Since AMBI was published originally in 2000, it has been used in an increasing number of investigations with monitoring purposes, or to analyse impacts on soft-bottom macrobenthic communities. Some guidelines for its correct use were published in 2005; however, a main issue remained without an answer — which are the minimal area and number of replicates necessary, to obtain a precise estimate for AMBI? In this study, new methodologies such as bootstrap techniques have been applied to this particular problem. Data were obtained from sampling carried out in 1995, within the framework of the Littoral Water Quality Monitoring and Control Network of the Basque Country (northern Spain). The sampling strategy consisted of 11 intertidal estuarine sampling stations (0.25m 2, sampled for each of six replicates) and 17 subtidal estuarine and coastal sampling stations (0.125m 2, sampled for each of six replicates). Two replicates have been established as being sufficient, both for intertidal and subtidal sampling stations, to classify 80% of the pseudosamples into the same disturbance level, in terms of AMBI, for 64% of the stations. For the minimal area, it has been determined also (for both intertidal and subtidal sampling stations) that 0.25m 2 is sufficient to classify 80% of the iterations into the same disturbance level, for 64% of the stations.