The data type required for all attributes in a data set in order to perform statistical correlation is _________.

• A. Boolean
• B. Categorical
• C. Text
• D. Numeric

Answer: (D)

Correlation looks at how numeric values of two attributes move together. To compute it, you need to perform arithmetic on the data: means, deviations, and products of the values. That kind of calculation only makes sense when the attributes are numeric. If an attribute is categorical or text, you can’t derive a meaningful correlation in the standard sense without first converting it into numbers (and even then you’re changing what you’re measuring). Boolean values can be represented as 0/1, but the typical requirement for calculating a standard correlation is numeric data across attributes.