Effect-directed analysis (EDA) has been frequently and successfully used to identify key toxicants in sediment extracts. However, by disregarding bioavailability this approach may lead to a biased prioritisation of fractions and toxicants with respect to hazards and risks. To overcome this problem the present EDA of sediment components from the Bílina river (Most Czech Republic), that inhibit growth of the green algae Scenedesmus vacuolatus, applies a novel partition-based dosing technique to prioritize and identify major toxic fractions and compounds in comparison to conventional solvent dosing. The novel dosing technique is based on partitioning from loaded silicone rods to the aqueous phase similar to partition processes that determine exposure in native sediment-water systems. In the present study the application of partition-based dosing had a big influence suggesting polar compounds such as triclosan as key toxicants while polycyclic aromatic hydrocarbon (PAH) fractions did not exhibit significant effects. In contrast, conventional dosing prioritized mainly PAHs in agreement with previous studies. For both approaches individual toxicants could be confirmed quantitatively based on the index of confirmation quality (ICQ), which compares the effect of fractions and artificial mixtures of identified and quantified toxicants over the full range of effect levels.