Table 4.
Cluster analysis summary, centrality measures.
Note: This table summarizes the results of our clustering analysis. Bias is on the y-axis of the plots and percent missing is on the x-axis. Each case (network, measure, missing data type) was placed into a cluster based on the pattern of bias across different levels of missing data. We then summarized what types of networks, measures and missing data went into each cluster. Note that this table only includes results for the strong positive and strong missing data types. A positive correlation means central nodes are more likely to be missing.