@article{jmmss 847, author = {Dawn Iacobucci, Rebecca McBride, Deidre L. Popovich, Maria Rouziou}, title = {In Social Network Analysis, Which Centrality Index Should I Use?: Theoretical Differences and Empirical Similarities among Top Centralities}, volume = {8}, year = {2018}, url = {http://journals.librarypublishing.arizona.edu/jmmss/article/id/847/}, issue = {2}, doi = {10.2458/v8i2.22991}, abstract = {<p>This research examines four frequently used centrality indices—degree, closeness, betweenness, and eigenvectors—to understand the extent to which their clear theoretical distinctions are reflected in differences in empirical performance. Even for stylized networks in which one centrality index may seem more relevant than the others, the four indices are frequently highly correlated. This result can be interpreted as good news: it does not diminish the conceptual distinctions, yet it suggests the indices are rather robust, yielding similar information about actors’ positions in networks, which can be reassuring given their widespread use by applied network analysts who may not appreciate the theoretically distinct origins and definitions. This research also compares computational speed across the centrality indices as another practical element that may help determine the choice of centrality index.</p>}, month = {8}, pages = {72-99}, keywords = {centrality,degree,closeness,betweenness,eigenvector centrality,social networks}, issn = {2159-7855}, publisher={University of Arizona Libraries}, journal = {Journal of Methods and Measurement in the Social Sciences} }