Re: Question on Paired T-Test of Degree Centrality Across Two Stages
|
Hi Yvonne, this is a difficult question that a number of us have been discussing offline. To me, the big issue is that fundamentally, you are asking if nodes increased their degree. This means, you are asking whether density increased between T1 and T2. Ie., there are simply more ties. So, if the reason for this is not artifactual, I would approach that via ? Snijders, T. A., & Borgatti, S. P. (1999). Non-parametric standard errors and tests for network statistics.?Connections,?22(2), 161-170. ? This is a method of comparing densities. It is implemented in UCINET. ? As you point out, a t-test of any kind would seem inappropriate because of non-independence. If node A increases degree by 1, then so does the node that A connected to. ? I am sending you privately specific comments by the Davids ¨C David Dekker and David Krackhardt ? steve ? ?
toggle quoted message
Show quoted text
From: [email protected] <[email protected]> On Behalf Of eating.hk11 via groups.io
Sent: Tuesday, September 3, 2024 11:03
To: [email protected]
Subject: [ucinet] Question on Paired T-Test of Degree Centrality Across Two Stages? I hope this email finds you well. I am working on a hypothesis that "the degree centrality of all nodes will change significantly during the T1 or T2 stages of the NGO's mention network." I have already measured centrality data for all nodes in both stages. While a paired t-test would be the obvious approach if the data were independent, I am not sure how to account for the dependencies inherent in network data within UCINET. So far, I have two possible approaches in mind: 1) Import the centrality data from the two stages into SPSS and conduct a paired t-test.
2) Use UCINET's Tools > Test Hypotheses > Node-Level > T-Test function. I would greatly appreciate any suggestions or advice on how best to proceed with this analysis. Thank you so much for your time and assistance in advance.
|
