2.5. Statistical analyses
Three-way ANOVAS with a split-plot design were carried out to
test the main and interactive effects of block, plant diversity, and
grazing by large herbivores on the plant Shannon–Wiener index
and insect abundance in experimental year. General Linear Model
(GLM) with a Tukey test was used to examine the statistical
difference in the mean values of the treatments and specific
comparisons between different groups of treatments were
performed using LSMEANS statement of this procedure. Effects
of blocks were tested together with the treatments in all the above
analyses, but they were not discussed in this study due to
insignificant effects on measured variables. Data were further
analyzed the effects of grazing breed using one-way ANOVA within
each plant diversity level. The relationships between plant
Shannon–Wiener index and the abundance of total insects and