I need advice on how to set up a multilevel model. I have a mulitnomial outcome (TX_decision) with three levels of treatment decision (unchanged, reduced dose, Increased dose). These data are at the patient level but patients are nested within providers (and providers are nested within clinics but I only have 4 clinics so I am ignoring clinic as a Level III variable). My predictors are at all levels: continuous (patient age), binary (gender) and multinational (such as disease type ). Bivariate analyses (such as cross tabs) show some interesting relationship between the the predictors and the outcome. However, I feel that I need to take into account the fact that patients are nested within providers. I have about 175 patients and about 10 providers. I'd like to run the analyses in SPSS but not sure how to set this up within the GEE procedure. Thanks for any help you might be able to provide. Bill William N. Dudley, PhD 437-L Coleman BuildingProfessor - Public Health Education The School of Health and Human Sciences The University of North Carolina at Greensboro Greensboro, NC 27402-6170 See my research on ResearchGate VOICE 336.256 2475 |
Administrator
|
Bill, I'm curious about why you want to use GEE. My first thought was to use
GENLINMIXED. You could take a look at chapter 6 in "Multilevel Modeling of Categorical Outcomes Using IBM SPSS", by Heck, Thomas & Tabata. The chapter title is: "Two-Level Models with Multinomial and Ordinal Outcomes". http://www2.hawaii.edu/~ltabata/mlm/HTT2012.html That reminds me, I think your outcome variable is actually ordinal, not multinomial: reduced < unchanged < increased. HTH. William Dudley-2 wrote > I need advice on how to set up a multilevel model. I have a mulitnomial > outcome (TX_decision) with three levels of treatment decision (unchanged, > reduced dose, Increased dose). These data are at the patient level but > patients are nested within providers (and providers are nested within > clinics but I only have 4 clinics so I am ignoring clinic as a Level III > variable). > > My predictors are at all levels: continuous (patient age), binary (gender) > and multinational (such as disease type ). > > Bivariate analyses (such as cross tabs) show some interesting relationship > between the the predictors and the outcome. However, I feel that I need > to > take into account the fact that patients are nested within providers. I > have about 175 patients and about 10 providers. > > I'd like to run the analyses in SPSS but not sure how to set this up > within the GEE procedure. > > Thanks for any help you might be able to provide. > > Bill > > > -- > William N. Dudley, PhD > Professor - Public Health Education > The School of Health and Human Sciences > The University of North Carolina at Greensboro > 437-L Coleman Building > Greensboro, NC 27402-6170 > See my research on > GoogleScholar > <https://scholar.google.com/citations?user=ZiYmyb4AAAAJ&hl=en> > ResearchGate <https://www.researchgate.net/profile/William_Dudley> > VOICE 336.256 2475 > > [image: email signature image example.png] > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA > (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD ----- -- Bruce Weaver [hidden email] http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." NOTE: My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
--
Bruce Weaver bweaver@lakeheadu.ca http://sites.google.com/a/lakeheadu.ca/bweaver/ "When all else fails, RTFM." PLEASE NOTE THE FOLLOWING: 1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above. 2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/). |
In reply to this post by William Dudley-2
There are very strong parallels between complex samples and multilevel
models. They grew up in different statistical cultures. They definitely ate used on the same kind of data. When I retired in 2001 one of my hopes was to explore when and IF the two approaches lead to different estimates and/or conclusions. I'll never have time to to do this. It would be interesting if people for either stat culture were to run their data through the two approaches. There has been a long tradition of showing that what were thought to be different stat methods were special instances of a more general model. Maybe somebody would want to compare/contrast complex samples approaches to multilevel model approaches. ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Art Kendall
Social Research Consultants |
Free forum by Nabble | Edit this page |