Skip to main content
. 2015 Aug 15;3:19. doi: 10.1186/s40462-015-0046-5

Table 1.

AICc model selection results for path analytic models examining factors affecting flight duration across the ocean and along the coast

Path model equations K AICc ΔAICc W Cumulative W Fisher’s C
tailwind ~ stage + age 16 71.64 0 0.56 0.56 22.09
flight duration ~ tailwind + tailwind2
tailwind ~ stage + age 17 72.37 0.73 0.39 0.95 17.97
flight duration ~ age + tailwind + tailwind2
tailwind ~ stage + age 18 76.69 5.05 0.04 0.99 17.11
flight duration ~ stage + age + tailwind + tailwind2
tailwind ~ stage + age 19 79.00 7.36 0.01 1 13.86
flight duration ~ stage + age + tailwind + tailwind2 + crosswind
tailwind ~ stage + age 20 83.79 12.15 0 1 12.68
crosswind ~ age
flight duration ~ stage + age + tailwind + tailwind2 + crosswind
tailwind ~ stage + age 21 86.73 15.09 0 1 9.19
crosswind ~ stage + age
flight duration ~ stage + age + tailwind + tailwind2 + crosswind
tailwind ~ stage + age 14 120.04 48.40 0 1 79.31

k represents the number of parameters in each path model. W and Cumulative W represent Akaike weights and cumulative model weights, respectively. Fisher’s C statistic = −2*ln(model likelihood). The null model is defined by the top model with paths to flight duration removed. 2 indicates curvilinear parameter coefficient