A method that allows the determination of the lunar and solar diurnal tidal components from randomly distributed poor geophysical data, together with the corresponding level of significance, is proposed. The comparison of the results obtained separately from the analysis of a long series of continuous data and these provided by a shorter series of data excepted from the former in a random way and processed through the proposed method shows the validity of the latter, which provided a significantM2 term in seismic activity at Hawaii, Phlegrean Fields and Vesuvius.