Sampling Methods in Research: Types, Techniques, and How to Choose

A sampling method is the procedure a researcher uses to select the people, objects, or records that will be studied from a larger group. In most studies, collecting data from an entire population is impractical, so researchers study a sample and draw conclusions about the population from it. How the sample is chosen determines how far those conclusions can be trusted, which is why the choice of sampling method is one of the most important decisions in research design.

This article covers the following:

• What a sample and a population are, and why sampling methods matter
• The two broad families of sampling techniques: probability sampling and non-probability sampling
• The main sampling techniques in research, with examples of each
• How sampling differs in qualitative and quantitative research
• How to choose a sampling method, and the common mistakes to avoid

Population and Sample

The population is the entire group a study is interested in, such as all adult patients with a particular condition in a country. The sample is the subset of that population from which data are actually collected. A good sample represents the population, so that findings from the sample apply to the whole group.

The difference between the two also determines what a study can claim. Results from a sample that was selected by chance can be generalized to the population with a measurable degree of confidence. Results from a sample chosen for convenience or judgment are more limited, and the manuscript should say so. How a study defines and bounds its population is closely tied to its scope, which Scope and Delimitations in Research discusses.

Probability Sampling and Non-Probability Sampling

The types of sampling used in research fall into two families. In probability sampling, every member of the population has a known, nonzero chance of being selected, and selection is random. This allows statistical estimates of how well the sample represents the population. In non-probability sampling, selection depends on the researcher’s judgment or on availability, so the chance of selection is unknown. Non-probability sampling is common when no complete list of the population exists, when the population is hard to reach, or when the aim is depth of understanding and not statistical generalization.

Probability Sampling Techniques

Simple random sampling. Every member of the population has an equal chance of being selected, for example by using random numbers to pick participants from a complete list. It is the most straightforward technique and requires a full list of the population, called a sampling frame.

Stratified sampling. The population is divided into subgroups, or strata, such as age groups or regions, and a random sample is drawn from each. This ensures that each subgroup is represented and is useful when the groups are expected to differ on the outcome of interest.

Cluster sampling. The population is divided into naturally occurring groups, such as schools or hospitals, and a random selection of whole clusters is studied. It is practical when the population is spread over a wide area and a full list of individuals is not available.

Systematic sampling. Participants are selected at a regular interval from an ordered list, for example every tenth person. It is simple to carry out, but it can introduce bias if the list has a hidden pattern that matches the interval.

Non-Probability Sampling Techniques

Convenience sampling. Participants are chosen because they are easy to reach, such as students in the researcher’s own class. It is quick and inexpensive, but the sample may not represent the population, and the results cannot be generalized with confidence.

Purposive sampling. The researcher deliberately selects participants who have the characteristics or experience that the study needs. Also called judgment sampling, it is widely used in qualitative research, where the aim is to learn from people who can speak directly to the topic.

Snowball sampling. Existing participants recruit further participants from among their contacts. It is useful for hard-to-reach groups, such as people with a rare condition, but it tends to produce samples of people who know one another, which limits diversity.

Quota sampling. The researcher sets quotas for certain characteristics, such as sex or age, so that the sample matches the population’s proportions, and then fills the quotas by non-random means. It resembles stratified sampling, but the final selection is not random, which separates the two.

