The idea of statistical significance is meaningless unless some random phenomenon is involved. In many instances in delinquency research, however, there is no obvious source of randomness. William W. Wattenberg and James Balistrieri justify their use of chi-square tests with the statement that these tests "establish the degree of statistical reliability with which the null hypothesis could be rejected." Some statisticians would extend the logic of drawing a sample from a finite, existing population to assuming that a sample drawn has come by a random process from an infinite hypothetical universe. Properly conducted, the replications are vital in the task of inductive inference, the formulation and testing of general propositions from particular sets of data. The use of statistical tests of signficance was considered justified as 1952 is a sample of years and San Diego a sample of cities which might have been used.