Measuring Within-Person Variation in Written Communication Patterns Across Social Contexts

Abstract Personality psychology increasingly recognizes within-person variation across contexts as a core feature of personality functioning, yet empirical measurement relies primarily on self-report and prompted experience-sampling methods. Naturalistic written communication offers a complementary behavioral data source, but no standardized analytic framework exists for quantifying within-person variation in everyday text. This paper introduces a transparent, fully specified pipeline for measuring context-linked variation in written communication patterns using longitudinal messaging data. We analyzed naturalistic messages authored by a single individual across five distinct social environments and four romantic dyads, spanning multiple years of interaction. Communication indicators - including emotional language, interpersonal orientation, cognitive elaboration, and humor markers - were derived using deterministic lexicon-based extraction with complete specification of all scoring procedures. Results show systematic differences in communication patterns across relational contexts, with stability within environments and divergence between environments. Convergent analysis using an independent lexicon framework (LETA) replicated key context-linked differences, supporting the robustness of observed patterns across measurement approaches. The contribution is methodological: a reproducible, fully transparent framework for extracting personality-relevant behavioral indicators from naturalistic written communication. The pipeline complements existing measurement approaches by operating on unprompted behavioral data and provides a foundation for multi-person validation and extension to applied settings.