As the most widely distributed freshwater fish worldwide, common carp Cyprinus carpio can be either invasive or ‘naturalised’ in most areas of introduction. This leads to different levels of perception regarding its role in freshwater ecosystems, with experimental research focusing either on its ‘middle-out’ impacts or overall function in limnological processes. At the same time, the large scales at which carp dynamics operate may severely limit the validity of laboratory and, oftentimes, field experiments in extrapolating results to real-world ecosystems. In this study, 129 laboratory, field and ‘natural’ experiments were systematically reviewed through causal criteria analysis, and within an historical/biogeographical and risk assessment context. Of the 19 countries where experiments were conducted, only four were considered as ‘low risk’ and one as ‘no risk’, the other being ‘medium’ to ‘high risk’. Experimental findings from 373 component-wise assessments supported the framework of effects on water quality, vegetation, invertebrates and vertebrates, with the latter including also amphibians and waterfowl, previously unreported. Stronger evidence was provided by natural and field relative to laboratory experiments, reflecting the reductionism of the latter. Critical biomass for an impact was highly dependent on experimental set-up, even though the overall threshold of ≈200 kg ha−1 under natural conditions supported recent findings. Management of carp should reflect the level of current and potential risk posed by the species in its different areas of distribution, thereby accounting for projections of further spread but also for unsuccessful colonisation. Future experimentation should favour a holistic→reductionist over a reductionist→holistic approach.