## SOCNET@LISTS.UFL.EDU

#### View:

 Message: [ First | Previous | Next | Last ] By Topic: [ First | Previous | Next | Last ] By Author: [ First | Previous | Next | Last ] Font: Proportional Font

Subject:

Re: rem.dyad terms in relevent in R

From:

Date:

Sun, 24 Mar 2013 11:16:49 +0100

Content-Type:

text/plain

Parts/Attachments:

 text/plain (152 lines)
 ```***** To join INSNA, visit http://www.insna.org ***** Hi Mark, Let's say we have three nodes, A, B, and C. In the following sequence of relational events, you can observe that ties tend to be reciprocated directly. There are exceptions after the second and after the fourth event, but otherwise we can observe reciprocal turn-taking. This means that we would get a positive PSAB-BA coefficient because we can observe reciprocity in 5 out of 7 transitions: A-B B-A A-C C-A B-A A-B B-A A-B In contrast, consider these two event sequences where A repeatedly sends information to B. In the first case, there are three other events that happen in the meantime. If this is a general pattern in the data, we will get a negative coefficient for RSndSnd because ties are established only after quite some time has passed. In the second case, A-B ties seem to be much more conditioned by recent A-B ties; one time there are two consecutive A-B ties, and twice we can observe one intermediate non-A-B tie. In this case, we would get a positive coefficient for RSndRSnd: A-B A-C B-C C-B A-B A-B A-C A-B A-B C-B A-B To be more precise, we do not count the number of intermediate extra-dyadic events, but we take the inverse of this value, for example 1/3 if the same tie could be observed in the third-previous round. And if we could never observe the same tie before, the value amounts to 1/infinity. If we only care about A-B ties in the previous example, we would have values of 1/infinity + 1/4 for A-B ties in the first sequence and values of 1/infinity + 1/2 + 1 + 1/2 for A-B events in the second sequence. In the following sequence, we would get a positive RRecSnd coefficient, but a negative PSAB-BA coefficient: A-B C-A B-A A-C A-B C-A There is always just one intermediate extra-dyadic tie, which means that ties tend to be based on fairly recent reciprocal ties (that is, 1/Inf + 1/Inf + 1/2 + 1/2 + 1/2 + 1/2 for RRecSnd). But we cannot observe any direct turn-taking in the conversation (that is, 0 + 0 + 0 + 0 + 0 + 0 for PSAB-BA). This explanation may have been a bit imprecise, but if you are interested in more details, you could look up the definition of these sufficient statistics in Carter Butts' 2008 article "A Relational Event Framework for Social Action" in Sociological Methods 38(1): 155-200. Finally, let me note that the most recent version of my texreg package for R is able to convert the model output of multiple rem.dyad objects to beautiful LaTeX or HTML tables or even MS Word documents. This is also true for ergm, stergm, and lnam objects as well as various multi-level models. There is also a function for displaying the output of multiple models side-by-side in the R console for easy model comparison. Take care, Philip Am 24.03.2013 00:27, schrieb Mark Tranmer: > Dear All, > > Can someone please help me with this query about the interpretation of > terms in the rem.dyad model in relevent in R? > > I am particularly interested in the distinction between this term: > > PSAB-BA: P-Shift effect (turn receiving) – AB->BA (dyadic) > > And these two terms: > > RRecSnd: Recency of receipt of actions from v′ affects v’s future rate > of sending to v′ > RSndSnd: Recency of sending to v′ affects v’s future rate of sending to v′ > > My guess is the PSAB-BA refers specifically to the very next event that > happens and whether this is a reciprocated action, whereas the other two > terms RRecSnd and RSndSnd are more general exchanges for recent events, > which makes me wonder how exactly "Recency" is defined. > > Can someone please give me an example of the difference of their > interperpration and/or point me to a relevant reference online or elsewhere? > > Thanks a lot! > > Cheers, > > Mark. > > Dr Mark Tranmer > Senior Lecturer > CCSR > Social Statistics > School of Social Sciences > Humanities Bridgeford Street Building > University of Manchester > Manchester > M13 9PL > > tel.: +44 (0) 161 275 4744 > fax: +44 (0) 161 275 4722 > > http://www.ccsr.ac.uk/staff/markhome1.htm > > > > > -- Dr. Philip Leifeld Eawag: Das Wasserforschungsinstitut des ETH-Bereichs Swiss Federal Institute of Aquatic Science and Technology Universität Bern, Institut für Politikwissenschaft University of Bern, Institute of Political Science Eawag, Überlandstr. 133, 8600 Dübendorf, Switzerland Phone: +41 58 765 5673 | E-mail: [log in to unmask] Office: FC-E18 | Homepage: http://www.philipleifeld.de _____________________________________________________________________ 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.```