No way of knowing about the strength of this research, from what you’ve reported. I could see how if you’d asked a questionnaire full of organizational items you could get three factors (which she calls “hubs”, “gatekeepers” and “pulse-takers”) through a factor analysis, each one containing a number of items. It is the responsibility of the researcher to name the factors and interpret their meaning, based on the items significantly loading. So here there would be a significant cross-loading between factors 1 (“hubs”) and 2 (“gatekeepers”), given that gatekeepers are described as more self-aware hubs. However, there would we a lot of interpretive art in this, and that’s fine- I’ve done that, too.
On the other hand, what you’ve reported doesn’t show data, so who knows. I guess it could be what we used to call armchair observation.
The part from the news report that I didn’t include, as to the research approach:
Through confidential surveying and painstaking analysis, Stephenson and her 20-member staff try to determine a company’s hidden makeup by asking simple questions like: Whom do you talk to? Whom do you go to for expert advice, or quick decisions? Which old-timers know where the bodies are stashed? Which employees are creating information “bottlenecks”, bolstering their own power by keeping colleagues in the dark?
When collected and converted into graphic images, the resulting data resemble DNA strands, densely encoded with names, job titles and crosscrossing lines indicating personal connections.
I doubt she’s doing factor analysis with these questions. You need interval level data. You can use a few nominal level as conditions but this looks more like qualitative research, unless she’s counting references to specific employees and having a means to define which of her three conditions they represent. I don’t know about this particular graph method mentioned. A graph presumes numbers, so they must be counting mentions of individuals among all the people they are surveying.
I like qualitative interpretive stuff (like hermeneutics and phenomenology) but it’s a lot more subjective, while factor analysis would be a lot more objective. Or at least as a quantitative empiricist a would see it. Seems more like humanities style research, and less social sciences style.