Sampling Methods at a Glance

<table style="width:100%; border-collapse:collapse;">
  <thead>
    <tr>
      <th style="border:1px solid #ddd; padding:10px; text-align:left;">Technique</th>
      <th style="border:1px solid #ddd; padding:10px; text-align:left;">Family</th>
      <th style="border:1px solid #ddd; padding:10px; text-align:left;">Selection</th>
      <th style="border:1px solid #ddd; padding:10px; text-align:left;">Typical Use</th>
      <th style="border:1px solid #ddd; padding:10px; text-align:left;">Main Limitation</th>
    </tr>
  </thead>
  <tbody>
    <tr>
      <td style="border:1px solid #ddd; padding:10px;">Simple Random</td>
      <td style="border:1px solid #ddd; padding:10px;">Probability</td>
      <td style="border:1px solid #ddd; padding:10px;">Random selection from a complete list</td>
      <td style="border:1px solid #ddd; padding:10px;">Surveys with an available sampling frame</td>
      <td style="border:1px solid #ddd; padding:10px;">Needs a full list of the population</td>
    </tr>
    <tr>
      <td style="border:1px solid #ddd; padding:10px;">Stratified</td>
      <td style="border:1px solid #ddd; padding:10px;">Probability</td>
      <td style="border:1px solid #ddd; padding:10px;">Random selection within subgroups</td>
      <td style="border:1px solid #ddd; padding:10px;">Studies where subgroups differ</td>
      <td style="border:1px solid #ddd; padding:10px;">Requires information about the subgroups</td>
    </tr>
    <tr>
      <td style="border:1px solid #ddd; padding:10px;">Cluster</td>
      <td style="border:1px solid #ddd; padding:10px;">Probability</td>
      <td style="border:1px solid #ddd; padding:10px;">Random selection of groups</td>
      <td style="border:1px solid #ddd; padding:10px;">Widely spread populations</td>
      <td style="border:1px solid #ddd; padding:10px;">Higher sampling error than individual random sampling</td>
    </tr>
    <tr>
      <td style="border:1px solid #ddd; padding:10px;">Systematic</td>
      <td style="border:1px solid #ddd; padding:10px;">Probability</td>
      <td style="border:1px solid #ddd; padding:10px;">Every nth member of a list</td>
      <td style="border:1px solid #ddd; padding:10px;">Large ordered lists</td>
      <td style="border:1px solid #ddd; padding:10px;">Hidden patterns can bias the sample</td>
    </tr>
    <tr>
      <td style="border:1px solid #ddd; padding:10px;">Convenience</td>
      <td style="border:1px solid #ddd; padding:10px;">Non-probability</td>
      <td style="border:1px solid #ddd; padding:10px;">Easily available participants</td>
      <td style="border:1px solid #ddd; padding:10px;">Pilot studies, exploratory work</td>
      <td style="border:1px solid #ddd; padding:10px;">Weak representativeness</td>
    </tr>
    <tr>
      <td style="border:1px solid #ddd; padding:10px;">Purposive</td>
      <td style="border:1px solid #ddd; padding:10px;">Non-probability</td>
      <td style="border:1px solid #ddd; padding:10px;">Chosen for specific characteristics</td>
      <td style="border:1px solid #ddd; padding:10px;">Qualitative research</td>
      <td style="border:1px solid #ddd; padding:10px;">Depends on researcher judgment</td>
    </tr>
    <tr>
      <td style="border:1px solid #ddd; padding:10px;">Snowball</td>
      <td style="border:1px solid #ddd; padding:10px;">Non-probability</td>
      <td style="border:1px solid #ddd; padding:10px;">Participants refer others</td>
      <td style="border:1px solid #ddd; padding:10px;">Hard-to-reach groups</td>
      <td style="border:1px solid #ddd; padding:10px;">Limited diversity</td>
    </tr>
    <tr>
      <td style="border:1px solid #ddd; padding:10px;">Quota</td>
      <td style="border:1px solid #ddd; padding:10px;">Non-probability</td>
      <td style="border:1px solid #ddd; padding:10px;">Non-random filling of quotas</td>
      <td style="border:1px solid #ddd; padding:10px;">Surveys where the sampling frame is unavailable</td>
      <td style="border:1px solid #ddd; padding:10px;">Selection within quotas is not random</td>
    </tr>
  </tbody>
</table>

Sampling in Qualitative and Quantitative Research

Quantitative research sampling aims to generalize findings to a population, so it generally favors probability sampling and larger samples, because the statistical methods used to analyze the data rely on random selection. Qualitative research sampling typically aims for depth, so it commonly uses purposive and snowball sampling, and samples are usually smaller. In qualitative work, the sample is often judged by whether it gives enough information to answer the research question and not by whether it is statistically representative. For the broader design choices that shape these decisions, see What is Research Design?.

How to Choose a Sampling Method

The right method follows from the research question and the practical situation.

• Match the method to the aim. If the goal is to generalize to a population, probability sampling is the stronger choice. If the goal is to understand experiences in depth, purposive sampling may fit better.
• Check for a sampling frame. Random methods need a list of the population. If none exists, a non-probability method may be the only option.
• Consider the population’s accessibility. Hard-to-reach populations often require snowball or purposive approaches.
• Plan the sample size. A sampling method does not replace a justified sample size. In quantitative work, the required size is usually calculated from the expected effect, the variability of the data, and the desired confidence level.
• Account for resources. Time, cost, and access shape what is feasible, and a realistic plan is better than an ideal one that cannot be completed.

Because the data collected also depend on how participants are recruited and handled, Best Practices for Collecting and Handling Data is a useful companion.

Common Mistakes to Avoid

• Choosing a convenient sample and then generalizing the results to a whole population
• Using a probability method without a complete sampling frame
• Confusing quota sampling with stratified sampling, although only stratified sampling selects randomly within groups
• Overlooking selection bias, where the way participants are recruited makes the sample differ systematically from the population
• Using snowball sampling without acknowledging that it may produce a narrow, interconnected sample
• Failing to describe the sampling method clearly in the manuscript, leaving reviewers unable to judge how far the findings apply

Final Thoughts

The choice of sampling method shapes everything that follows in a study, from the kind of conclusions that can be drawn to the way the work is judged by reviewers. A method that fits the research question, is carried out carefully, and is described clearly strengthens a manuscript far more than a large sample chosen without thought. Describing the method and its limits precisely is part of reporting a study well, and AJE’s editing services help researchers communicate their methods and findings clearly to editors and reviewers.