Before Validation: How Transition Operationalization Constructs Bibliometric Process Phases

This preprint presents a diagnostic study of measurement-path dependence in bibliometric phase construction, using delayed-recognition ("Sleeping Beauty") citation histories as a phase-explicit hard-case testbed. Working from a fixed candidate space of 1,351 delayed-recognition papers constructed independently of the transition operators under comparison, the study evaluates whether three literature-grounded boundary operators (Ke et al., 2015; Du & Wu, 2018; Fang, 2018) locate the same pre-awakening/post-awakening transition when applied to identical citation histories. Exact pairwise boundary agreement ranges from 26.8% to 41.2%, and disagreements in transition location propagate into changes in event membership and disciplinary/documentary composition in the large majority of cases. Two independent calibration frameworks—temporal-exchangeability permutation testing and same-displacement reference comparison—are used to assess whether the observed compositional divergence is substantively unusual. The findings distinguish phase-membership sensitivity from aggregate representational magnitude and argue that the operational stability of a reconstructed bibliometric phase is a measurement question that should be established prior to any substantive interpretation of the phase's content. This is a preprint; the manuscript has not yet completed peer review.