Fixed and random effects used in the GLMM (sections A-F) and GLM (section G) analyses of nest predator abundance, nest predation, nest concealment and the level of parental caution during nest visits.

N (sample size) = number of corvid point counts (sections A-B) or number of nests examined (sections C-G).Covariates: recording date (sections A-B), clutch sequence (sections C-D), nestling age and brood size (section G).aAnalysis carried out with all nest predators together (see Tables 2 and 3) and with each nest predator species separately (see Results section).bAnalysis carried out with corvid point counts conducted during the shrike settlement period in 2009 and 2010.cAnalysis carried out with corvid point counts conducted during the shrike settlement and breeding periods in 2010 to examine seasonal change in nest predator abundance.dAs nest concealment was estimated in only part of shrike nests, we used a simpler model structure than in section C to explore how nest concealment influences nest predation in both habitat types.Fixed and random effects used in the GLMM (sections A-F) and GLM (section G) analyses of nest predator abundance, nest predation, nest concealment and the level of parental caution during nest visits.