Random Samples in Research Studies

The term "random sample" comes up a lot when you're reading about medical research. Understanding this term can help you interpret those health studies you come across in the news and get a better grasp of how they may, or may not, apply to you.

Simply put, a random sample is a subset of individuals randomly selected by researchers to represent an entire group as a whole. The goal is to get a sample of people that is representative of the larger population.

For example, if researchers were interested in learning about alcoholic use among college students in the United States, the larger population (in other words, the "group of interest"), would be made up of every kid in every college and university in the country. It would be virtually impossible to interview each and every one of these people to find out if they drink, what types of alcohol they drink, how often, under what circumstances, how much (a beer or two per week versus enough to get intoxicated every weekend), and so forth. Instead of undertaking such a gargantuan task, scientists will pull together a random sample of college students to represent the total population of college students.

How Researchers Create Random Samples

Random sampling can be costly and time-consuming. However, this approach to gathering data for research does provide the best chance of putting together an unbiased sample that is truly representative of an entire group as a whole.

Going back to the imaginary study of alcohol use among college students, here's how random sampling might work. According to the National Center for Education Statistics ( NCES), approximately 19.7 million students were enrolled in U.S. colleges and universities in 2020, the most recent statistics available. These 20 million individuals represent the total population to be studied.

For the purpose of drawing a random sample of this group, all students must have an equal chance of being selected. For example, scientists conducting the study would need to make sure that the sample included the same percentage of men and women as the larger population. According to the NCES statistics, 11.3 million of the total population of college students are female and 8.5 million are male. The sample group would need to reflect this same ratio of women to men.

Besides gender, researchers would also want to go through the same process for other characteristics—for example, race, cultural background, year in school, socio-economic status, and so forth, depending on the specific purpose of the study. For instance, if they wanted to home in on alcohol use among Asian students, they would create a random sample consisting only of Asian students. By the same token, if the study was focused on how much students drink during the week, they would create a questionnaire or other method for finding only kids who drink on weekdays for their research.