Skip to main content

Table 3 Summary of statistical results (and associated Test numbers as identified in the Results section) for analyses of bryozoan hosts and effects of infections on hosts and patterns of covert infections. Statistical results presented for generalised linear modelling (GLM) and generalised linear mixed modelling (GLMM) are presented as Likelihood Ratio Test (LRT) results with the test statistic as D

From: Persistence, impacts and environmental drivers of covert infections in invertebrate hosts

Focus Test Statistical test (error distribution) Test statistica df P Statistical analysis
Bryozoan hosts A1 GLM (Quasi-Poisson) 80.683 1 < 0.001 Live colony size varies over time in all three rivers
A2 GLM (Binomial) 498.68 2 < 0.001 Proportion of statoblast-producing colonies varies amongst rivers
Covert infections B1 GLM (Binomial) 82.826 2 < 0.001 Proportions vary between rivers
B2 GLM (Binomial) 22.486 3 < 0.001 Proportions vary over time within rivers
B3 GLM (Binomial) 40.248 6 < 0.001 Proportions vary amongst roots within rivers
B4 GLMM (Binomial) Random structure: River/Root/Trip 9.777 1 0.002 Less likely in larger, statoblast-producing colonies
B5 GLMM (Binomial) Random structure: River/Root/Trip 4.740 1 0.029 More likely in conditions of low productivity, low temperature and high dissolved oxygen in small colonies not producing statoblasts
B6 GLMM (Gaussian) Random structure: River/Root/Trip 8.194 1 0.004 High infection severity more likely in conditions of low productivity, low temperature and high dissolved oxygen
B7 GLM (Gaussian) 5.444 1 0.007 Infection severity likely to be reduced in statoblast-producing colonies
B8 GLM (Gaussian) 2.539 1 0.042 Infection intensity likely to be reduced in statoblast-producing colonies
B9 GLM (Gaussian) 0.434 1 0.401 Infection intensity not influenced by colony size
  1. aThe test statistics for Chi-square tests and GLMMs with a binomial error distribution are χ 2 values