The data of 30-year investigations of macroscopic fluctuations are generalized in this review. The results of measuring processes of different nature (from biochemical reactions to radioactive decay) fluctuate. They are presented by certain discrete values. The discreteness of corresponding spectra of states can be probably explained by both pure mathematical causes and cosmophysical factors. The fluctuation amplitude and the shape of the spectrum of the states of objects with different nature depend on the universal external cosmophysical factors. General problems arising in connection with the analysis of the "scatter in the results" and "nonreproducibility of measurements" are discussed.