Fig. 2From: Using a Bayesian approach to reconstruct graph statistics after edge samplingDifference between the expected (using Eq. 2) and true number of nodes removed in the edge sampling process for two networks of the same power-law degree distribution but different degree assortativities. The mean is shown over 1000 sampled graphs for each value of p with standard deviation error bars (one deviation above and one below.)Back to article page