Principal Component Analysis in MCDM: An exercise in pilot selection

Abstract

Assignment of criteria weights during group multi-criteria decision-making (MCDM) processes is a challenging and time-consuming process often rife with subjectivity. This is particularly evident in hiring or personnel selection processes. Both the military and corporate aviation communities have required a large influx of pilots recently due to an increasing pilot shortage. This paper outlines a study conducted to evaluate the effectiveness of principal component analysis (PCA) as an objective weight assignment method to establish a rank order of United States Army pilots based upon their emotional intelligence, safety attitude, and safety citizenship scores. PCA results were compared to an algorithm tracking the rank order of pilots using every possible criteria weighting combination to test PCA's validity as a criteria weight assignment method. Regression analysis demonstrated a strong correlation (p < 0.000) between the algorithm ranking results and PCA’s ranking results, reinforcing PCA's validity as a quantitative approach to criteria weight assignment in group MCDM applications. Implications of this study include increased efficiency and reduced subjectivity in group-based MCDM pilot selection processes.