Recent advances in instrumentation, digital computer technology and mathematical theory promote the error analysis of Earth-tide data. Various statistical techniques developed and used in other fields are applicable in the study of Earth tides, and the accuracy of the Earth's rigidity constants determined from the tides will be greatly improved with the help of these techniques. The fundamentals of the statistical techniques of autocorrelation, crosscorrelation, convolution, statistical means, bandpass filtering, correlation coefficients, power spectra, coherency and equalization are described, and their principal applications in the Earth-tide analysis summarized. Examples of effective application of these techniques in the elimination of the errors in the tidal data such as those introduced from instrumental drift, phase differences between the observed and predicted tides, etc. are discussed. This work is an attempt to introduce statistical analysis into the Earth-tide study.