Skip to main content

Table 3 Investigating the performance of the null hypothesis randomisation test. Sets of artificial data sets were created using four different approaches

From: Explaining the geographic spread of emerging epidemics: a framework for comparing viral phylogenies and environmental landscape data

 

Path model

Approach for creating artificial data sets:

straight-line

least-cost

random walk

 (i) branch durations randomly permutated among branches

6 %

3 %

3 %

 (ii) branch durations equal to spatial distance

0 %

0 %

5 %

 (iii) branch durations equal to environmental “weight”

100 %

100 %

100 %

 (iv) branch durations equal to a value between (ii) and (iii)

95 %

100 %

100 %

  1. In each case, artificial data sets were applied to the “elevation” raster treated as a resistance factor. Values equal the percentage of 100 artificial data sets for which the p-value of the null hypothesis test was < 0.05