Skip to main content

Advertisement

Table 6 Model fitting summaries for meta-analyses containing a single fixed-effects explanatory variable

From: A systematic review of phenotypic responses to between-population outbreeding

Model Fixed effects DIC Vs DIC Δ DIC Range in DIC Study variance component Residual variance component Min. effective samples Max. auto-correlation
1 ~ Intercept −576.1 - - 1.7 0.0200 0.0110 889 0.058
2 ~ Generation −573.4 −576.1 2.7 1.7 0.0204 0.0110 1000 0.027
3 ~ Fitness class −610.9 −576.1 −34.8 1.8 0.0187 0.0102 1000 0.044
4 ~ Trait type −619.4 −576.1 −43.4 1.3 0.0145 0.0099 908 0.032
5 ~ Trait timing −586.6 −576.1 −10.5 0.4 0.0196 0.0107 1000 0.050
6 ~ Taxon category −571.0 −576.1 5.0 1.0 0.0187 0.0111 883 0.061
7 ~ log(physical distance) −467.8 −464.6 −3.2 0.6 0.0246 0.0107 856 0.077
8 ~ Lifespan category −575.6 −576.1 0.5 1.2 0.0194 0.0110 1000 0.028
9 ~ Mating system −574.5 −576.1 1.5 1.2 0.0205 0.0110 1000 0.044
10 ~ Population status −576.7 −576.1 −0.7 1.2 0.0201 0.0109 770 0.041
11 ~ Observation environment −573.6 −576.1 2.4 2.4 0.0178 0.0110 840 0.086
12 ~ Quality score −575.8 −576.1 0.3 3.4 0.0200 0.0109 1000 0.032
  1. DIC (column two) gives the deviance information criterion. “Vs. DIC” gives the DIC for the corresponding model including only an intercept term. “Δ DIC” gives the difference in DIC between columns two and three. Range in DIC gives the range in DIC among three replicate model runs, as an indication of the extent of Monte Carlo error in DIC estimates. Study and residual variance components are variance estimates for study and residual random effects. “Min. effective samples” gives the minimum number of effective samples for the parameters in the model. “Max. auto-correlation” gives the maximum absolute sample autocorrelation at lag 1 for any parameter in the model.