1987 年 29 巻 1 号 p. 48-59
In the previous studies, using Pearson's product moment correlation coefficient, we found that in most cases of cancers, statistically significant positive correlations were observed between natural background radiation exposure rate and crude cancer mortality rate over the period 1950-1978. Furthermore, we found that the statistical significance of correlation between natural background radiation exposure rate and the age-adjusted cancer mortality rate in the same period mostly disappeared. We studied the cause of this apparent correlation and found that the prefecture with a higher natural background radiation exposure rate had a greater component ratio of older people. In Japan, a number of prefectures with a higher natural background exposure rate are located in relatively thinly populated districts which have been experiencing an outflow of the younger generation to more highly industrialized and urbanized areas. Therefore, statistically significant positive correlations were observed for almost all cancers between natural background radiation exposure rate and crude cancer mortality rate. In the present investigation, we statistically tested the frequency distributions of natural background radiation exposure rate and age-adjusted cancer mortality rate, and calculated Spearman's rank correlation coefficient between natural background radiation exposure rate and the age-adjusted cancer mortality rate. The frequency distribution of the natural background radiation exposure rate and that of the age-adjusted mortality rate appeared normal in most cases of cancer, and the statistical significance of correlation between natural background exposure rate and the age-adjusted cancer mortality rate did not differ much on the whole, even though we used Spearman's rank correlation coefficient between them.
