A new method of stock assessment, dynamic response analysis, allows a qualitative assessment of stock status relative to its level of maximum net productivity to be carried out with minimal data. This paper evaluates the performance of dynamic response analysis on a simulated population under variable conditions with uncertain data. In the no-harvest case the data consist simply of a temporal sequence of relative population sizes. Dynamic response analysis is most sensitive to the number and precision of the population estimates and least sensitive to environmental variability and the intrinsic population growth rate. Significance levels must be chosen carefully, since some combinations of parameters and error levels result in an unacceptably low proportion of correct assessments. Dynamic response analysis can be a useful stock assessment technique for the management of marine mammals, but attention must be paid to the quantity and quality of the data.