It seems like one problem with surveying people in an organization about their informal networks is concerning the validity of their answers. Will people really tell management who they talk to? Will they admit to how much time they spend getting coffee and chatting? Even if they are talking about the latest bug or getting information about the newest build when they chat, management may call this "wasted time".

One way to bypass this problem of validity is to measure artifacts, rather than surveying people. In the past, artifacts may have included the carbon copies of phone message books and the names on interoffice envelopes. By looking at who was getting calls and who was getting memos, one could determine who was at the center of communication in the organization. Modern artifacts include email server logs, telephone logs, and time-lapsed video cameras of the workspace.

If you are studying a computer system (AKA software testing), you might look at server logs. This could be a helpful artifact during usability testing, performance testing, or functional testing.

If you are studying communication patterns on software teams, like I am, you might look at emails, bug reports, phone logs, chat rooms, or other communication artifacts from the team. You might determine if anyone is left out or is a bottleneck in the process. You might figure out where people congergate so you can determine where to post an information radiator (it's probably where the bagels are.) If you think you know where the common area is, you might look at the number of soda cans and Starbucks cups thrown away in trash cans to make sure you really have the most popular place.