Direct Factor Analysis of Sociometric Data

Peer influence groups: identifying dense clusters in large networks

“…First, I apply the computationally feasible measures to the large simulated networks. Second, I use data on two smaller networks, a portion of the MacRea-Gagnon (MacRea, 1960) prison network (g = 39) and a private Northeastern high school (g = 790) to compare with other group detection methods.…”

“…I provide two examples. First, I use data taken from MacRea (1960), based on work of Gagnon, on friendships in a prison. This example provides a small network that is not overly clustered and thus allows me to compare a non-obvious case between three different methods.…”

Inferring network structure from cascades

“…The network consists of 67 prisoners(nodes) which have 182 friendships (edges) [22]. Success rates are reported in Table II.…”

The Economics of Social Networks

“…The citation network consists of the network of citations among all papers that have cited Milgram's (1967) paper or have the phrase "small worlds"in the title, and is from Gar…eld [?]. The prison data record friendships among inmates in a study by MacRae [122], the ham radio data record interactions between ham radio operators from Killworth and Bernard (1976), and the high school romance data collected romantic relationships between high school students over a period of a year and a half in a US high school and is from Bearman, Moody, and Stovel [12]. The number of nodes, average degree, and clustering numbers are as reported by the studies.…”

“…With probability p have this link be part of the graph, and with probability 1 p have that link be absent from the graph. This random graph formation process was explored in detail by Erdös and Rényi (1959, 1960, 1961 and has been studied extensively since then (see Bollobás (2001). 25 There are a number of interesting properties that such Bernoulli networks have.…”