The bounded effect concentration as an alternative to the NOEC

Abstract This paper criticizes the No Observed Effect Concentration (NOEC) defined as the highest concentration for which the difference from the control group is not statistically significant. It is shown by example and theoretically that NOECs defined in such a way will often be located at concentrations where effects of considerable magnitude occur. Moreover, large and precise experiments lead to lower NOECs than smaller and imprecise experiments. This means, the less one knows, the higher the ‘safe’ value. This seems an illogical pattern. A two-step procedure is proposed that involves finding the concentration where the effect might be 25% at the most, followed by linear extrapolation to a concentration where the effect is acceptably small, e.g. 10%. The resulting statistic is called a 10% Bounded Effect Concentration (BEC10). The BEC10 lies below the EC10, the difference between the two becoming smaller if the experiment becomes larger. The 10% Bounded Effect Concentration can therefore safely be assumed to cause at the most 10% effect to the species being tested, under laboratory test conditions. Therefore, it provides a better starting point for risk analyses.