Print

Print


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

In our current research, we are using the variables of structural holes 
and cohesion.  We have two indicators of each.  I have questions about 
how to aggregate each set of indicators and, based on this aggregated 
measure, how to determine cutpoints for creating categories (e.g. 
high/low).

FOR COHESION, we use indicators of egocentric density and mean strength 
of ties (valued ties).
1)  What is the appropriate way to "aggregate" these two indicators for 
cohesion into one measure?  Density is indexed 0 to 1, while mean 
strength of ties has a range of 0-4.
2)  Based upon this aggregated measure, we need to first identify an 
actor's cohesion as low vs. high.  We also need to be able to identify 
the actors' cohesion into three categories:  low, medium, high.  What 
are the appropriate methods for determining 1/2 and 1/3 cutpoints or 
thresholds for cohesion?   Is is appropriate to use the mean or median? 
 Other suggestions would be appreciated.

FOR STRUCTURAL HOLES, we use efficiency and constraint.
1)  What is the appropriate way to "aggregate" these two indicators for 
structural holes into one measure?  I assume we will need to somehow 
reverse code the constraint indicator so that both indicators have a 
positive association with structural holes.  Can we simply take an 
average of the two?

2)  Based upon this aggregated measure, we need to first identify an 
actor's use of structural holes as low vs. high.  We also need to be 
able to identify the actors' cohesion into three categories:  low, 
medium, high.  What are the appropriate methods for determining 1/2 and 
1/3 cutpoints or thresholds for structural holes?   Is is appropriate 
to use the mean or median?  Or, given the indicators are indexed 0 to 
100, is it appropriate to use the half-way point of 50?  Other 
suggestions would be appreciated.

If you have any suggestions for research that has been published that 
might address these issues, I'd appreciate any leads you can provide.
Thanks,
Julie



Julie M. Hite
Brigham Young University
[log in to unmask]

_____________________________________________________________________
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.