Final answer:
The statement that is not correct is the first statement: We always remove variables that contribute less than 5% to the model's R-squared. In backward stepwise selection, we consider removing the variable that contributes the most to the model, not the least.
Step-by-step explanation:
The statement that is not correct is the first statement: We always remove variables that contribute less than 5% to the model's R-squared. In backward stepwise selection, we consider removing the variable that contributes the most to the model, not the least. The goal is to eliminate variables that have the least impact on the model's R-squared.
For backward stepwise selection, the first step is to run the model with all the variables and check the R-squared. Then, in the second step, we try removing each variable in the model, one at a time, and record the R-squared value each time a variable is removed. This process continues until there are no more variables to remove or the change in R-squared becomes insignificant.