Table 2
Although there is a significant correlation between most of the shape attributes and the two measures of the regression line, the percent variability “explained” by these relationships is generally fairly low, with all four attributes together accounting for only 4.1% and 20.2% of the variability in intercepts and slopes respectively