The size and density of graphs interact powerfully and subtly with other gr
aph-level indices (GLIs), thereby complicating their interpretation. Here w
e examine these interactions by plotting changes in the distributions of se
veral popular graph measures across graphs of varying sizes and densities.
We provide a generalized framework for hypothesis testing as a means of con
trolling for size and density effects, and apply this method to several wel
l-known sets of social network data; implications of our findings for metho
dology and substantive theory are discussed. (C) 1999 Elsevier Science B.V.
All rights reserved.