Determining Relative Importance with Independent Variable Groups: An Alternative Dominance Analysis Method
Journal of Behavioral Data Science · 18.01.2026 · 6 (1) · 41-67
DOI: 10.35566/jbds/luchman ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Behavioral Data Science, online erschienen am 18.01.2026. Verfasst von Joseph Luchman. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 20 Literaturangaben.
Abstract (Originalsprache)
Grouping independent variables (IVs) in relative importance analyses like dominance analysis (DA) can reduce computation time and improve analysis feasibility. A side effect of grouping IVs is that determining the importance of individual IVs within groups is not possible. This work proposes an extension of DA where the researcher can group IVs yet still compare individual IVs. The proposed extension extends from Owen values, a variant of the Shapley value solution concept from Cooperative Game Theory. Owen values are used to generate within-group DA statistics and designations which can be used for relative importance analysis of individual IVs in IV groups. This manuscript provides an analytic example with data in the R statistical computing environment that shows how within-group DA statistics and designations are determined. The manuscript concludes by discussing how to group IVs into IV groups and potential extensions to the method to simplify the IV grouping process.
Zitieren
APA
Luchman, J. (2026). Determining Relative Importance with Independent Variable Groups: An Alternative Dominance Analysis Method. Journal of Behavioral Data Science, 6(1), 41-67. https://doi.org/10.35566/jbds/luchman
Vancouver
Luchman J. Determining Relative Importance with Independent Variable Groups: An Alternative Dominance Analysis Method. Journal of Behavioral Data Science. 2026;6(1):41-67. doi:10.35566/jbds/luchman
BibTeX
@article{Luchman2026,
author = {Luchman, Joseph},
title = {Determining Relative Importance with Independent Variable Groups: An Alternative Dominance Analysis Method},
journal = {Journal of Behavioral Data Science},
year = {2026},
volume = {6},
number = {1},
pages = {41--67},
doi = {10.35566/jbds/luchman}
}
EUVIMED