Prior to statistical analysis, skewed variables were transformed
to approach normality and to reduce the influence of a few large
values, using the angular transformation for proportional data and
the logarithmic transformation for continuous variables (Table 1).