Suppose that there are three candidate predictors, x1 , x2, and x3, for a final regression model. Suppose further that the intercept term, β0 is always included in all the model equations. How many models must be estimated and examined if one applies all possible regressions approach?
Thanks!
Hint
Our models vary based on which predictors are selected to be in a particular model.
How many ways can we make a model with $0$ of the predictors? With $1$? With $2$? With all $3$?