A Note on the Problem of Statistical Calibration