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

I have done this to predict alter level outcomes controlling for ego and other alter characteristics (mostly predicting unprotected sex with a sex partner) with Stata using the GLLAMM package:  http://www.gllamm.org/

 

You need to set up the data set with repeated ego level observations for each alter level observation. The command that I have used is something like this:

 

gllamm alter_outcome   ego_var1 ego_var2 ego_var3 alter_var1 alter_var2 alter_var3, i(ncsid) family(binomial) link(logit) adapt nolog trace eform

 

alter_outcome = the dependent dichotomous variable such as unprotected sex

ego_var1-3: ego characteristics, such as beliefs about HIV, ego’s substance use, etc.

alter_var1-3: alter characteristics or ego’s report of relationship with specific alter, such as alter gender, age, relationship commitment between ego and alter, etc.

ncsid: unique variable identifying  ego.

 

You can read explanations of the model options (adapt, nolog, trace, eform) here:  http://www.gllamm.org/gllamm.hlp

 

Here are some examples where I used this approach:

 

 

Kennedy DP, Wenzel SL, Brown R, Tucker JS, Golinelli D. Unprotected Sex Among Heterosexually Active Homeless Men: Results from a Multi-level Dyadic Analysis. AIDS and Behavior. 2013;17(5):1655-1667.

 

Brown RA, Kennedy DP, Tucker JS, Golinelli D, Wenzel SL. Monogamy on the Street: A Mixed Methods Study of Homeless Men. Journal of Mixed Methods Research. 2013;7(4):328-346.

 

Kennedy DP, Tucker JS, Green HD, Golinelli D, Ewing BA. Unprotected sex of homeless youth: Results from a multilevel analysis of individual, social network, and relationship factors. AIDS and Behavior. 2012;16(7):2015-2032.

 

de la Haye K, Green HD, Kennedy DP, et al. Who Is Supporting Homeless Youth? Predictors of Support in Personal Networks. Journal of Research on Adolescence. 2012;22(4):604-616.

 

Kennedy DP, Wenzel SL, Tucker JS, et al. Unprotected Sex of Homeless Women Living in Los Angeles County: An Investigation of the Multiple Levels of Risk. Aids and Behavior. 2010;14(4):960-973.

 

 

From: Social Networks Discussion Forum [mailto:[log in to unmask]] On Behalf Of Yawo Kokuvi
Sent: Tuesday, May 16, 2017 7:50 AM
To: [log in to unmask]
Subject: [SOCNET] GEE and Egocentric Networks

 

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

Dear SOCNET terse,

 

Good afternoon. 

 

I am trying to fit a logistic regression model based on data from an egocentric/personal network. 

 

I've noticed that GEE is increasingly being used to  deal with non-independence of ego-alter observations.

 

I want to fit my model using GEE. But there is very little guidance out there on how to set up the procedure in SPSS/STATA (when network data is involved). 

 

I will be very grateful if anyone here some leads, references on how to fit this model in SPSS, Stata, R or SAS.

 

Thanks in advance,

 

Cy

_____________________________________________________________________ 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.


__________________________________________________________________________

This email message is for the sole use of the intended recipient(s) and
may contain information that is sensitive, proprietary, and/or privileged.
Any unauthorized review, use, disclosure or distribution is prohibited. If
you are not the intended recipient, please contact the sender by reply
email and destroy all copies of the original 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.