I am not sure how a local version of the measure can be implemented, but could the unit of analysis be the number of three paths where the focal node is either node 2 or 3 (in other words, in the middle). The numerator would be the number of these paths that are “closed” (part of a 4-cycle), and the denominator could be the total number of these paths. What do you think?

Tore

]]>I have not implemented a local version of Robin and Alexander’s (2004) coefficient. The global one is implemented in the reinforcement_tm-function. Without looking up the paper again, is a local coefficient implemented?

As mentioned on Clustering in Two-mode Networks (https://toreopsahl.com/tnet/two-mode-networks/clustering/), Robin and Alexander’s coefficient is similar to reinforcement of a one-mode tie between two primary nodes. Could you project a binary two-mode network and then study the link weights to answer your question?

Best,

Tore