02072nas a2200181   4500000000100000000000100001008004100002260001500043100002400058700001700082700002700099700002000126245007500146856004600221300002200267520158700289022001401876       2025                            d  c2025-12-141 aSarah E. Wolfensohn1 aVera Baumans1 aF. Josef van der Staay1 aSaskia S. Arndt00aImplications of stress for animal welfare and animal-based research†  uhttps://doi.org/10.1177/00236772251396310  a002367722513963103 aBetter animal welfare is essential for better animal-based science. Poor welfare induces stress (and vice versa), which in turn can affect or even severely confound research results. To ensure the validity of scientific results, those who work with laboratory animals must take responsibility for their care and welfare. A harm–benefit analysis can be used to weigh the welfare and quality of life of animals in scientific studies against the resulting benefits to other animals or humans, provided that animal welfare can be validly assessed. This review considers the identification and characterisation of stress by physiological, hormonal, immunological and behavioural measures and by assessment of the physical condition of the animal. It addresses controllability, predictability, chronicity, duration and severity, and the intrinsic and extrinsic factors that modulate animal behaviour and coping mechanisms in response to stress. EU Directive 2010/63 requires procedures to be classified as mild, moderate or severe. However, as long as some researchers use purely subjective assessments without a supporting structure and reference scales, or no assessment method at all, these terms may have limited practical value. The challenge is to understand the state of the animal from the information available. Welfare assessment tools can be used to demonstrate the true impact of research procedures and their refinement to protect animal welfare. If any doubt exists about the harm–benefit evaluation of the experiment then the welfare of the animal should take priority.  a0023-6772