An essential first step in any analysis is to become familiar with the data. While total costs, length of stay, and birthweight are mentioned in the problem statement and are of primary interest, it is also valuable to provide a descriptive analysis for the other variables. This gives the analyst additional insight for interpreting the linear regressions in the problem context. For example, what is the distribution of severity of illness or the type of payment?

Figure 12.3 Descriptive Analysis for Total Costs, Length of Stay, and Birthweight shows a table of descriptive statistics and histograms for the three variables of primary interest – total costs, length of stay and birthweight (in pounds). It is good practice to include units of measure with column names so that they will appear in tables and graphs.

Figure 12.4 Descriptive Statistics for Other Variables shows descriptive statistics for the other variables in the data set. These tables were created with the Tabulate function with the variables grouped by categories such as demographics, admission, diagnosis, and payment.

For some variables, all observations have the same value. In some cases this is due to the subset being analyzed (e.g., the Hospital County is Clinton for all observations since we are only looking at one hospital). In other cases, such as Emergency Department Indicator, variation is possible but for CVPH in 2014 no newborns were admitted through the Emergency Department. It is tempting to omit the description of Emergency Department Indicator because there is no variation, but including this information adds insight in the problem context.