With respect to the management of dredged sediments, a crucial issue is whether the removed materials (watered and/or processed) are disposed of or reused in an environmentally sound manner. In this context, the number of studies dealing with hazard or risk assessment has exponentially increased. This emphasis has resulted in the promotion and application of a very large variety of ecotoxicological tests. As a consequence, there is a clear need to highlight the scope and limitations of these tests for their appropriate selection and interpretation. In this paper we discuss the choice, implementation and interpretation of laboratory tests carried out on aquatic organisms at various levels of biological organization. We examine some experimental tools and methods in order to determine how suitable they are in regard to the objectives for which they are employed. To make this discussion more consistent our paper is based on results from research programmes conducted for governmental organizations and industrial partners. The Laboratoire des Sciences de l'Environnement (LSE) was involved with Cemagref de Lyon in a first methodological programme on risk assessment of scenarios of dredged sediment deposition. Another programme for a chemical company was implemented to assess the benefits of a physico-chemical treatment applied to contaminated dredged sediments in a scenario of reuse or deposit in gravel pits. Currently, the LSE is working on a programme of risk assessment for road sediments in valorization scenarios. From these programmes, we expose how single-species tests, as well as more complex bioassays and microcosm tests, can be used in an iterative step of risk assessment. Concerning microcosm tests, we also introduce a more realistic system that has been designed to simulate natural hydraulic conditions of gravel pits to assess the effects of toxicants on gravel pit aquatic biota during the sediment immersion phase and the sediment post-deposition phase (paper in preparation). The benefits of these ecotoxicological approaches are underlined, but limits are discussed with regard to several criteria: ecological relevance, realism, use for decision-making, cost and complexity of methods involved.