***** To join INSNA, visit http://www.sfu.ca/~insna/ ***** Dear colleagues, The formula proposed by Rich Rothenberg has published earlier: Broese van Groenou, M.I., van Sonderen, F.L.P., & Ormel, J. (1990). Test-retest of personal network delineation. In C.P.M. Knipscheer & T.C. Antonucci (Eds.), Social network research: Substantive issues and methodological questions (pp. 121-136). Lisse: Swets & Zeitlinger. The exact formula (page 124) is: % overlap = [(a/b + a/c)/2]*100 where a=overlap T1-T2; b= number identified T1; c= number identified T2 However, in a study by Van der Zouwen & Van Tilburg (2001) we adopted this formula in our first analyses and were not satisfied with the preliminary results. Therefore, we analysed three variables: "Based on the data concerning the links, the joint network size at two observations could be broken down into three parts: the number of network members identified at both observations ('stable'), the number of members not identified at the first observation ('gained'), and the number of members identified at the first observation, but not identified at the second observation ('lost'). The T1 and T2 network sizes and the numbers of stable, of lost, and of gained network members are of course not independently measured. Therefore, in the analyses three variables to be explained will be used: the network size at T2, the proportion of T1 members not identified at T2 ('lost members'), and the number of T2 members not identified at T1 ('gained members'), as proportion of the T1 network size. The T2 network size correlated -.20 with the proportion of lost members and .14 with the proportion of gained network members; the latter two correlated -.37." (page 43) van der Zouwen, J., & van Tilburg, T.G. (2001). Reactivity in panel studies and its consequences for testing causal hypotheses. Sociological Methods & Research, 30, 35-56. best wishes, Theo van Tilburg ------------------------------------------------------------------------ Associate Professor Social Gerontology Vrije Universiteit Amsterdam Dept. of Sociology & Social Gerontology De Boelelaan 1081c, NL-1081 HV Amsterdam, The Netherlands phone +31-20-4446870, fax +31-20-4446810 Internet: http://home.scw.vu.nl/~tilburg/ Email: [log in to unmask] - - - - - - - - - - - - - - Original Message - - - - - - - - - - - - - - On Tue Feb 18 15:42:40 2003, "Richard Rothenberg" <[log in to unmask]> wrote: >***** To join INSNA, visit http://www.sfu.ca/~insna/ ***** > >Skye: > >At the simplest level, you want to know if the same people stay in the >network over time. A problem in calculating stability is that network >size may also change. A measure proposed by Morgan is to use the >proportion of persons who are present at time one and time two, divided >by all the different people present in the network at both times. We >used this to look at the stability of drug using networks (low) and >sexual networks (higher) in the same population, and accompanied it with >description of the network characteristics (frequency of cycles etc.). > This is a lot more primitive than what you may be looking for, but it >is a beginning. > >Morgan DL, Neal MB, Carder P: The stability of core and peripheral >networks over time. Social Networks 1995;17:273-297 > >Rothenberg RB, Potterat JJ, Woodhouse DE et al. Social network dynamics >and HIV transmission. AIDS 1998;12:1529-1536 > >Rich Rothenberg > >Skye Bender-deMoll wrote: > >> ***** To join INSNA, visit http://www.sfu.ca/~insna/ ***** >> >> hi folks, >> I'm working with some "dynamic" network data on classroom >> interactions collected by Dan McFarland. One of the things we are >> interested in looking at is some measure of "stability" or "repeating >> patterns" in interaction networks over time. >> >> There are of course all sorts of interesting theoretical questions in >> how (or if) one ought to aggregate dyads to create networks. But >> assuming that one has a series of matrices describing network change >> over time, is it possible to generate a "stability statistic" without >> an explicit model of the underlying network processes? (stable in >> the sense of repeating relations and sequences, rather than the sense >> of "fragility" or susceptibility to perturbations - 'tho they may be >> linked) Simple correlations of successive networks seems not to be >> so effective as it is very sensitive timescale of the patterns one is >> looking for, and sociomatrices tend to be very sparse and zero >> dominated. >> >> If anyone has experience/thoughts/references on this topic I'd >> greatly appreciate hearing about them, and will of course compile and >> resubmit refs, >> >> thanks, >> -skye >> >> _____________________________________________________________________ >> SOCNET is a service of INSNA, the professional association for social >> network researchers (http://www.sfu.ca/~insna/). To unsubscribe, send >> an email message to [log in to unmask] containing the line >> UNSUBSCRIBE SOCNET in the body of the message. >> > >-- >Richard Rothenberg, MD >Professor >Department of Family and Preventive Medicine >Emory University School of Medicine >69 Jesse Hill Jr. Drive SE >Atlanta, GA 30303 >email: [log in to unmask] >tele: 404-616-5606 >fax: 404-616-6847 > >_____________________________________________________________________ >SOCNET is a service of INSNA, the professional association for social >network researchers (http://www.sfu.ca/~insna/). To unsubscribe, send >an email message to [log in to unmask] containing the line >UNSUBSCRIBE SOCNET in the body of the message. > - - - - - - - - - - - - End of Original Message - - - - - - - - - - - - _____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers (http://www.sfu.ca/~insna/). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.