25.01.2015 Views

CONNECTIONS - INSNA

CONNECTIONS - INSNA

CONNECTIONS - INSNA

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>CONNECTIONS</strong><br />

Multiplicity and Redundancy<br />

The last section of Table 5 summarizes the<br />

precise configurations of ties. Recall in our<br />

discussion of Figure 4, there were potentially<br />

eight types of ties. However, when aggregated,<br />

only seven distinct patterns of social affiliations<br />

are represented– corporate only, university only<br />

and social club only (note there were no ties that<br />

were only based on shared affiliations through<br />

museums since all of those ties were structurally<br />

redundant). Table 5 shows that only a quarter of<br />

the ties were from corporate ties alone. While<br />

fewer than 3 percent of the ties were formed<br />

solely through shared university ties, over a<br />

quarter (26.28 percent) of the ties were based<br />

exclusively on mutual social club memberships.<br />

The bottom section of Table 5 also shows that<br />

over a third of the ties between these 20<br />

directors were constituted by both corporate and<br />

social club ties. Equally important from the<br />

perspective of emphasizing the role of noncorporate<br />

ties in integrating our corporate<br />

directors, 7 ties were based on two or more<br />

noncorporate ties. Finally, Table 5 indicates that<br />

6 of the 137 ties were from three or more<br />

sources, including at least one corporate tie. In<br />

short, the results in Table 5 underscore the<br />

message from Figure 4 – non-corporate ties<br />

constitute a substantial portion of the social<br />

connections between our example directors, with<br />

social club ties being most abundant.<br />

Thus far, the discussion has focused on how to<br />

represent the multiplicity of ties between the<br />

directors in this example network. However, it<br />

would also be useful to further aggregate these<br />

various configurations of ties back to the<br />

individual directors. To accomplish this<br />

aggregation, the analyses revert to the original<br />

set of 207 ties. However, rather than<br />

aggregating on the basis of the tie ID itself (1.02,<br />

1.16 or 1.08), this time we need to aggregate by<br />

the vertex number (making sure that both<br />

sending and receiving vertices are represented).<br />

Therefore, by shifting the level of analysis back<br />

to the directors themselves, each director will<br />

possess a set of ties that can be characterized by<br />

being completely or predominantly corporate.<br />

In contrast, a director’s set of ties can be<br />

primarily grounded in other organizational<br />

affiliations. Consider again John A. Barr. This<br />

time however, Table 6 presents all of his 33 ties<br />

with the 19 other directors.<br />

By sorting the ties by type, it is straightforward<br />

to calculate both the sum of strengths which<br />

reflects the 33 individual ties, and the sum of the<br />

strengths for each type of tie. 3 Table 6 shows<br />

that of the 33 ties, 9 are based on shared<br />

corporate board memberships, 2 on mutual<br />

membership on university Boards of Regents or<br />

Trustees, and 22 on shared social club<br />

memberships. This means that 27.3 percent of<br />

John Barr’s ties were corporate, 6 percent were<br />

university and 66.7 percent were social club<br />

based. When this procedure is applied to the<br />

remaining 19 directors, we can then obtain<br />

summary measures of the configuration of ties<br />

possessed by each director. Table 7 presents the<br />

summary statistics for the 20 directors in this<br />

example.<br />

Table 7. Configuration of Ties for Directors<br />

Average Proportion of Corporate Ties 0.4990<br />

Average Proportion of Museum Ties 0.0215<br />

Average Proportion of University Ties 0.0540<br />

Average Proportion of Social Club Ties 0.4255<br />

Therefore, on average, the twenty directors are<br />

characterized by half of their ties being<br />

corporate based, and nearly another half are<br />

based on social club ties. Finally, Table 7<br />

indicates that the average percent of ties based<br />

on museum ties is only 2.15 percent and the<br />

average percent of ties based on shared<br />

university ties is 5.4 percent. In other words,<br />

this example of 20 directors is tied together by<br />

both corporate and club ties in roughly equal<br />

proportions. While this conclusion is far from<br />

3 Calculating the degree of each director on the basis<br />

of each of the four affiliation matrices would yield<br />

the number of different directors tied to John Barr<br />

and as a consequence, we would lose the multiple<br />

corporate ties between this director and David<br />

Kennedy, and the multiple club ties between John<br />

Barr and G.A. Freeman Jr., Paul W. Goodrich, R.S.<br />

Ingersoll, David Kennedy, H.J. Livingston, Brooks<br />

McCormick, James F. Oats Jr., and J. Harris Ward.<br />

18

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!