GEN-REPR v2.1: A Generalized Reproducibility Framework for Experimental Materials Research (Prototype Implementation Version)
Note on this version: This record contains GEN-REPR v2.1, updated to include Section 4.1 (Reference Implementation). This section details the working software prototype (application version 0.2.0) developed to operationalize the 23 modules, including JSON schema validation and automated reproducibility classification safeguards. This specific version is deposited here to establish a permanent DOI for the reference implementation. The manuscript is currently under peer review at Research Integrity and Peer Review. Abstract:Reproducibility in experimental research depends not only on repeating an experiment, but on preserving sufficient information to reconstruct the material, process, measurement conditions, analytical pathway, and provenance of a reported result. Existing reporting practices address important components of this problem, but requirements may remain distributed across methods descriptions, supplementary information, data policies, specialist standards, and repository practices. Incomplete reconstruction records undermine research integrity across disciplines: they impede independent verification, obscure selective reporting, and weaken the evidence base on which policy and practice depend. This paper presents GEN-REPR v2.1, a proposed generalized reproducibility framework organized as a version-controlled, traceable architecture rather than as an isolated checklist. Although the architecture is illustrated with experimental materials research—including a retrospective mapping to a published laser powder bed fusion study—the Core is designed for cross-domain use through domain-specific annexes that adapt operational detail without altering the shared evidentiary pathway. The framework contains 23 Core modules spanning scope and material identity, feedstock, processing, thermal history, specimen preparation, characterization, property measurement, statistics, instrumentation, environmental conditions, deviations, acceptance criteria, negative and partial results, provenance, raw-data preservation, computational transparency, version control, reproducibility classification, independent validation, deposition, and standardized reproducibility statements. A central design principle is to distinguish documentation from demonstrated reproducibility. The revised manuscript adds a prespecified validation program (V0-V9) comprising structural consistency testing, controlled information-loss tests, retrospective literature audits, cross-framework comparison, inter-assessor reliability testing, machine-readable conformance tests, independent implementation, controlled replication, and usability assessment.
Authors
- Raymond Scales
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-13
- DOI
- https://doi.org/10.5281/zenodo.24193548
- Primary Topic
- Research Data Management Practices
- Type
- preprint