***** To join INSNA, visit http://www.insna.org ***** I'm just a PhD student, I would advocate for the open number question: - the number of degree is a key parameter, and the distribution is very uneven; - you should always compare things that are similar, however arbitrary, to understand what your bias are. The main case I've seen this discussion is for sexual relation network; however painful, the tedious discussions with the most active subjects proved decisive to tackle STD. Maybe forcing the revelation of a certain number of contact is a good way to measure non-symetric relations, or triggering memories --- but I wouldn't use it as a roof, not with relevant hubs around. More precisely: - the relevance of degree distribution and hubs to your study, - directional ties, - statistics vs. interpretation all favor an unbouded criterion; in comparision to: - the key-role of the ends, and a constraint on degree - non-directional ties - interpretative analysis prefer a given number of ties -type of study. I'd love to have any reference though, because I'm supposed to work on a "What are your three main partner" type of survey, and I don't really nkow how to deal with it. 2006/10/23, Timothy Huerta <[log in to unmask]>: > ***** To join INSNA, visit http://www.insna.org ***** > > Hi all! > > Got a question. While most network surveys I've seen tend to be in the > "here's a list of all objects within a boundary condition, please indicate > your relationship with any or all of them", I've also seen the "list the top > XX number of people who meet this boundary criteria and then, please > indicate your relationship with any or all of them". > > I see a number of pros and cons to each of the forms. The former is more > complete and less likely to have missing or underreported results. The > latter has a reduced responder burden and may be useful when the number of > possible objects in the network is very large. The latter is actually a > series of ego networks that are overlapped and fused into a network map. The > former may result in overreporting of results which one might use reciprocal > ties for, but in doing so losing data on directionality. > > So, where does the SNA community stand on this issue? Are there studies that > look at the issue? > > Tim > > _____________________________________________________________________ > 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. > -- Bertil Hatt 224 rue de Rivoli - 75001 Paris --- 01 42 61 06 92 -- 06 85 49 78 99 Mon agenda : www.tinyurl.com/no774 ou www.google.com/calendar _____________________________________________________________________ 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.