Walter S D, Elwood J M
Br J Prev Soc Med. 1975 Mar;29(1):18-21. doi: 10.1136/jech.29.1.18.
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.
本文描述了一种用于检测流行病学事件季节性的统计显著性检验方法。该方法与爱德华兹的方法类似，但能够考虑人群风险的任意变化模式，以及季节周期中不同时间段的不等长情况（例如一年中的月份）。通过该检验，可以估计季节性变化的幅度以及假设的简单谐波波动中最大值出现的时间；可以使用拟合优度检验来评估这种曲线对数据描述的充分性。文中给出了使用无脑儿数据进行计算的数值示例，并将结果与其他检验方法的结果进行了比较。
