What gets lost when gaze data get lost: archival reproducibility and interpretive accountability in eye-tracking research

Purpose Complex digital research records do not remain scientifically useful merely because files survive. This paper argues that reproducibility depends on preserving the evidentiary chain connecting recorded traces, processing decisions, derived variables, and published claims. Eye-tracking research is used as a revealing case because gaze data acquire scientific meaning only through multiple layers of documentation and transformation. Design/methodology/approach This conceptual methodological paper synthesizes eye-tracking reporting guidance, research-transparency principles, and a reflexive legacy case in which the original fixation-level evidence could no longer be reconstructed. No participant-level findings from the legacy project are reported or reanalyzed. Findings Four preservation requirements are identified: retain raw and derived gaze data separately; document transformations from gaze streams to analytic events; preserve construct-level distinctions among attention, interest, choice, and attitude; and maintain participant linkage and flow records across modalities. A minimal archival checklist is provided. Research limitations/implications The paper develops a conceptual and practice-oriented framework rather than empirically testing archival interventions. Future research can evaluate how preservation practices affect the reproducibility and reuse of complex digital behavioral data across research settings. Practical implications Researchers should preserve not only data files but also the metadata, processing decisions, linkage information, and construct definitions needed to reconstruct how recorded behavior became analytic evidence. Originality/value The paper reframes archival continuity as a problem of documentation and interpretive provenance rather than file retention alone. It introduces the evidentiary-chain perspective: a scientific record remains reusable only when future researchers can reconstruct how recorded behavior became an analytic variable and, ultimately, a substantive claim.