Sent this to SOCNET. Posted here too.I have been chewing on the idea that in a given intra-organizational network the process of knowledge creation is going to be enabled or constrained by the number of distinct, non-overlapping sub groups. Also, that rather than totally unconnected fragments, what is key is that these groups are connected by to each other.
One way to capture this idea, I thought, would be to look at the number of non-overlapping subgroups of strong relations in an advice-seeking relationship. But what is the mechanism of linking? Options include brokers in a different network of strong relationships, brokers in a different network of weak relationships, or brokers in the same network, but using weak relationships. So, if you are a subgroup with a broker, you are more connected (and more likely to spread your knowledge) than one without. A related question than becomes who “counts” as a broker? Is there some good cut-off value? Or maybe use the five individuals with the highest brokerage score?
So, three questions (and thank you paid forward)-
1) Is this similar to some work out there already?
2) Does the clustering coefficient already capture what I am describing as connected subgroups?
3) Does one of the options above sound better to you? (Advice subgroups with brokers in communication, Advice subgroups with brokers in weak communication, or advice subgroups with brokers in weak advice relations)?