On a more precise calculation of the electric conductivity in the photosphere and in sunspots