4.2.3 Joint descriptive statistics for two or more variables

The descriptive statistics presented above were applied to the variables of interest separately, for demonstration purposes. This rarely happens in practice – instead, two or more variables are generally jointly presented to assess the type of relationship between them.

Such relationships are generally considered between one variable as the outcome variable, and considered between one or more variable(s) as the exposure variable(s). We are interested in how the values taken by the outcome variable vary for different values of the exposure variable(s). How the relationship is explored depends on the type of variables.

• The relationship between one categorical variable and one (or more) categorical variable(s) can be displayed in a contingency table or cross-tabulation. An example contingency table for jointly describing two or more variables is shown in this separate PDF file.
• The relationship between one numeric variable and one (or more) categorical variable(s) can be displayed in a summary table, as shown in Figure 4, or graphically (see course Summarising and presenting data).
• Last, the relationship between a numeric variable and another numeric variable can be displayed graphically (if it is inconvenient to display such variables in a table).