Sampling, Determining Size

Populations

In an ideal word, all research would be done on a population or the entire group of people a researcher is studying. This is because when a researcher conducts research on the entire group of people he or she wants to study, every person is represented in the results. When everyone is represented the researcher does not have to use statistics to make inferences about the population. When a researcher does not have to make inferences, there is no concern that the results contain statistical error. The researcher, for example, does not have to prove that the sample is similar enough to the population to draw the conclusions made from the data. This makes the conclusions easier to defend because there will be fewer questions about the process.

However, studying entire populations is typically not practical and is often impossible. A great example is the U.S. Census; a great deal of time and money goes into tracking down every person in the United States every 10 years, and without fail not everyone is included in the report. A researcher must remember that no matter how hard he or she tries to contact everyone on a list, people change contact information, people travel, some people refuse to participate, and the list of practicality issues goes on and on.

There are even more practicality issues when a list of people in a population does not exist. A researcher first has to attempt to create a list, track people down, and then convince them to all participate in the study. This is not easy and sometimes not even possible. For example, a researcher who wants to study the effect of eye color on SAT scores is not going to find a list that includes eye color for all of the people that have taken the SAT test.

Other lists such as members of a specific fraternity across the country may be more easily created, but even if it is possible, it is going to take a great deal of time and money to do so. The researcher would have to search online and make many phone calls just to create the list. Then the researcher would have to find contact information for all of these people and get each of them to participate. This all assumes that everyone on the list is still alive and able to participate. Given that research using populations is filled with feasibility issues, most researchers rely on probability theory, which allows them to use samples rather than populations.

...