A Heuristic Method for Estimating the Relative Weight of Predictor Variables in Multiple Regression

Area of Science:

• Statistics
• Psychometrics
• Regression Analysis

Background:

• Assessing the relative weight of predictor variables in multiple regression is complex due to intercorrelations.
• Existing measures of relative importance are often computationally intensive and difficult to implement with more than five predictors.

Purpose of the Study:

• To propose a computationally efficient method for determining the relative weight of predictor variables in multiple regression.
• To demonstrate that the proposed method yields results comparable to more complex, established techniques.

Main Methods:

• A novel computational approach for calculating the proportionate contribution of each predictor to R-squared.
• Comparison of the proposed method's results with those from existing, more complex relative weight measures.

Main Results:

• The proposed method is computationally efficient and applicable to any number of predictors.
• Results from the new method closely align with those from established, albeit more complex, techniques.

Conclusions:

• The developed method offers a practical and efficient solution for assessing predictor relative weight in multiple regression.
• Guidelines are provided for the appropriate application of this new procedure in various research contexts.