Question

Which of the following statements is not true?

1.An interaction term can contain any combination of discrete and continuous variables
2.When we are making decisions about whether to keep an interaction term in our model, we can use changes in R-squared
3.When we are making decisions about whether to keep an interaction term in our model, we can use changes in prediction accuracy
4.In order to form an interaction term, one of the variable has to be binary and the other has to be continuous

Answer 1

An interaction term does not require one variable to be binary and the other to be continuous. Changes in R-squared and prediction accuracy are used to decide whether to keep an interaction term in a model.

Step-by-step explanation:

The statement that is not true is: 4. In order to form an interaction term, one of the variables has to be binary and the other has to be continuous.

An interaction term is a statistical concept used in regression analysis to capture the combined effect of two or more variables on the outcome variable. It does not require one variable to be binary and the other to be continuous; the variables can be of any type (discrete or continuous).

When deciding whether to keep an interaction term in a model, changes in R-squared and prediction accuracy are considered, as mentioned in statements 2 and 3.