POPULATIONS AND SAMPLING FRAMES
    Review of Populations and Sampling Frames

    Whereas I've been talking about taking samples from "populations," it's now time to recognize that populations don't stand around waiting to be sampled for the most part.  Typically, we take our samples from lists of the members comprising a population we are interested in studying.  Such a list is the sampling frame.

    If we wanted to select a sample of students at your college, we would need to obtain a student roster, perhaps from the registrar's office.  Sometimes such lists exist in a printed form, such as a student directory.  From the standpoint of sampling, it's more useful if the list is in an electronic form, so we can let the computer do the sampling.

    Whatever it's form, a sampling frame is seldom a perfect list of the population of interest.  Some members will be omitted from the list.  Other people will have stayed on the list after having left the population in question.  It's important, therefore, that you pay some attention to how accurately the sampling frame represents the population.

    Sometimes convenience leads inexperienced researchers to use sampling frames that only approximate the populations they are interested in.  A good example would be the use of a telephone directory to select a sample of voters.  Ultimately, this practice allows you to talk about how the people listed in the telephone directory will vote, excluding voters without phones and those with unlisted numbers.