Now, let’s learn about four of the most common probability methods used to select a sample and consider which is the best fit for your project.
Simple random sampling is the most basic method of random sampling. Having a random sample means that anyone in the population has an approximately equal probability of being selected for the sample. To obtain a random sample, you can randomly generate a list of names in a database, randomly select addresses, or approach people at random in a place where anyone would have an equal chance of participating in your study.
Stratified sampling involves categorizing the members of your sample and then picking a randomly from each category. This allows you to have sufficient representation from each desired category for data collection.
For Example…
Let’s say you’re doing a survey related to the relationship between language spoken in homes and access to information on COVID-19 in Monterey County, CA. If you know that 40% of the population of Monterey County are Spanish speakers, but your simple random sample is only polling roughly 10% Spanish speakers, you would want to stratify your sample because the survey answers from Spanish speakers are important to your research. Thus, you would break down your larger sample by language spoken at home (as listed by the census, perhaps) and then draw names from each sub-sample. Creating sub-samples allows you to hone in on the population you want to study.
Cluster sampling is useful if your population is particularly large or generic. When using this method, you first randomly select a portion of your population of interest. From there, you can take a random or stratified sample out of that portion (also called a cluster).
For example, if you’re studying students participating in Greek Life in universities across the United States, you might choose to narrow it down to a sub-sample or cluster of a single school’s Greek system. Then, you may take a simple random sample or stratified sample from within that cluster. Ultimately, cluster sampling helps you keep your sample size manageable while remaining representative of the subpopulation.
Systematic sampling is easier to conduct than simple random sampling, but still yields many of the same benefits. It requires the researcher to select members of a randomly ordered population at a regular interval (Thomas, 2020).
For example, if you are looking for a sample of 500 volunteers from a nationwide NGO that has 5000 volunteers, you might use systematic sampling to select every 10th volunteer in that entire population (QuestionPro, n.d.). You could do so by simply going down an alphabetical list of volunteers and selecting every 10th person.
