Results were expressed as percent of control in order to reduce statistical variance and to allow comparison across treatments and trials.
Statistical comparisons of variance and mean values for all three groups were made using one-way ANOVA and post hoc Tukey test by Statistical Package for the Social Sciences (SPSS, Version 21, IBM co).. Paired t tests were also used to compare the distances between right and left side, and the maxillary occlusal cants to other measurements.
For statistical comparison, when the variances were equal in Levene's test for homogeneity of variance, one-way analysis of variance (ANOVA) followed by post hoc least significant difference (LSD) test was used; otherwise, Dunnett's T3 test was applied.
A statistical comparison of these variance estimates is then made and a P value obtained.
Statistical comparison was performed by variance analysis, followed by the Student-Newman-Keuls test, using the GraphPad Prism software.
Logarithmic transformation of the mean to enable statistical comparisons gave greater variance stability.
For statistical comparisons, analysis of variance (ANOVA) was used with the Bonferonni corrections (Instat for IBM, version 2.0; GraphPad Software, San Diego, CA, USA).
However, as there was only one sample from the uninvaded period from the coastal Erebus Cove core, statistical comparisons of compositional variance in the uninvaded and Olearia zones are very weak and non-significant (PERMANOVA: F1,25 = 1.47, P = 0.18 and multivariate generalized linear models Wald statistic = 7.38, P = 0.22).
Additional tests were performed that enabled statistical comparison of item error variances and the factor covariance to be identical across gender.
This analysis should be used primarily as a statistical reference for comparison of within plot variance vs. the site means.
Statistical group comparisons (repeated measures analyses of variance (rANOVAs) with treatment duration as covariate).
