Random Sample

It is not always possible or feasible to collect data from every member of a population because of time constraints, lack of resources, or difficulty in identifying or locating all members of a specified group. In these cases, the researcher must select a sample from the larger group known as the target population to whom the researcher wants to generalize. Random samples allow the researcher to use inferential statistics to estimate the extent to which the sample differs from the population.

The target population is the group to which the researcher wishes to generalize, and the sample is a finite subset of the population. For example, if the researcher wished to generalize about the effect of comprehension strategies on first-grade students' reading ability, first-grade students would constitute the population. This group of first-grade students may be considered a sample of present and future first-grade students (sometimes referred to as a time/place sample) in one school, a school district, or state.

Usually one or more samples from a target population are used in a study. In these cases, researchers should follow systematic procedures for identifying the population, selecting the sample (subjects), and assigning subjects to treatment and control groups if findings are to be generalized beyond the study group back to the target population. A representative sample helps to ensure that the generalizations or inferences made from the sample to the population are free of systematic bias and present a fair view of the target population. The best way to achieve a representative sample is through random sampling techniques. Different procedures may be used to select a random sample. Random sampling is often referred to as simple random sampling to distinguish it from other kinds of sampling, such as stratified random sampling, cluster sampling, and systematic sampling.

A simple random sample is the most basic kind of probability sample. Probability sample selection procedures are those in which each unit or subject in the population has a known, nonzero chance of being selected. For simple random sampling, each case in a finite population has an equal chance of being selected, and the selection of any one case is independent of the selection of any other case. For a sample to be truly random, systematic procedures should be employed to select the subjects. The recommended method for selecting a sample using simple random procedures is to use a table of random numbers. Tables of random numbers are usually computer generated, and they are often included in textbooks and on the Internet. To use a table of random numbers, the researcher first assigns an identification number to each subject in the population. Next, the researcher randomly selects a starting place on the table. This should be done by blindly placing a pointer on the [Page 840]table to identify the row and column from which to start selecting numbers. Only those subjects whose identification number matches the number in the table are included in the sample. For example, to select a sample of 100 students from a list of 600, the researcher might number the students from 000 to 599, refer to the table, and systematically move through the table selecting only those students whose identification numbers fell between 000 and 599 until 100 students were selected.

...