The analysis of organic pollutants in environmental water samples requires a pre-concentration step. Pre-concentration techniques such as stir bar sorptive extraction (SBSE) have gained popularity since they minimise the use of toxic organic solvents and can be considered as green analytical techniques. Similar to other pre-concentration techniques, one of the problems when SBSE is used is the matrix effect, which often occurs during the analysis of environmental water samples such as estuarine or wastewater samples. The present work studied the matrix effect during SBSE coupled to in-tube derivatisation-thermal desorption (TD)-gas chromatography-mass spectrometry for the determination of several endocrine disruptor compounds, such as alkylphenols, bisphenol A, estrogens and sterols, in environmental water samples, after optimisation of the major variables affecting the determination. Variables such as the addition of methanol or an inert salt to the donor phase, the extraction temperature, the volume of the donor phase, the stirring rate and the extraction time were studied during the SBSE optimisation. In the case of the in-tube derivatisation and TD step, the volume of the derivatisation reagent (N,O-bis(trimethylsilyl)triufloroacetamide with 1% of trimethylchlorosilane (BSTFA + 1% TMCS)) and the cryo-focusing temperature were fixed (2 mu L and -50 A degrees C, respectively) according to a consensus between maximum signal and optimal operation conditions. Good apparent recovery values (78-124%) were obtained for most of the analytes in Milli-Q water, except for 4-tert-octylphenol (4-tOP), which showed apparent recovery values exceeding 100%. Precision (n = 4) was in the 2-27%, and method detection limits were in the low nanogrammes per litre level for most of the analytes studied. The matrix effect was studied using two different approaches. On the one hand, Milli-Q water samples were spiked with humic acids, and apparent recovery values were studied with and without correction with the corresponding deuterated analogue. On the other hand, estuarine water and wastewater samples were spiked with known concentrations of target analytes, and apparent recoveries were studied as explained above. In general, the matrix effect could be corrected with the use of deuterated analogues, except for 4-tOP and nonylphenols for which [H-2(4)]-n-nonylphenol did not provide good corrections.