Does N Equal One? Random and Nonrandom Sampling
In a recent blog post, Karen Sternheimer refers to sociological data collection as “systematic and grounded in theory.” Therefore, reflecting on your own thoughts and opinions does not qualify as “systematic” data collection. First you are—to use sociological lingo—an N of one! In other words, your entire study consists of just one person, total! You. Your experience is not a random sample of experiences; it is not a sample by any stretch of the imagination.
Do you know what a sample is? Before we get to samples, consider that the term population refers to all of the cases that a researcher is interested in; this is true regardless of whether the researcher is interested in individuals, groups, or things. For example, if you decided to do some research on the students in your class—let’s say you wanted to learn about their career plans—then the population of interest to you is your class. What about if you were interested in the same subject, but wanted to know the career plans for all students at your school? In this second case, your population of interest is the whole school.
How many students are there in your class? So collecting data—let’s say you decide on a survey as your method of data collecting—from the entire class might be feasible. Now what about from your entire school? Is it too big for you to be able to collect data from all the students? Maybe you could collect data from everyone if you had help, but usually no one wants to help you unless you pay them. While that is reasonable, what would you do if you still wanted to answer to your research question, but could not afford to get data from the entire population? If you face the problem of not being able to get information from the entire population you’re interested in, you’re sharing a predicament faced by many other researchers.
What do we do? We study a sample—a subset of the total cases in which we are interested. We collect data on a sample of the population, knowing that if every member of the population has an equal chance of being in the sample, we can generalize the information from, for example, your random sample to the entire school (the population).
How will you define your sample? In order to decide, you should know that there are two broad categories of sampling in quantitative research: random and nonrandom. Let’s look at three types of nonrandom sampling. Accidental or convenience samples are those that researchers select based on what or who is convenient. If you are using an accidental or convenience sample, you might hand your survey to whoever you encounter around your school until you get the number of surveys you want completed.
A quota sample means that you would select your sample in proportion to some aspect of the population. For example, if your school is 60 percent female and 40 percent male, you might want that proportion of females to males reflected in your sample. If your school has 600 females and 400 males, and you decide that your sample will be made up of 100 students, your quota sample would include 60 females and 40 males.
The final nonrandom type of sample is a judgment or purposive sample. In this case, as the researcher you decide what characteristics are of interest to you and your sample consists of people (in this case) who meet that criteria. If you are interested only in learning about the career plans of students who intend to pursue graduate degrees, for example, these are the only people who would be in your sample.
I’ll discuss three types of random samples here. Simple random sampling is akin to putting all the names of the students in your school into an enormous hat and then drawing names; those people will make up your random sample. Unless you have a tiny school such a process is unwieldy; a good alternative is to use a table or a website of random numbers to select those who will be in the sample. Systemic sampling means that you would select every nth person from a list of every student at your school.
Usually, we divide the number in the population by the sample size to find n: From a school of 1,000 students for which we want a sample size of 100, we would choose every 10th student. (You could use a randomly selected number as your starting point.) If you chose to use a stratified sample, you would divide the student population into subgroups based on one or more variables of interest. You might want to examine career plans by student major; in that case you could group students by major and then take a simple random sample of each major.
Next time you hear someone representing their views as shared by “most people,” consider their sample size. Does N equal 1?