Thus, to find out which variable had the strongest association with laying date, we ran linear mixed models in Proc Mixed, SAS 9.3, with laying date in a given year as response variable and mean or maximum temperature as predictor (since including both will cause intercollinearity due to their significant correlation).
Analysis used generalised linear mixed models in R with a two-level hierarchical structure children, nested within school communities.
To address these complications, we used a novel multi-step process with zero-inflated generalized additive mixed models in a Bayesian framework.
Using adjusted mixed models, in high-risk offspring, there was no significant difference in cortisol levels across repeated sampling years (CAR: p = 0.45, logAUC: p = 0.40, logT3: p = 0.66).
For data analysis, we emphasize proper implementation of mixed models, in terms of both distributional assumptions and specification of fixed and random effects to explicitly recognize multilevel data architecture.
Using adjusted mixed models, in high-risk offspring with a lifetime mood disorder (n = 26/53), clinical characteristics including lifetime psychotic symptoms, comorbid SUD, or parent lithium response did not significantly predict baseline CAR, logAUC, or logT3 cortisol (all p > 0.05) (Additional file 1: Table S3).
Hence, these models are developed using either linear or non-linear mixed models in which site, stand, competition variables but also regional units or geographic coordinates are used as covariates (Lappi 1997; Eerikäinen 2003; Calama and Montero 2004; Mehtätalo 2004; Nanos et al. 2004; Hökkä, 1997; Schmidt et al. 2011).
One procedure used to deal with correlated observations is to fit mixed models, in which the variability between the sampling units can be explained by including random parameters, which are estimated at the same time as the fixed parameters (Lappi 1997; Calama and Montero 2004).
For growth, there must also be a within-species standard-deviation, called the residual, which we assumed to be constant for all species (the same assumption is used is mixed models in lme4); there is no residual in a survival model because the binomial distribution defines the variation.
All logistic mixed effects models were fitted using the lmer function for fitting linear or generalized linear mixed models (in the lme4 package for R [81]).
We attempted to analyse between-location and between-species differences in nestling infection (infected vs. uninfected) by using Generalized Mixed Models, in which the nest was included as a random factor and species as a fixed factor.
