Print

Print


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

Using standard General Linear Model tests with network data is inferentially risky because the most fundamental assumption - independence of observations - is explicitly undermined by the use of a network perspective. If observations are independent, then there is no network; if observations are dependent, standard inferential tests are not valid. So, in that sense, a standard statistical test used with network data is a contradiction.  There are plenty of network statistical methods and models that can be applied.

Professor Garry Robins
Melbourne School of Psychological Sciences
University of Melbourne
http://www.psych.unimelb.edu.au/people/garry-robins

Melnet website: http://www.sna.unimelb.edu.au/

Check out our new book on ERGMs: Lusher, D., Koskinen, J., & Robins, G. (2012). Exponential random graph models for social networks: Theory, methods and applications. Cambridge University Press. (http://www.cambridge.org/us/knowledge/isbn/item6897868/?site_locale=en_US)
Look inside the book: (http://www.amazon.co.uk/Exponential-Random-Models-Social-Networks/dp/0521141389#)

From: Social Networks Discussion Forum [mailto:[log in to unmask]] On Behalf Of kamal badar
Sent: Tuesday, 20 November 2012 11:33 PM
To: [log in to unmask]
Subject: Implications/limitations of applying inferential statistics to co-authorship network data

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

According to Hanneman & Riddle (2005) " Social network analysts rarely use samples in their work. Most commonly, network analysts identify a population and conduct a census of that population. The boundaries are those imposed by the researcher or even created by the actors themselves. Social network studies, therefore often draw the boundaries around a
population that is known, a priori, to be a network" (Page 5).

Talking about co-authorship networks, we collect bibliometric data from databases according to boundaries imposed (geographic location of researchers, disciplinary fields, journals within the fields or individual institutions or departments ect). If the co-authorship network understudy is considered a population, what implications/limitations can we have while applying inferential statistics (e.g. correlation and OLS regression) to a specific phenomenon (for e.g. examining the association of centrality and academic performance)? Doesn't the exercise of inferential statistics provide types of estimates of population parameters and characteristics based on a sample of that population not the population itself? How can we defend if we do indeed apply inferential statistics to co-authorship network?

Hoping to get important insights from the experts.

Regards
Kamal Badar
Doctoral Student
Asian Insitute of Technology
Thailand.
_____________________________________________________________________ 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]<mailto:[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.