top of page

The Master Network

The Master Network, shown in Figure 1, graphs all the interactions of the Salem Witch Trials based on the 1692-93 court records.  In these documents, any clear or implied interaction is counted as a relationship.  For example, when Mercy Lewis accuses Elizabeth Procter of witchcraft, there is a relationship between the two.  The relationship may have occurred prior to 1692, but it also might not have since analysis of the documents suggests not all accused suspects were known to their accusers.  At the very least, the relationship occurred in 1692.  The other classification of relationships are the implied ones.  The social nature of the trials and legal proceedings created relationships through the process of accusations.  In this network, people who played in active role in the proceedings while in the same room are connected.  For example, William Wormwood and Elizabeth Hubbard both actively participated in the examination of George Burroughs through Wormwood's testimony and Hubbard's accusation.  Since both people were definitively in the same room at the same time to play an active role in the examination, a network formed that matters in the explanation of the witch trials.    The people who charged neighbors, and even strangers, with witchcraft created recognizable networks as well as those who defended their family members and friends.

Master Network data-

Nodes: 1,465

Edges: 9,929

Average degrees: 13.52

Maximum degrees: 260 (Hathorne, John)

Average Geodesic Distance: 3.27

Maximum Geodesic Distance: 7

Average Betweenness Centrality: 1619.48

Maximum Betweenness Centrality: 120131.327 (Bradbury, Mary)

Average Eigenvector Centrality: .001

Maximum Eigenvector Centrality: .013 (Hathorne, John)

Number of Groups by Cluster: 22

Figure 2 shows the graph with the nodes and edges at the lowest proportions to emphasize the range of edges in the network.  While the nodes and some edges are too small to study in this image, the major pathways between groups appear with much more ease in this image.  This allows a more in-depth discussion of intergroup relationships during the witch trials. 

Figure 3 shows the groups of the network without the intergroup edges.  This view portrays the details of each clusters isolated from the rest of the graph.  There are 22 groups in the Master Network, each of which take on their own shapes.  The first 9 of 22 groups are numbered here to designate names.  The groups formed with various themes based on chronology, geography, and case file as denoted by their names:

1. The Advancing Group

2. The 1693 Trials

3. Mary Bradbury's Case

4. The Andover in October 1692

5. The Topsfield Group without Wildes

6. The Procter and Nurse Group

7. The Lynn and Jurors Group

8. Dorcas Hoar's Case

9. Roger Toothaker's Coroner's Jury

10-22. Deposition based groups

bottom of page