Samples vary in how well they reflect a population. A sample has a bias when a part of a population is overrepresented or underrepresented. A bias is a systematic error introduced by the sampling method.
Public Opinion A newspaper wants to find out what percent of the city population favors a property tax increase to raise money for local parks. What is the sampling method used for each situation? Does the sample have a bias? Explain.
Who are the people in the sample?
The people in the sample are only those who might be selected. In this case, only those who call the newspaper.
A newspaper article on the tax increase invites readers to call the paper and express their opinions.
This is a self-selected sample. It might have a bias, depending on who calls the newspaper. The people who call may overrepresent or underrepresent some views. For example, some property owners who are against the tax might organize a campaign to get friends and neighbors to call in.
A reporter interviews people leaving the city's largest park.
This is a convenience sample, since it is convenient for the reporter to stay in one place. Because the location is near a park, the sample may overrepresent park supporters and the results will have a bias.
A survey service calls every 50th listing from the local phone book.
This is a systematic sample because the phone listing is ordered alphabetically. The regular sampling interval is every 50 listings. This sample may have a bias if there is some link between people who are listed (or not listed) in a phone book and people who pay property taxes.
One way to collect sample information is to perform a study.
In an observational study, you measure or observe members of a sample in such a way that they are not affected by the study.
In a controlled experiment, you divide the sample into two groups. You impose a treatment on one group but not on the other “control” group. Then you compare the effect on the treated group to the control group.
In a survey, you ask every member of the sample a set of questions.