Sample

The key to using samples in statistical analysis is to be sure that they are random. A random sample is one in which every member of the population has an equal chance of being selected for the sample. For example, a researcher could not choose 11th grade students for a sample if they all came from the same city, from the same school, were of the same sex, or had the same last name. In such cases, the sample chosen for study would not be representative of the total population.

Many systems have been developed for selecting random samples. One approach is simply to put the name of every member of the population on a piece of paper, put the pieces of paper into a large fishbowl, mix them up, and then draw names at random for the sample. Although this idea sounds reasonable, it has a number of drawbacks. One is that complete mixing of pieces of paper is very difficult. Pieces may stick to each other, they may be of different sizes or weight, or they may differ from each other in some other respect. Still, this method is often used for statistical studies in which precision is not crucial.

Today, researchers use computer programs to obtain random samples for their studies. When the United States government collects statistics on the number of hours people work, the kinds of jobs they do, the wages they earn, and so on, they ask a computer to sift through the names of every citizen for whom they have records and choose every hundredth name, every five-hundredth name, or to make selections at some other interval. Only the individuals actually chosen by the computer are used for the sample. From the results of that sample, extrapolations are made for the total population of all working Americans.