15.5. Rational approaches for dealing with incomplete data

In a rational approach, you use mathematical or logical relationships among variables to attempt to fill in or recover the missing values. A few examples will help clarify this approach.

In the sleep dataset, the variable Sleep is the sum of the Dream and NonD variables. If you know a mammal’s scores on any two, you can derive the third. Thus, if there were some observations that were missing only one of the three variables, you could recover the missing information through addition or subtraction.