[SNA & Statistics] How to Test Network-Based Hypotheses Correctly
- Related Articles
[SNA & Statistics] Why Standard Statistics Fail on Network Data
So how can we avoid biased results caused by network autocorrelation?
The key is to choose a statistical approach based on what your hypothesis is about and what form your data takes. In network analysis, this usually means distinguishing between two cases: analyzing relationships themselves, and analyzing attributes of nodes.