The establishment of reliable values for concentrations of 16 priority pollutant polycyclic aromatic hydrocarbons (PAH) in a suite of marine sediment reference materials included an examination of the methods used to determine such compounds. Results from the five techniques used indicate large method biases. The biases arise from chemical interferences in methods which use non-selective measuring systems with separation techniques which do not first completely resolve the analytes; e.g., single-parameter optical detectors with high performance liquid chromatographic (HPLC) techniques or a flame ionisation detector (FID) with capillary column gas chromatographic (GC) methods. Measurement by a mass spectrometer with HPLC and GC removes much of the method bias. Results for a representative sediment are discussed to illustrate these conclusions.