Sinha T, Benedict R
School of Business, Bond University, Gold Coast, Queensland, Australia.
Cancer Lett. 1996 Feb 6;99(2):225-31. doi: 10.1016/0304-3835(95)04063-3.
We investigate the relationship between latitude and incidence of melanoma among Whites in different age groups using data from different continents. Relevant data for Whites were obtained for 59 regions around the world. A statistical analysis was carried out using regional dummy variables to eliminate spurious statistical correlation due to clustering. Simple correlation between latitude and incidence of melanoma is strongly negative for almost all age groups. However, once the regional dummies were introduced in the analysis, the relation between latitude and incidence rates disappeared for all age groups but the explanatory power of the regression equation increased substantially.
我们使用来自不同大陆的数据，研究不同年龄组白人中纬度与黑色素瘤发病率之间的关系。获取了全球59个地区白人的相关数据。使用区域虚拟变量进行统计分析，以消除由于聚类导致的虚假统计相关性。几乎所有年龄组中，纬度与黑色素瘤发病率之间的简单相关性均为强负相关。然而，一旦在分析中引入区域虚拟变量，除一个年龄组外，所有年龄组中纬度与发病率之间的关系均消失，但回归方程的解释力大幅提高。
