To account for potential differences in bioavailability (and toxicity) due to different soil organic matter (OM) contents in natural and artificial soil (AS), in the current European environmental risk assessment (ERA) a correction factor (CF) of 2 is applied to toxicity endpoints for so called lipophilic pesticides (i.e. log Kow > 2) generated from laboratory tests with soil invertebrates. However, the appropriateness of a single CF is questioned. To improve the accuracy of ERA, this study investigated the influence of soil OM content on the toxicity to the earthworm Eisenia andrei of five active substances used in pesticides covering a wide range of lipophilicity. Laboratory toxicity tests were performed in AS containing 10 %, 5 % and 2.5 % peat, and a natural LUFA 2.2 soil (4.5 % OM), assessing effects on survival, biomass change and reproduction. Pesticide toxicity differed significantly between soils. For all pesticides, toxicity values (LC50, EC50) strongly correlated with soil OM content in AS (r2 > 0.82), with toxicity decreasing with increasing OM content. Obtained regression equations were used to calculate the toxicity at OM contents of 10.0 % and 5.0 %. Model-estimated toxicity between these soils differed by factors of 1.9–3.6, and 2.1–3.2 for LC50 and EC50 values, respectively. No clear relationships between pesticide lipophilicity and toxicity-OM relationships were observed: the toxicity of non-lipophilic and lipophilic pesticides was influenced by OM content in a similar manner. The results suggest that the CF of 2 may not be appropriate as it is based on incorrect assumptions regarding the relationships between lipophilicity, OM content and toxicity. Further research should be conducted to understand the mechanistic link between toxicity and soil OM content to better define more chemically and ecologically appropriate CFs for ERA.