This is an “either or” situation. Attribute data is binary only. One either records that a process gives an attribute or it doesn’t. Basically, the equivalent of a yes or no question.

Note: This definition is highly debated. Some experts agree with the definition above and some say that attribute data has the same definition as discrete data. Try not to get confused by this.

Example 1: Let’s say Rudy is a unicorn breeder. She has a special breed of horse that can possibly become unicorns. A horse is only categorized as a unicorn if it has the attribute of a big fat horn growing out of its forehead.

One male unicorn, named Fred, that she uses for breeding produces 60% horned offspring when crossed with another female unicorn.

Her other male unicorn, named Brad, that she uses for breeding produces 35% horned offspring when crossed with another female unicorn.

The attribute data for Fred is: 60% of offspring has the attribute of a horn and 40% do not have the attribute of a horn.

The attribute data for Brad is: 35% of offspring has the attribute of a horn and 65% do not have the attribute of a horn.

This kind of data is important to Rudy because if she can’t sell a dang horse as a unicorn if it doesn’t have the dumb horn!

 

Example 2: 99.7% of the beer bottles coming off a bottling line have the convenient attribute of a cap. If 1,000,000 6-packs are sold, then a few less than 3,000 people need to watch out for sticky car seats.