Controlled flexibility for agronomic and natural-science research embedded in Living Labs safeguarding scientific rigor under adaptive conditions

CONTEXT Living Labs (LLs) bring agricultural research into real-world settings where researchers and practitioners co-create, test and adapt solutions as conditions change. This flexibility is valuable, but creates a less visible problem: if reference conditions (comparators), sampling, measurements or research questions change during the process, the scientific claim may no longer be supported in the same way. OBJECTIVE We propose controlled flexibility as a framework for keeping evidence scientifically interpretable while preserving the adaptiveness and co-creation that make LLs useful. METHODS This Perspective focuses on quantitative, experimental and structured observational research embedded within LLs. It distinguishes two questions that are often conflated: did the LL generate useful learning or change, and can the resulting data support a specific scientific claim?, thus differentiating between “evaluation for impact” and “evidence for publication”. Systems thinking helps match the measurements to the claimed level of change. RESULTS AND CONCLUSIONS The framework separates a protected evidential core (research question and claim, analytical unit, comparator or case-selection logic, core indicators, sampling rules, metadata, analysis plan and model of change) from elements that may be adapted locally. Decision logs, deviation records, shared scientific rationale and data audits make adaptations visible. If a change affects the evidence supporting the claim, a new claim version is defined before further analysis. SIGNIFICANCE Controlled flexibility offers a practical way to remain adaptive without weakening scientific credibility. It helps LLs produce evidence that is transparent, proportionate to the claim and publishable, without sacrificing co-creation and situated learning.