Several dead time correction methods were compared experimentally with the exact correction method and their limits were discussed. These correction methods were applied to neutron activation analysis of a biological sample. A special electronic circuit and an additional counting equipment were used to obtain the fractional dead time with a suficiently high frequency.