A Nonparametric Approach to Composition-Adjusted Time-to-First-Citation Analysis

This paper introduces a standardised nonparametric framework for the comparative analysis of time-to-event data in heterogeneous bibliometric settings. Focusing on time-to-first-citation, the proposed approach combines stratified Kaplan-Meier with standardised representations to enable composition-adjusted comparisons across journals with different editorial and disciplinary profiles. The resulting indicator provides a bounded summary of early citation uptake. Variability is assessed through a journal-level resampling procedure that preserves within-journal dependence. The framework is deliberately model-free and avoids restrictive parametric assumptions. An empirical application to citation records from leading journals in bibliometrics and research evaluation illustrates its practical relevance and its advantages over conventional immediacy-based measures.