Survey Research in Anthropology: Part 3 – Finding a Population & Sample

So, in a previous blog post, you learned that sometimes Anthropologists use surveys during their research, to collect data on the people they are studying. In other previous blog posts, you learned how to find a research topic, and then narrow that topic down into a research question. This blog post will be focusing on how to refine your research question by deciding on a population and a sample.

Deciding on a population means deciding what group of people you want to study. People in a population share some sort of a characteristic, for example, they may be the same age, sex, or from the same country or city, or work at the same company.

But who do you choose? Here are some things to consider. First, your background research should show you who other researchers chose, and explain why they chose that population. Second, your background research may show that studies have not been done on a certain population, so maybe you want to choose that one. Another thing to consider is where you live and if you are able to travel to do your research. If you want to study people in Uganda, but you don’t have funds to travel, that may not be a good choice.

Let’s go back to our made-up influenza example, where we are looking at the perceptions of influenza. In this pretend scenario, no one had considered the perceptions of immigrants in large cities in the USA. But this is still too broad of a question. You have to choose a location—you can’t study every large city in the USA. Say that you live in or near Seattle in Washington state, where there is a large population of immigrants from other countries, and so you decide to choose that city. But, immigrants in Seattle is still a huge group of very diverse people— it would be difficult to do research on all those different groups of people. You should narrow that population to something smaller. So, you do some research on immigrants in Seattle, and find out that one of the top 5 immigrant groups is from Ethiopia. You’ve always been interested in people from East Africa, and so you decide to choose Ethiopian immigrants as your population to study.

So you have decided on your study population. Now, you are not going to be able to give a survey to every single Ethiopian immigrant in Seattle. But how do you decide who to give a survey to? You need to choose a sample of people from the population you are studying. But the sample needs to be representative of the population, meaning it is like your population. But how do you do that?

There are several ways to sample a population, and here are just a few ways. First, you can do a random sample. That is where everyone in the population has an equal chance of being selected as a member of the sample. For example, you could list out all the people in the population, and then assign a number to each person on the list. Then use a random number generator to choose which numbers will be part of the sample.

Then there is systematic sampling, where you do something like choose every 10th name on a list of people, or every 10th street address on a map. Then those people become part of the sample. But how do you know how many to choose? Well, take the population size, and divide it by the sample size you want, and the result is the number you will use. So, if your population was 10,000 people, and you want a sample of 400 people, then take 10,000 and divide by 400, and that equals 25. So then, you need to choose every 25th person or every 25th street address for your sample in this scenario.

There is also stratified sampling, where you divide the population into certain groups, called strata, like males and females, or lower class and higher class, and choose a certain number of people from each group at random. The number of people chosen from each group is based on the number of total people in that group.

Another method is cluster sampling, where a population is divided into different clusters, and then a number of clusters is selected randomly, and then a number of people within the clusters are chosen randomly.

But how do you know when you have chosen enough people in your sample? How many is good enough? Well, there is a formula for that. You can search online for a sample size calculator, and just plug in your numbers, and it will tell you how big of a sample you need. For example, say you want to be 95% percent sure that your results are representative of the population (which is the amount typically used in research) and say your population is 500 people— then you need a sample of 217 people. But this doesn’t mean that all you need to do is send out 217 surveys—some people will not take part and won’t complete a survey, so you need to give out extra so that you have 217 completed surveys.

So, in this blog post, you have learned how to refine your research question by deciding on a population and a sample. Your research question is now “Perceptions of Influenza Among Ethiopian Immigrants in Seattle, Washington.” In the next blog post, I’ll talk about doing some background research through a literature review.

Want to Learn More About Survey Research in Anthropology?

Just join my Udemy class, “Exploring Surveys in Anthropology Research: Anthropology 4U.”

Thanks for reading!