Social Networking Assignment - Subgroups and Block Models

The purpose of this exercise is to learn how to use UCINET to identify cliques and conduct a structural equivalence blockmodel.

Monastery network: In 1968, Samuel Sampson wrote a doctoral dissertation about 18 men living in a New England monastery who were training to become monks. The network data for this assignment come from the fourth time that Sampson asked the men to choose three others whom they liked most. (Two men decided to name four others; even monks don't always follow instructions!)

Sampson identified four factions, that is, subgroups of novice monks: (1) The Loyal Opposition, a group of five conservatives who had entered the monastery earlier; (2) the Young Turks, a group of seven who entered later and questioned many monastery practices, which the Loyal Opposition defended; (3) three Interstitials, who wavered between the two sides of the conflict; and (4) three Outcasts, who were not accepted by anyone else. The names of the men and the factions to which they belong are listed on the next page.

Near the end of Sampson's research, a "crisis in the cloister" erupted, resulting in the expulsion of four monks (Ambrose, Bonaven, Elias, Simp), because they were "not in conformity with the spirit of the order." Five others almost immediately left voluntarily (Albert, Amand, Basil, Romul, Victor). In the end, only Hugh, John, Louis, and Peter remained at the monastery. Could a graph of the liking network, also using the monks' betweenness centralities, a clique analysis, and a blockmodel give insights into those occurrences?

Datasets:

Monastery_Like4 is an undirected UCINET data file that was symmetrized using the maximum criterion (i.e., if one monk named a second monk, they are assumed to like one another).

Monastery_Attr is a UCINET attribute file with each monk classified into one of the four factions identified by Sampson: (1) Loyal Opposition; (2) Young Turks; (3) Interstitials; (4) Outcasts; and their normalized betweenness centrality scores. The table on the next page displays the contents of the attribute file.

Perform these Steps with UCINET and NetDraw:

1. In UCINET, use "Match Multiple Datasets" to match the Monastery_Like4 and the Monastery_Attr files, putting the default suffix -Matched at the end of both output files.

In NetDraw, create the graph of Monastery_Like4-Matched. You may need to move a few points slightly so some lines and labels don't squash together. Remove the arrowheads from the graph. Open the Monastery_Attr-Matched file. Change the colors of the nodes to represent Sampson's four factions. Change the sizes of the nodes to represent the actors' betweenness centrality scores. Save this color graph, as a screen capture or as a .jpg file, for your report.

2. In UCINET, use "Network/Subgroups/Cliques" to find all cliques with three or more members in the Monastery_Like4 network. Instruct UCINET to save the "(Output) Co- membership matrix:" (with the default name "CliqueOverlap") for further analysis. From the log output, copy and save for your report (a) the list of cliques found; (b); the Actor- by-Actor Clique Co-membership Matrix; and (c) the Hierarchical Clustering of Overlap Matrix.

In UCINET, match the CliqueOverlap network with the Monastery_Attr file, putting the default suffix -Matched at the end of both output files.

In NetDraw, open CliqueOverlap-Matched to graph the ties among monks who belonged to the same cliques (this dataset is identical to the "Actor-by- Actor Clique Co- membership Matrix" in the output log). You may need to move a few points slightly so some lines and labels don't squash together. Remove the arrowheads from the graph. Open the Monastery_Attr-Matched file. As in Step #1, change the colors of the nodes to represent Sampson's four factions (use the same colors you used in Step #1).

Change the sizes of the nodes to represent the actors' betweenness centrality scores. Save this color graph, as a screen capture or as a .jpg file, for your report.

3. In UCINET, conduct a blockmodel analysis of Monastery_Like4, using "Network/Roles & Positions/Structural/Concor/Standard" and specifying "Max depth of splits (not blocks):

= 2" to obtain a four-block solution. From the output log, copy and save the Blocked Matrix and the Density Matrix for your report. Use "Network/Cohesion/Density/Density Overall" to find the mean density of Monastery_Like4 for your report.

Write a brief report (print color graphs and tables, with single-spaced text) describing the main findings of your analyses; include the following:

1. Display the color graph of the monastery liking network in which you modified the nodes to indicate the novice monks' attributes. Describe what you observe. In particular, comment on the locations of members of the Loyal Opposition, the Young Turks, and others in the graph, the connections within and between faction members, and any other noteworthy features such as the actors' betweenness centralities. To what extent does this graph reveal the opposing political factions identified by Sampson?

2. Display the list of cliques found, the Actor-by-Actor Clique Co-membership Matrix, the Hierarchical Clustering of Overlap Matrix, and the color graph of the CliqueOverlap network. (You may cut-and-past from UCINET output, but use Courier font to keep columns aligned!) Describe the major findings of this analysis, commenting on whether your subgroup analysis identifies the monastery's opposing political factions as described by Sampson. (Suggestion: Color-code the members' names in the list of cliques, using the same four colors you use in the graph.)

3. From the blockmodel analysis, display the 4-block matrix and the density matrix. Using the mean network density, manually create the image matrix for the network, where 1 means block density is ≥ network mean and 0 means block density < network mean. Describe the blockmodel's major features, such as the within- and between-block ties and whether the monks' block memberships are consistent with Sampson's factions.

4. How well or poorly do the network color graph, clique analysis, and blockmodel analysis together help to explain and understand the expulsions and departures of the novice monks during the "crisis in the cloister?"

Attachment:- Assignment.rar