Africa WEurAsia EEurAsia IPacific NoAmerica SoAmerica |
The 'closest neighbors' in this sample are constructed so that each numerically adjacent pair in the series is as similar as is possible in a linear alignment (Murdock and White 1969). This allows any variable to be correlated with its neighbor's score in the same variable as a measure of the extent to which sampling variance is not random, but regionally clustered. The higher this coefficient, the more the statistical significance of correlations with this variable need to be deflated. When you draw lines from society 1 (starting in South Africa) to 2 and so on through each successive number (ending in Tierra del Fuego at the tip of South America), you are traversing a path of similar cultures. |