As global conservation actions become more urgent, informed decision-making requires robust analyses of the costs and benefits of policy options, based on available evidence. Recovery planning for threatened or endangered species must assume a cause-and-effect relationship between proposed management interventions and population responses. However, a significant portion of current knowledge about threatened or endangered species is derived from observational studies because experiments that fully meet random and controlled design criteria are largely infeasible or unethical. Large-scale field experiments are becoming more common, yet the greater uncertainty generated by what remain fundamentally observational studies can lead researchers to weak inferences about causal mechanisms, creating debate and confusion among decision-makers, planners and stakeholders. This has been an acute problem facing conservationists and governments as they struggle with the successful recovery of species in decline. In other domains where experimental evidence is difficult to collect, causal modelling has been adopted to identify causal relationships from observational data, based on a set of strong assumptions and identification rules. In Canada, significant and ongoing efforts have had limited success in reversing the population decline of woodland caribou (Rangifer tarandus caribou). We examine the scientific framework for woodland caribou recovery efforts through the lens of causal modelling, highlighting feasible steps that could be taken to improve the rigour of causal inferences.