The establishment of causal links between chemical contamination and observed ecotoxic effects of environmental samples is a major challenge in ecotoxicology. Effect-directed analysis is an approach combining fractionation with toxicity testing before chemical nontarget analysis to separate and subsequently identify toxic compounds in environmental samples. A final confirmation step in this procedure evaluates how much of the observed toxicity in the environmental sample can be attributed to the mixture of identified toxicants. The aim of the present study is to advance the methodology for the confirmation of mixtures in effect-directed analysis by using tools for the assessment of mixture toxicity accounting for unknown modes of action and heterogeneity of concentration-response curves. For this purpose, toxicants were identified in sediment extracts from two different sites. All identified compounds were tested both individually as pure compounds and in mixtures at ratios equal to those found in the sediment extracts. The observed extract toxicity was then compared with the expected combined effects calculated according to the models of concentration addition and independent action as well as with the observed toxicity of the synthetic mixture. Drawbacks of the established approach using toxic unit summation are demonstrated, and the Index of Confirmation Quality, an easy-to-read representation that allows a quantitative measure of confirmation over a range of different effect levels, is introduced.