A variable is an attribute of an object. The value of a variable can vary from one thing to another.

There are two different types of variables:
  • Quantitative (or numeric) - data are numeric values of a quantity.
  • Qualitative (or categorical) - data differs only in kind.

Record quantitative data as the numeric value.

Record qualitative data as a numeric coding (for example, 0, 1, 2) or labels (for example, Low, Medium, High). Numeric codings are often preferred as they succinctly represent the values and allow quicker data entry. However, codings can be difficult to understand by anyone other than the person who assigned them. To make their meaning clear, you can also assign a label to each numeric coding.

Data are ordered based on numeric value, numeric coding, or if a numeric coding is not used into alphabetical label order. For some statistical tests, the sort order determines how observations are rank-ordered and can affect the results. Often alphabetic sort order does represent the correct ordering. For example Low, Medium, High is sorted alphabetically into High, Low, Medium which does not represent their correct, natural order. In such cases, you must explicitly set the order before analyzing the data.