Quant #07: One-item and Composite Variables

A variable in survey research refers to a collection of attributes gathered as responses to specific questions. The collected attributes must vary for that item to count as a variable. For example, if all responses to a specific question are represented by one number (say 5), then this is not a variable; it is a constant. In contrast, if an attribute has responses spanning the range of 1 through 10, for example, then it is a variable.

When a variable can be reflected or represented by one question or item, it is referred to as a one-item variable.  One-item variables can include attributes such as gender, income, and others (as long as they vary). Comparatively, when a variable can only be represented by two or more questions or items, it is referred to as a multi-item composite index or variable. Examples of composite variables can include concepts such as attitude whose data can be collected through two or three questions. Each of these questions could be asked by use of a five-point scale (1 through 5).