Working with attribute data requires much larger samples (often hundreds of items) than are used with variables. This is because, for the charts to work correctly, there must be a reasonable probability that each sample will contain at least one instance of the characteristic under examination.
Attribute charts are less sensitive, less informative, and rely on criteria that are less objective than variable charts. As such, it is always preferable wherever possible to use measurements rather than simple attributes. However, in situations where measurement is not possible, attribute charts can still prove useful, particularly in identifying areas of concern during a process of improvement.
One advantage of attribute charts is the ability to record multiple characteristics on the same chart, in preparation for Pareto analysis. We shall look at this later in the section on u-Charts.