalizability
How to Choose the Right Sampling Technique for Your Study
Follow these five steps:
Identify your research approach. Quantitative studies typically need probability sampling. Qualitative studies typically use non-probability sampling.
Check if you have a sampling frame. Do you have a complete list of your population? If yes, probability sampling is possible. If no, you likely need non-probability sampling.
Consider your budget and timeline. Probability sampling takes more time and resources.
Assess your generalizability needs. Do you need to make claims about the entire population? Use probability sampling.
Evaluate your population’s accessibility. Hidden populations require snowball or purposive sampling.
Examples by Discipline in Kenya
| Discipline | Common Sampling Technique | Justification |
|---|---|---|
| Education (quantitative) | Stratified random | Ensure representation across schools or grade levels |
| Public Health | Cluster sampling | Study households across different counties |
| Business (qualitative) | Purposive | Interview successful entrepreneurs |
| Sociology (hidden population) | Snowball | Access hard-to-reach groups |
| Psychology (pilot study) | Convenience | Test instruments with available students |
Common Mistakes Kenyan Students Make With Sampling
Avoid these frequent errors in your thesis proposal:
Claiming random sampling without using random selection. Saying “random” doesn’t make it random. You must describe your random selection method.
Using convenience sampling but claiming generalizable results. Be honest about your limitations.
No justification for your sampling technique. Every sampling choice needs a rationale.
Sample size too small for statistical tests. Consult a statistician or use sample size formulas.
Ignoring sampling bias in your limitations section. Every study has limitations; acknowledge them.
Confusing quota sampling with stratified sampling. The key difference: quota sampling is non-random; stratified random sampling is random.
Conclusion
Choosing the right sampling technique is one of the most important decisions you will make in your research methodology. Probability sampling—including simple random, stratified random, cluster, and systematic sampling—is your best choice when you need generalizable, statistically valid results. Non-probability sampling—including purposive, snowball, convenience, and quota sampling—is more practical for qualitative research, exploratory studies, and hard-to-reach populations.
The most important rule is this: whatever you choose, justify it. Your supervisor needs to understand why your sampling technique aligns with your research objectives, population, and resources.
If you’re feeling stuck on your research methodology chapter—whether it’s sampling, sample size calculation, or data analysis—you don’t have to figure it out alone. At Proposal Writers Kenya, we help students across Kenya craft clear, well-justified thesis proposals that impress supervisors and get approved. From Chapter One to Chapter Three, we provide expert guidance tailored to your specific study.
Ready to move forward with your thesis proposal? Contact us today for a free quote and let’s get your research on the right track.