The minority of papers only is devoted to the study of impact of the uncertainties in nuclear data on the nuclear concentration received during the solution of the problem of fuel burn-up in the reactor facility. On the other hand, uncertainties of known reaction rates, neutron fluxes, etc. can lead to considerable distortions of the results obtained therefore it is important to be able to assess the impact of such uncertainties on nuclear concentration of nuclides. In this paper we consider the problem of the impact assessment of uncertainties in nuclear data on reactor functionalities as applied to isotope kinetics which represents the well-known Cauchy linear problem. The most exact approach is the statistical approach of the randomized selection of input parameters in using different distribution laws. But the simplest method of the analysis of sensitivity of model to perturbation parameters is the following (it has several names in the literature: one-at-a-time sensitivity measures, 1% sensitivity method): by varying one of the input parameters of the task for the small amount (for example, for 1%) when other parameters are constant, the corresponding response of output parameters is defined (variation approach). Our results show that in burn-up calculations the mean square deviations of nuclear concentrations obtained using statistical approach coincide with the variations of nuclear concentrations obtained in the variation approach.