Table 4. Binomial logistic mixed effect model results for the
analysis of latitude x seed type on germination rates at the two common
gardens (Hazel Dell and Horse Rock Ridge). Coefficient estimates are
expressed as log-odds values, using contrasts =
c(”contr.sum”,”contr.poly”) to compare each group against the grand
mean, where the intercept is the grand mean and others are deviations
from that. For random effects: σ2 = residual variance
of logistic regression model; τ00 = between-subject
variance partitioned to each random effect in the model; ICC =
intraclass correlation coefficient; N = number of levels to each random
effect. OLRE = observation level random effect. Marginal
R2 provides the variance explained by the fixed
effect(s) only while Conditional R2 provides the
variance explained by both the fixed and random effects, as calculated
in Nakagawa et al. (2017).