Relevance relations for the concept of reproducibility

Area of Science:

• Scientific methodology
• Systems science
• Biomedical research

Background:

• Reproducibility is a fundamental scientific principle, but its application is questioned in complex systems and biomedical research.
• Interdisciplinary research presents unique challenges to reproducibility due to varying descriptive levels of systems.
• Distinguishing relevant features across different descriptive levels is crucial for accurate reproduction of results.

Purpose of the Study:

• To propose a general framework for establishing 'relevance relations' between system complexity and description granularity.
• To introduce 'contextual emergence' as an operational method for implementing these relevance criteria.
• To develop level-specific, consistent criteria for scientific reproducibility in complex, interdisciplinary studies.

Main Methods:

• Developed a formal scheme for a 'relation of relevance' linking system complexity and descriptive granularity.
• Defined an interlevel relation, 'contextual emergence,' to translate between descriptive levels.
• Proposed a procedure for constructing level-specific reproducibility criteria within a consistent framework.

Main Results:

• The proposed 'relevance relations' and 'contextual emergence' provide a formally sound and empirically applicable method.
• This approach allows for the translation between different descriptive levels, enabling consistent reproducibility criteria.
• The framework challenges the notion of a single fundamental ontology, offering a nuanced view of scientific understanding.

Conclusions:

• The 'relation of relevance' and 'contextual emergence' offer a robust solution to reproducibility issues in complex and interdisciplinary science.
• This methodology facilitates the creation of specific, yet consistent, reproducibility standards across diverse scientific domains.
• The approach avoids both rigid universalism and relativistic fragmentation, promoting a more integrated scientific landscape.