Types of Samples

In the video on populations and samples, we learned about probability samples, and specifically about simple random samples, and about non-probability samples. However, there are different types of probability and non-probability samples and it's important to understand them. These different types of samples are briefly introduced below.

Probability Sample

• Everyone in the population has some probability of being in the sample.
• Provides representative samples
• Allow inferences to me made about the population

There are many types of probability samples. The video introduced simple random sample, but there are others that might be more appropriate given our population, interests, or limitations to how we can sample. Below are some examples of probability samples.

Simple random sample

• Every unit in the population has the same probability of being selected into the sample.

Stratified random sample

• Units of the population are split into mutually exclusive groups based on some characteristic, and then a sample is drawn from each group using simple random sampling.

Cluster sample

• Units of the population are split into mutually exclusive groups based on some characteristic, and then a sample of the groups are taken using simple random sampling.

For more details about probability samples see hereLinks to an external site. or hereLinks to an external site..

Non-Probability Samples

• Not every unit in the population has a chance of being in the sample
• Not representative of the population
• Cannot be used to make inferences about the population

While most of the time we prefer some type of probability sample, it's not always feasible to accomplish this. If this is the case, we typically rely on some type of non-probability sample. Some examples of non-probability samples are described below.

Convenience sample

• Units in the sample are chosen because they are the most accessible to the researcher.
• For example, for professors conducting research on people, students in their class can be a convenience sample.

Voluntary response sample

• Sample is composed of units that voluntarily enter into the sample.
• For example, a researcher sends an email to a survey to all students at UCSD, but only those that want to fill it out do so.

Purposive sample

• A researcher chooses specific units of the population based on their knowledge about the population.
• For example,  a researcher wants to study the strategies activists use for organizing protests, so they purposefully select a subset of activists organizing protests to interview.

Snowball sample

• Units in the sample are recruited by other participants.
• For example, a researcher wants to study local politicians, and after each interview the researcher asks the participant to recruit other local politicians they know.

For more details about non-probability samples see here or here.