Sources of error in nuclear instrument calibration are analyzed in the article. Transformation functions are derived for several specific applications. Correction methods are determined to be more accurate than the least-squares method for identification of objects, processes, and calibration of instruments. It is recommended that calibration of measuring instruments be performed according to the minimum of the sum of the squared deviations, taking into account the form of the dependence and the measurement errors of both variables being compared. 2 refs., 3 figs.