ABSTRACT

Before computing any inferential statistics, it is important to do exploratory data analysis (EDA), as outlined below. This chapter helps you understand your data, helps you to see if there are any errors, and helps you to know if your data meet basic assumptions for the statistics that you will compute. To enable you to understand your data, you will undertake different types of analyses and create different types of plots depending on the level of measurement of the variables. As discussed in Chapter 3, this program labels the levels of measurement nominal, ordinal, and scale. Remember that we think it is useful to distinguish between nominal and dichotomous, and we think that normal is more descriptive than the term scale. Keep in mind that there are times when whether you call a variable ordinal or scale might change based on your EDA. For example, a variable that you initially considered to be ordinal may be normally distributed and thus may be better thought of as normal. Recall from Chapters 2 and 3 that making the appropriate choice requires that you understand your data; thus EDA should help guide your selection of a statistic.