Data Analysis
A three stage balanced nested design was used to integrate the indicators measured at
different scales, and among sites (Figure 2-2). Hypothesis testing for differences between means
was accomplished by using two-sample t-test with an alpha of 0.05 and a two-tailed confidence
interval. The sampling of nine distinct reference locations produced a dataset where the
assumptions for analysis of variance (ANOVA) was not ensured; therefore, non-parametric tests
were used to detect any significant differences among the reference sites and among the distinct
forest age class segments (SAS, 2002).
Trends over time and between variables were obtained from linear regression using the
general linear model (PROC GLM) (Yang et al. 2006; SAS, 2002). Plant species indicator
analysis (IndVal) was used to measure the level of relationship between a given plant species to
categorical units such as pine flat subtypes or forest age classes. It calculates the indicator value
d of species as the product of the relative frequency and relative average abundance in each
categorical cluster. Indicator species analysis is used to attribute species to particular
environmental conditions based on the abundance and occurrence of that species within the
selected group. A species that is a "perfect indicator" is consistent to a particular group without
fail. Indicator values range from 0 to 100, with 100 being a perfect indicator score. Because
indicator species analysis is a statistical inference, a test of significance is applied to determine if
species are significant indicators of the groups with which they are associated (Dufrene and
Legendre, 1997). This is achieved by the Monte Carlo permutation test procedure (1000
iterations), where the significance of a P-value is determined by the number of random runs
greater than or equal to the inferred value (oc=0.10). Accuracy is defined from the binomial 95%
confidence interval (Strauss, 1982).