In stock assessment modeling, the process of testing the sensitivity of model results in relation to errors and uncertainties in the input parameters. For example, a virtual population analysis (VPA) might be used to determine the fishing mortality rates over several years. The results might be conditioned on an assumed annual natural mortality rate (M) of 0. 2. The sensitivity of this assumption might be examined by redoing the VPA based upon a different M, perhaps M=0. 15 to 0. 3. From the sensitivity analysis one can determine the importance of particular parameters to the overall scientific advice.