Print

Print


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

Qian

Yes we reported the apparent robustness of certain centrality measures under conditions of missing data in:

Costenbader, E., & Valente T.W. (2003). The stability of centrality measures when networks are sampled.  Social Networks, 25, 283-307.

And in-degree centrality is quite robust to missing data.

-Tom

 

From: Social Networks Discussion Forum [mailto:[log in to unmask]] On Behalf Of Qian Hu
Sent: Monday, February 11, 2013 7:56 PM
To: [log in to unmask]
Subject: Missing network data

 

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

Dear SOCNET members,

I have two questions about the missing network data. 

what can be done if we cannot get an ideal response rate (say less than 70 percent, or even lower)?

  1. I remember somewhere in the literature that says some centrality measures such as the indegree centrality, are relatively robust for network data that captures over 50% of the data. Am I correct?
  2. The existing research seems to suggest two methods to deal with missing data: exponential random graph models, and imputation methods. If we know the key actors are captured in the imperfect data set. Can we leave them as they are without using any additional modeling process? Any literature to support this argument?

 

 

 

Thank you very much. 

Qian Hu

University of Central Florida 

_____________________________________________________________________ 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.

_____________________________________________________________________ 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.