FIRE: an SPSS program for variable selection in multiple linear regression analysis via the relative importance of...

Area of Science:

• Statistics
• Psychometrics
• Data Analysis

Background:

• Variable selection is crucial for building accurate multiple linear regression models.
• Existing methods may not fully capture predictor importance or ensure model generalizability.

Purpose of the Study:

• To present an SPSS program for variable selection in multiple linear regression.
• To implement techniques based on the relative importance of predictors.
• To provide a comprehensive approach for model development and validation.

Main Methods:

• Optimal data splitting for cross-validation.
• Sequential selection of predictors based on relative importance.
• Assessment of the selected model using standard statistical indices and procedures.

Main Results:

• The developed SPSS program facilitates a structured approach to variable selection.
• The methodology ensures that selected predictors contribute meaningfully to the regression model.
• Cross-validation enhances the generalizability of the final regression model.

Conclusions:

• The provided SPSS program offers a practical tool for researchers conducting multiple linear regression.
• The relative importance approach enhances the interpretability and reliability of regression models.
• This methodology supports robust model building and validation in statistical analyses.