Print

Print


*****  To join INSNA, visit http://www.insna.org  *****

Dear all,

 

I am working on a case study of gang member presence in drug distribution networks in Montreal. I completed a series of centrality analyses, but, after reading de Nooy, Mrvar, and Batagelj's book on Pajek, I decided that a triad census analysis would the most appropriate way to finish the case study. 

 

I need some advice on my results (see below). The presence of forbidden triads in the de Nooy, Mrvar, and Batagelj's examples were minimal. My network is different in that it has a very high presence of one of the forbidden triads (11-201) and a high presence of two of the other forbidden triads. At the same time, there is also a higher presence of other triads that lead me to conclude that the best fit is the hierarchical clusters model. 

 

How do I proceed? Does such a high presence of forbidden triads suggest that a balanced-theory framework and triad census should be avoided or should I proceed beyond the forbidden triads and toward a discussion of the hierarchical clusters model? 

 

I hope that I've been clear in presenting the problem - any help will be greatly appreciated...

 

Carlo 

 

 

Results:

 

   Type    Number of triads (ni)        Expected (ei)             (ni-ei)/ei     Model

------------------------------------------------------------------------------------------------------

   3 - 102                 80523              4509.61                  16.86     Balance

  16 - 300                  2959                 0.01              369376.13     Balance

------------------------------------------------------------------------------------------------------

   1 - 003                740189            651161.19                   0.14     Clusterability

------------------------------------------------------------------------------------------------------

   4 - 021D                  148              4509.61                  -0.97     Ranked Clusters

   5 - 021U                  222              4509.61                  -0.95     Ranked Clusters

   9 - 030T                   31               433.34                  -0.93     Ranked Clusters

  12 - 120D                  161                10.41                  14.47     Ranked Clusters

  13 - 120U                  122                10.41                  10.72     Ranked Clusters

------------------------------------------------------------------------------------------------------

   2 - 012                 31589            187717.24                  -0.83     Transitivity

------------------------------------------------------------------------------------------------------

  14 - 120C                   86                20.82                   3.13     Hierarchical Clusters

  15 - 210                   878                 1.00                 876.67     Hierarchical Clusters

------------------------------------------------------------------------------------------------------

   6 - 021C                  352              9019.21                  -0.96     Forbidden

   7 - 111D                 1596               433.34                   2.68     Forbidden

   8 - 111U                 1487               433.34                   2.43     Forbidden

  10 - 030C                    5               144.45                  -0.97     Forbidden

  11 - 201                  2576                10.41                 246.44     Forbidden

------------------------------------------------------------------------------------------------------

 Chi-Square: 1095831831.7789***

 2 cells (12.50%) have expected frequencies less than 5.

 The minimum expected cell frequency is 0.01.

 

 

 

 

 


_____________________________________________________________________
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.insna.org). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.