1. Choose a network dataset for which you also have some actor-level attribute data as well. (Go to UCINET help menu, choose Standard Datasets, and look at the descriptions of each. Or use your own data.)
2. Compute the following centrality measures: degree, closeness, betweenness, eigenvector, flow betweenness, information. How are they different? (Helpful to examine network diagram to understand the differences in the measures.)
3. Join them all together into a single dataset in which the rows are individuals and the columns are centrality measures (Data > Join > Columns).
4. Extract the normalized versions of each measure into a new dataset (for measures that don't have a normalized version, just use the raw scores) (Data > Extract)
5. Correlate every measure with every other (Tools > Similarities > Correlation), then run non-metric mds on the correlation matrix.
6. Run Correspondence Analysis on file created in Step 4.
7. For each interval-level actor attribute, correlate each centrality measure with the attribute. For each categorical actor attribute, run analysis of variance use each centrality measure as dependent variable. How does centrality relate to the attributes? Which way does the causality run?