Fig. 5

The plots show the mean widths of 95% confidence intervals obtained from bootstrapped sub-samples of a network. The mean widths of density and transitivity increase with lower sample size, which indicates increasing uncertainty around the point estimate of the network metrics. However, the pattern is reversed for mean strength and diameter because the values for these two metrics are directly affected by the number of nodes present in the network. Therefore, we consider scaled versions of these two metrics where the number of nodes at each level scales the values