Organic total extracts play an important role in soil and sediment risk assessment. Beside a routine application in analytical chemistry, they are used in bio-analytical investigations as a "worst-case scenario" or, e.g., in order to simulate chronic intoxication, and as samples for effect-directed analysis. While theoretically providing highly reliable data and good reproducibility, the whole process of sample handling and extract preparation can lead to extracts that might fail to accurately represent a toxic potential of their corresponding sampling site. This review identifies and discusses the most important possible alterations that have the potential to lead to over and, more often, underestimation of the effectiveness of extracts. Since incorrect data will compromise soil and sediment risk assessment as a whole, results from analytical and bio-analytical investigations of extracts demand cautious interpretation. Reliability of extract testing grows with reproducibility; experiments should therefore be repeated with independent extraction replicates. New or optimized extraction procedures should circumvent the issues mentioned here while being suitable for routine application.