A test for seasonality of events with a variable population at risk

Abstract

A statistical significance test to detect seasonality of epidemiological events is described. The method is similar to that of Edwards, but makes it possible to allow for an arbitrary pattern of variation in the population risk, and also for the unequal lengths of time sectors of a cycyle of seasons (e.g., months of a year). From the test it is possible to estimate the amplitude of seasonal variation and the time at which the maximum occurs in a postulated simple harmonic fluctuation; the adequacy of the description of the data by a curve of this kind may be evaluated using a goodness-of-fit test. A numerical example of the calculations is given using some anencephalus data, and the results are compared with those of alternative tests.