Peto J
Br J Prev Soc Med. 1974 Nov;28(4):241-5. doi: 10.1136/jech.28.4.241.
It is suggested that cigarette consumption can and should be reduced by tax increases. As this contradicts a recent communication, the differences of analysis and the underlying models are discussed. Straightforward curve-fitting to serial social or economic data can be very misleading. If there is random drift, it is better to analyse consecutive changes of the parameters rather than their actual values.
有人认为,可以而且应该通过提高税收来减少香烟消费。由于这与最近的一份报告相矛盾,因此讨论了分析差异和潜在模型。直接对系列社会或经济数据进行曲线拟合可能会产生很大的误导性。如果存在随机漂移,最好分析参数的连续变化而不是其实际值。