Society of Actuaries (SOA) PA Practice Exam

Which method focuses on assessing the improvements to model fit when adding variables?

1. Backward Selection

2. Forward Selection

3. Stepwise Selection

4. Least Squares Selection

The correct answer is: Forward Selection

The method that focuses on assessing the improvements to model fit when adding variables is the one known as Forward Selection. This approach begins with a model that includes no independent variables, then adds them one at a time based on specific criteria, usually related to how well they improve the fit of the model. Each potential variable is evaluated to determine if its inclusion significantly improves the model's explanatory power or predictive accuracy, often measured by metrics such as R-squared or adjusted R-squared. In contrast, other methods like Backward Selection start with a full model and sequentially remove variables, while Stepwise Selection combines forward and backward techniques to refine the model. Least Squares Selection is not primarily focused on the stepwise addition of variables and does not specifically assess the improvement of fit through the inclusion of those variables in the same manner as Forward Selection. This distinct focus on adding variables systematically makes Forward Selection particularly valuable for understanding the incremental benefits each variable can contribute to the model.