Skip to main content
. 2019 Feb 6;10:35. doi: 10.3389/fmicb.2019.00035

Table 2.

Model selection of PERMANOVA according to fledgling sampling unit using a weighted UniFrac distance matrix (a) and an unweighted Unifrac distance matrix (b) and according to age class (adults vs. fledgling) for a weighted UniFrac distance matrix (c) and an unweighted Unifrac distance matrix (d).

Model rank Fledgling sampling unit Age class d.f. logLik AICc Δ AICc AICc ω R2
a. Dataset with fledgling samples only; weighted UniFrac
1 + 9 –69.994 180.2 0.00 1.000 0.14
2 1 –124.672 257.3 77.11 <0.001 0.00
b. Dataset with fledgling samples only; unweighted UniFrac
1 + 9 –349.406 739.1 0.00 1.000 0.19
2 1 –426.972 861.9 138.89 <0.001 0.00
c. Dataset with fledgling and adult samples from 2015 only; weighted UniFrac
1 + 2 117.497 –222.9 0.00 1.000 0.26
2 1 93.184 –178.3 46.58 <0.001 0.00
d. Dataset with fledgling and adult samples from 2015 only; unweighted UniFrac
1 + 2 58.705 –105.3 0.00 1.000 0.24
2 1 35.970 –67.9 43.42 <0.001 0.00

The parameters included in the model are indicated by “+” and degrees of freedom (d.f.), log likelihood (logLik), Akaike Information Criterion for small sample sizes (AICc), delta AICc (Δ AICc), AIC weight (AICc ω), and R2 are shown. Models are ranked according to model support (from the smallest AICc to the largest).