Abstract

Purpose: This exploratory study developed and provided initial validation evidence for the Korean Stowell-based Coaching Leadership Scale (K-SCLS), a 16-item instrument measuring coaching leadership in Korean SMEs, while examining the structural nature of coaching leadership dimensionality through a Multi-Level Integration Framework. Method: A sample of 300 SME employees (Mage = 20s–60s; male = 51.3%) completed an initial 22-item instrument refined to 16 items via exploratory and confirmatory factor analyses. A split-half cross-validation (n₁ = n₂ = 150) addressed same-sample EFA-CFA concerns. Bifactor modeling, hierarchical CFA, and sequential measurement invariance testing were employed alongside HTMT-based discriminant validity analysis. Results: Cross-validation confirmed item selection stability (87.5% correspondence) and replicated the four-factor structure in an independent subsample (CFI = 1.000, RMSEA = 0.000). The correlated four-factor model demonstrated excellent fit (CFI = 1.000, RMSEA = 0.008; α = .938–.949; AVE = .822–.843). Bifactor modeling revealed a dominant general coaching leadership factor accounting for 90.1% of common variance (ECV = .901; ωH = .958), with negligible specific factor contributions (ωS = .041–.172). HTMT ratios (.843–.933) and interfactor correlations (.845–.943) indicated subfactors were not empirically distinct. Full measurement invariance across gender was achieved (scalar ΔCFI = 0.000). Conclusion: These findings reconceptualize the K-SCLS as a bifactor-structured instrument wherein four content domains reflect a superordinate general construct, reframing two decades of discriminant validity failures as a substantively meaningful integration phenomenon theoretically consistent with Korean cultural constructs including jeong, woori consciousness, and chemyon. Residual common method variance (43.0%) and the single-source design necessitate multi-source replication.