Refers to a quality characteristic that meets or does not meet product specification. These characteristics can be categorized and counted. Typical examples include sorting and counting the number of blemishes in a particular product (defects), or the number of nonconforming pieces (defectives).
Suppose you want to investigate the quality of a bag of M&Ms. You could rate each candy for different defects such as chips in the candy coating, legibility of the "m" printed on each candy, flat spots or misshaped candies. For example, evaluate chips in candy coating.
|
Attribute |
Rate as... |
Data |
|
Does the candy have a chip? |
Pass or fail |
Attribute with 2 categories |
|
How many chips per candy? |
0, 1, 2, 3, > 4 chips |
Attribute with 5 categories |
|
Note |
Attributes are often evaluated as either pass or fail, or are compared with visual standards with generally up to five different categories. |
Attribute data are normally collected when standard measurements are difficult to obtain. Compared with continuous variables that can take on infinite values, an attribute variable can only be classified into specified categories. In the M&M example, if you had measured the size of the chip in the candy coating, it would be continuous data. So while continuous measurements usually give much more information, attribute data are generally more convenient and easy to collect. Also, as seen in this example, attributes are usually subject to rater interpretation. For example, one rater may define a chip differently from other raters.