Introduction: Why Perform Simple Descriptive Analyses?

The procedures discussed in this chapter are useful for (at least) three important purposes. The first involves the concept of data screening. Data screening is the process of carefully reviewing data to ensure that they were entered correctly and are being read correctly by the computer. Before conducting any of the more sophisticated analyses to be described in this text, you should carefully screen your data to avoid computational errors (e.g., numbers that were accidentally entered, out-of-range values, numbers that were entered in the wrong column). The process of data screening does not guarantee that your data are correct, but it does increase the likelihood by identifying and fixing obvious errors.

Second, these procedures are useful because they allow you to explore the shape of your data distribution. Among other things, understanding the shape of your data helps you choose the appropriate measure of central tendency (i.e., the mean, mode, or median). In addition, many statistical procedures require that sample data are normally distributed, or at least that the sample data do not display a marked departure from normality. You can use the procedures discussed herein to produce graphic plots of the data, as well as test the null hypothesis that the data are from a normal population.

Finally, the nature of the research question itself might require use of a procedure such as PROC MEANS or PROC FREQ to obtain a desired statistic. For example, if your research question is “What is the average age at which women married in 1991?” you could obtain data from a representative sample of women who married in that year, analyze their ages with PROC MEANS, and review the results to determine the mean age.

Similarly, in almost any research study it is desirable to report demographic information about the sample. For example, if a study is performed on a sample that includes participants from a variety of demographic groups, it is desirable to report the percent of participants of each sex, the percent of participants by race, the mean age, and so forth. You can also use PROC MEANS and PROC FREQ to obtain this information.