The site requires Cascading Style Sheets (CSS) be enabled in your web browser

Survey Sampling


The purpose of survey sampling is to reduce the cost and effort that it would take to survey the entire target population. This document provides a brief introduction to the various methods of survey sampling.


Before we begin a sampling discussion, it is important to clearly define and understand three words that will be repeatedly used.


It is the complete set of observations or objects.

For example, if you are studying the behavior of ants, the population of ants is every single ant that exists on earth.


A set of observations or objects drawn from a population.

While studying the behavior of ants, you will use a sample of ants chosen from the population of ants.  A sample is selected as it is highly impractical to study the behavior of every respective ant (and impossible, too).


It is the process of selecting a subset of subjects from within a population to estimate characteristics of the whole population.  A well chosen sample will represent well the characteristics of the population.

In practical terms our "subjects" will be individuals invited to participate in a survey.

Probability versus Non-Probability Sampling

Sampling methods are broadly classified as either probability or non-probability.

Using a probability based sample, each member of the population has a known non-zero probability of being selected. Probability methods include random sampling, stratified sampling, and systematic sampling. Different probability based, random sampling techniques are suited to specific research situations, and are critical to effective research. These methods will be discussed below.

The benefit of probability sampling is that sampling error may be calculated. Sampling error measures the degree which a sample may differ from the population.  The error is expressed using a range of values. For example, polling a sample of voters, the poll's results may differ by +/-5% of the actual results of the election (e.g. the population).

In other words, such error measure tells us how likely the sample is to differ from the population. In contrast, when using non-probability sampling the degree to which the sample differs from the population is unknown.  There is no way of measuring their bias or sampling error.

In non-probability sampling, members are selected from the population in some non-random manner. This sampling method contains subjects of the population which have zero chance of selection. These include convenience sampling, quota sampling, purposive sampling, accidental sampling, judgment sampling, and snowball sampling. The main advantages of using a non-probability sampling method are convenience, cost and less technical know how (of statistics).

Most Ideal Sampling Method

The most ideal sample design is one which achieves your survey's objectives and does so within the limited resources provided to conduct the survey study. You may choose a design that provides the greatest precision without going over budget. Additionally, you should choose a sampling method that is within your comfort level of expertise.

Sampling Method Descriptions

A few of the most common sampling methods are described below.

Random Sampling
It is the purest and most common form of probability based sampling. Every member of the population is equally likely to be picked (sampled). They all have the same chance (probability) of selection. Random Sampling is also known as Simple Random Sampling. As the name implies, this method may be the ideal choice for those who are not statisticians or are using a population that is not extremely large.

Systematic Sampling
This method relies on arranging the target population according to some ordering scheme and then selecting elements at regular intervals through the target population's ordered list. Systematic sampling begins with a randomly selected item and then proceeds with the selection of every N-th element onwards.

For example, randomly open a page in a novel and in turn randomly pick a word as your first selection. Then pick every 10th word listed until you complete your sample selection count.

The systematic sampling method is simple to implement. It is also as good the Random Sampling method so long as the order of population's items are randomly listed.

Convenience Sampling
It is a non-probability method of sampling which involves the sample be drawn from a part of the population which is readily available and convenient. For example, such sample drawing may be through meeting an individual by way of a friend's introduction, in person, by phone or email. Scientific generalizations may not be drawn from such sampling method as it is not be representative enough of the population. This non-probability method is commonly used for preliminary survey efforts to get an estimate of the true, full fledged survey results, without incurring much cost or time. Convenience sampling is also referred to as opportunity sampling, grab sampling and accidental sampling.

Quota Sampling
With Quota sampling, the population is first divided into independent and mutually exclusive sub-groups. Next some discretion is applied to select the subjects from each sub-group based on a specified proportion. For example, a sample of 50 males and 30 females of Asian and Latin race/ethnicity may be selected from the population. The method of sampling is non-random. It is a non-probability sampling method which is similar to the Stratified Sampling method.

Stratified Sampling
This sampling technique is very similar to Quota sampling, with the notable exception that items are randomly selected from the population. When the population contains several distinct categories or groups respectively containing common characteristics, the population may be organized into separate "strata." Each stratum is then sampled as an independent sub-population, whereby individual subjects may be randomly selected. Examples of stratums are: males and females, students and teachers, buyers and sellers, to name a few.

The relevant stratums and their actual representation in the population are identified first. Next, random sampling is used to select a sufficient number of subjects from each stratum. In this context, sufficient means a sufficiently large sample size to be reasonably confident that the stratum represents the population. The larger the sample, the more likely it will be to represent the population. Stratified sampling is often used when one or more of the stratums in the population have a low incidence relative to the other stratums (e.g. homogeneous subgroups). Identifying strata and implementing such an approach may increase the cost and complexity of sample selection.

Panel Sampling
Panel sampling is the method of first selecting a group of participants through a random sampling method and then asking the group for the same information again using multiple surveys over a period of time. Panel sampling based studies provide the researcher to find out why changes in the population are occurring, since they use the same sample of people every time. That sample is called a panel. This type of sampling is common with health research. For instance, tracking the effectiveness of a drug over time on patients participating in the panel. Panel studies, while they can yield extremely specific and useful explanations, can be difficult to conduct. They tend to be expensive, they take a lot of time, and they suffer from high attrition rates. Attrition is when people drop out of the study.

Snowball Sampling
Snowball sampling is a non-probability sampling method whereby existing survey participants solicit and recruit prospective survey respondents from among their acquaintances.

There are a few advantages to this sampling technique. Online surveys are immediately accessible to social networking. It is good for reaching an audience the survey owner is not aware of or a population that is difficult to contact. It is possible for the surveyors to include people in the survey that they would not have known. One immediate negative is it introduces selection bias and it may be likely the sample will not be a good representation of the population.

Snowball sampling is also known as: referral sampling, chain-referral sampling, and chain sampling.

More Information

This web page provides a brief introduction to survey sampling and a few more the most common methods used for conducting surveys. Survey sampling is a lengthy subject and requires a deeper discussion for one to have a complete understanding of the subject. If you desire to learn more about this subject, we recommend you consult our book recommendation list.

Last updated: 2012 January 24