© John Wiley & Sons, Inc.

FIGURE 17-4: Observed versus predicted outcomes for the model SBP ~ Age + Weight, for the data in Table 17-2.

Watching Out for Special Situations that Arise in

Multiple Regression

Here we describe two topics that come up in multiple regression: interactions (both synergistic and

anti-synergistic), and collinearity. Both relate to how the simultaneous behavior of two predictors can

influence an outcome.

MODEL BUILDING

If you have a big data set with many variables, how do you plan which predictors to try to include in your multiple regression

model? Once you choose the ones you want to consider, how do you decide which ones to keep and which ones to remove

to achieve the best-fitting model? Biostatisticians approach model building using different methods, but the goal of all of these

is to achieve the best-fitting model that explains the relationship between the predictors and outcome, and to do it in a

transparent way. Chapter 20 includes a section explaining how to develop a modeling plan if you have a choice of many

potential predictor variables.