Informal influence

Post date: Nov 13, 2015 9:34:20 PM

If you ask "who do you go to for advice" or "who is influential here", then you can tabulate indegree (the number of incoming ties) to get a sense of informal influence. The only problem is, everybody's vote is counted equally. Wouldn't it be better if we could take into account the voters' own levels of influence? There is a measure in UCINET that will do that, called Bonacich Power or Beta Centrality. Just go to Networks|Centrality & Power|Bonacich Power (Beta Centrality). Fill out the form like this:

(where pv504 is the name of the network dataset). The result will be a score for each person that indicates the extent to which that person is named by people who are themselves often named. You could then use Excel to plot centrality (Y axis) against organization rank or tenure (X axis), like this:

Note that informal influence generally increases with increasing formal position. But there are some exceptions, and these are the interesting ones.