Structural micro forces in online social networking websites: Impact on friendship structure

There is already a modest body of research on structural properties of online social networking websites. However, there seems to be a big gap when it comes to important questions related to underlying network forces that create that structure. Drawing from empirically grounded theories of off-line communication networks, hypotheses about the biases (forces) creating the friendship structure in educational online social networking websites are set up, for dyadic and triadic levels of analysis. Hypothesized micro forces are then empirically investigated by building probabilistic models from the data of an online educational social network, using recent reformulations of a biased net modeling approach. In addition, magnitudes of forces are investigated longitudinally, based on five observations of a network. Results indicate the existence of all the hypothesized micro structural forces. Reciprocity is by far the strongest force acting in the network, followed by transitivity, similarity and lastly, closure. Reciprocity is shown to increase with the time, whereas transitivity and similarity decrease. Closure bias first decreases, and then steadily increases.