Print

Print


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

Assuming i am doing a network analysis on nominations among a class of 20 students.
If every student nominates 19 others, my maximum number of ties will be 20 x 19.
But in my dataset students nominated varying number of their peers - ie, some 8, some 5, some 10, some 19.

I have set up my dataset in spss as follows:
student name ==== nominee===helps with assignment
e.g.,
01 === 02 ====1
01 === 03 ====1
01 === 04 ====0
01 === 05 ====1
02 === 01 ====1
02 === 03 ====0

When I am doing my analysis (such as Anova or crosstabluation) in Spss, I focus on only valid ties - that is those that have a 1, code the 0s as missing cases. So respondent 01 for example will only have 3 valid cases since her 01 - 04 tie is set to missing. However under QAP regression in UCINET, they are counted as valid responses, and the total number of valid cases is 20x19.

For a complete network with directed ties, should regression and other multivariate analysis be based on the total number of possible ties, i.e, N(N-1) or the total number of actual ties - which may be less than but not greater than N(N-1)?  Is there any substantive validation for one versus the other?

have a lovely weekend ....

cheers - PintoN


---------------------------------
Yahoo! Mail Mobile
 Take Yahoo! Mail with you! Check email on your mobile phone.

_____________________________________________________________________
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.