My name is Anna and I am using software Ucinet 6 to analyze network constructed of group participant’s Liking scores of each other participant. I would like to calculate betweenness centrality of each participant in my group. My question concern SYMMETRIZATION of the adjacency matrix that i have. My data say who likes whom. So they give a DIRECTED network. Is it OK to symmetrize the matrix? Ucinet 6 gives such a possibility. I wonder if such symmetrization does not produce an untrue, false, twisted picture of the relationships in the group?
Does anyone know WHEN and UNDER WHAT CONDITIONS one can symmetrize a network? And what method of this symmetrization (through replacement with maximum, minimum or average value of the non-symmetrical ties) “harms” my data the least?
I am looking forward to your help.