Abstract
<p>This study presents the first longitudinal, population-census validation of the Passion Scale (Vallerand et al., 2003) in professional sport officials. A complete registered population of 619 Brazilian football referees (Federação Paulista de Futebol) was assessed across five measurement waves (T1–T5; 2010–2013), yielding 1,834 observations. Eight aims were addressed: (1) cross-sectional psychometric validation at T1 including CFA and ESEM; (2) longitudinal replication across T2–T5; (3) scalar measurement invariance across five waves; (4) within-person variability of harmonious (HP) and obsessive (OP) passion; (5) latent growth models of HP, OP, HP-OP Balance Index (IP), and Passion Criteria (PC); (6) role-specific passion dynamics (main referees vs. assistant referees); (7) internalization pathway analysis using pre-entry motivational data (mixed-methods approach); (8) exploratory factor analysis of antecedent instruments. Full scalar invariance was supported (ΔCFI = −.007). HP increased monotonically (d = 0.53). OP showed greater within-person variability than HP (d = 0.48; 66.1% of referees). Pre-entry motivational data mapped onto DMP-predicted internalization pathways at the item level, with antecedent EFA yielding three theoretically coherent factors. The census design eliminates sampling bias and allows population-parameter estimation. Findings extend DMP boundary conditions to part-time professional practitioners and provide the first longitudinal item-level examination consistent with the DMP antecedent sequence.</p>