Print

Print


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

Thanks for the replies - they are summarized below:

Tore Opsahl et al's study of an online community contains ways to measure network properties for valued networks - see http://toreopsahl.com/2009/03/06/article-patterns-and-dynamics-of-users-behaviour-and-interaction-network-analysis-of-an-online-community/ - also, his tnet R-package contains measurement methods for valued, directional data

Koka and Prescott's (2002) article on social capital - only had a quick glance at this one though.

I was recommended to check out the software at Elbirt Technologies - http://www.elbirttechnologies.com/index.html - and I tried using the Descriptives & Percentiles software, but it sort of hung. I was also recommended a software tool called Galileo, but I couldn't locate that one.

Finally, I got the suggestion to calculate the degree of asymmetry over time as well as the variance on the symmetric/asymmetric part as percentage of the total variance.

Though I sort of gave up on my analysis as it didn't really seem to produce any interesting results. The data was yearly global flows of wheat (quantities) - my hypothesis is/was that the structure of the network of global wheat trade would change as the global price of wheat has fluctuated. I did try measuring FlowBetweenness, checking how these attributional measures changed over the years, but the results didn't show any relation. I also average-symmetrized the data - b(i,j) = (a(i,j)+a(j,i))/2 - and calculated some other centrality measures, and I also tried some preliminary 6-iteration REGE-algorithms on some periodical data (3-year moving averages). If anyone wants to have a go at the data, just drop me a line.

Yours,
Carl


Dr Carl Nordlund
carl.nordlund(at)hek.lu.se
Human Ecology Division, Lund university
www.hek.lu.se
________________________________
Från: Social Networks Discussion Forum [[log in to unmask]] för Carl Nordlund [[log in to unmask]]
Skickat: den 21 januari 2011 09:41
Till: [log in to unmask]
Ämne: [SOCNET] Network-wide measures for valued and directional data

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

I'm looking for suggestions on how to measure network-wide properties for valued, directional data. What I got is relational data (valued and directional, that is) between a fixed set of actors for 15 different time-periods, and what I would like to do is to compare how the relational structure has evolved over time.

And yes, "network-wide properties" is quite a vague term, but it's intentional as I'm pretty much open for any ideas! My initial idea was to calculate actors' flow betweenness (which seems suitable for valued directional data) and then use the mean values as a network-wide measure of flow betweenness. Although I know of several network-wide measures of centrality, they typically are best applied to dichotomous symmetric data, and although I could transform my data to this, I rather not as any theoretical motivation for such a dichotomization would be somewhat straggling (it's trade flow data!). Symmetrizing the data should be okay though (from a conceptual point of view)

Though I'm pretty much open to any suggestion on network-wide measures, not just centrality-type ones. I am thinking about using Kansky's Pi-index (originally used to measure the "elongatedness" of transportation networks) but I doubt this has been implemented. I am also thinking about conducting role-analyses (REGE) on each set in the time-series and compare strata movement etc.

Yours,
Carl

Dr Carl Nordlund
carl.nordlund(at)hek.lu.se
Human Ecology Division, Lund university
www.hek.lu.se
_____________________________________________________________________ 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.