Descriptive Statistics

This chapter describes the methods and use of descriptive statistics (summary statistics). In the context of the scale of measurement, the relevant approach to describe the data is discussed. We begin with frequency tables, followed by crosstabulations for the data measured on categorical scale. A relevant and brief discussion on the Chi-squared test is also undertaken in the context of crosstabulations. We then describe the measures of central tendency (mean and median) and some selected measures of dispersion. We discuss how to generate each statistic separately as well as all the measures of central tendency and dispersion in one command. We discuss three approaches to examine the normality of data: Kolmogorov-Smirnov, Shapiro-Wilk and Jarque-Bera. We also explain the situations to which each of these tests is relevant. Normality tests may apparently seem like hypothesis testing, though it is also an approach to describe the data. A lot of these tests are sensitive to missing values. In each section, we discuss how to handle missing values, largely focusing on omitting the missing values. We end the chapter by demonstrating how to get summary measures as well as normality indicators, in just one command.