The statistical analysis of acoustic backscatter samples recorded by multibeam echosounders can be a valuable tool for remote seafloor characterization and interpretation. The present paper aims at analyzing the statistics of backscatter data values, both in "raw" status and after various averaging operations, using field data. It is shown that the statistics of the data can be adequately described by a Weibull distribution parametrized by the incidence angle and the level of applied processing: the distribution of the averaged backscatter amplitude, processed according to various schemes, varies from a Rayleigh law for raw data to lognormal and finally to Gaussian distribution after successive averaging operations. Based on these results, some recommendations for the calculation of the mean backscatter strength are presented. Finally, the influence of high-amplitude scatterers in the backscatter probability density function is addressed; a scheme is suggested to separate the contributions of the substrate from the contributions of the scatterers on the statistical distribution of sonar data samples.